WO2011114139A1 - Auto-antigen biomarkers for prostate cancer - Google Patents

Abstract

The presence of certain auto-antibodies indicates that a subject has prostate cancer. The auto- antibodies recognise antigens listed in Table 1 herein. These auto-antibodies and/or the antigens themselves can be used as biomarkers for assessing prostate cancer in a subject.

Description

AUTO-ANTIGEN BIOMARKERS FOR PROSTATE CANCER

This application claims the benefit of GB application 1004304.0, filed 15th March 2010, the complete contents of which are hereby incorporated herein by reference for all purposes.

TECHNICAL FIELD

This invention relates to biomarkers useful in the diagnosis, monitoring and/or treatment of prostate cancer.

BACKGROUND

Prostate cancer (PC) is a disease of the prostate, a gland in the male reproductive system. In a subset of men PC is aggressive and this form has a high mortality. It is currently difficult to determine at an early stage whether the cancer is relatively benign or aggressive. Failure to diagnose and treat aggressive forms can have negative consequences, but over-treatment of patients with relatively benign PC is undesirable.

Although remova l of prostatic tissue a nd pathologica l exa mi nation is currently the on ly accurate test for PC, it is preferable to minimise the number of avoidable surgical procedures. Thus a recommendation to perform a biopsy is normally given only after receiving the results of a digital rectal examination (DRE) and eva luating of the serum concentration of Prostate- Specific Antigen (PSA; kallikrein-3). PSA is currently the only molecular marker approved for use in the context of PC diagnosis. Tens of millions of PSA tests are performed annually worldwide but it has high false positive and significant false negative rates [1]. Reported specificities for the PSA test vary but in general are much less than 50%. A raised PSA level can indicate PC but it is also seen in other conditions of the prostate such as benign prostatic hypertrophy (BPH) and prostatitis. Thus the FDA recommends that the PSA test is used in conjunction with DRE.

Various forms of PSA assays have been developed to improve specificity, including comparison of bound versus free PSA and monitoring PSA concentrations over time (PSA velocity) [2], but the specificity and sensitivity of these tests are still low, resulting in many unnecessary prostate biopsies being performed every year.

The poor performance of PSA has resulted in the search for alternative biomarkers for PC e.g. the PCA3 or DD3 antigens [3-6], the serum markers of reference 7, the gene expression profiles of reference 8, the glycan profiles of reference 9, AMACR (alphamethylacyl CoA racemase), EPCA (early prostate carcinoma antigen), EPCA-2, gene promoter methylation, gene fusions including TMPRSS2:ERG gene fusions, peptide fingerprints, metabolites including sarcosine, etc. Still, however, no current test can discriminate between aggressive and non-aggressive cancers, although such a test would provide significant clinical benefit by enabling earlier active management of aggressive cancers while reducing unnecessary intervention for indolent cancers.

There is thus a need for further and improved new in vitro tests with better specificity and sensitivity to enable non-invasive diagnosis of PC. Preferably the discriminatory power of this diagnostic test should be sufficiently high to support population screening approaches, which PSA cannot achieve [10]. Idea lly it should also be useful for the detection of PC at a n early stage. It is an object of the invention to meet these needs.

DISCLOSURE OF THE INVENTION

The invention is based on the identification of correlations between PC and the level of auto- antibodies against certain auto-antigens. The inventors have identified antigens for which the level of a uto-antibodies ca n be used to indicate that a subject has prostate ca ncer. Autoantibodies against these antigens are present at significantly different levels in men with PC and without PC a nd so the auto-antibodies and their antigens function as biomarkers of prostate cancer. Detection of the biomarkers in a subject sample ca n th us be used to improve the diagnosis, prognosis and monitoring of PC. Advantageously, the invention can be used to distinguish between prostate cancer and other diseases of the prostate such as benign prostatic hypertrophy (BPH) and prostatitis where inflammation and raised PSA levels are common.

The inventors have identified 24 such biomarkers and the invention uses at least one of these to assist in the diagnosis of PC by measuring level(s) of auto-antibodies against the antigen(s) and/or the level(s) of the antigen(s) themselves. The biomarker can be (i) auto-antibody which binds to an antigen in Table 1 and/or (ii) an antigen in Table 1, but is preferably the former.

The invention thus provides a method for analysing a subject sample, comprising a step of determining the level of a Table 1 biomarker in the sample, wherein the level of the biomarker provides a diagnostic indicator of whether the subject has prostate cancer. Analysis of a single Table 1 biomarker can be performed, and detection of the auto- antibody/antigen can provide a useful diagnostic indicator for PC even without considering any of the other Table 1 biomarkers. The sensitivity and specificity of diagnosis can be improved, however, by combining data for multiple biomarkers. It is thus preferred to analyse more than one Table 1 biomarker. Analysis of two or more different biomarkers (a "panel") can enhance the sensitivity and/or specificity of diagnosis compared to analysis of a single biomarker. Each different biomarker in a panel is shown in a different row in Table 1 i.e. measuring both auto-antibody which binds to an antigen listed in Table 1 and the antigen itself is measurement of a single biomarker rather than of a panel. Thus the invention provides a method for a nalysing a subject sample, comprising a step of determining the levels of x different bioma rkers of Ta ble 1, wherein the levels of the biomarkers provide a diagnostic indicator of whether the subject has prostate cancer. The value of x is 2 or more e.g. 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or more (e.g. up to 24). These pa nels may i nclude (i) a ny specific one of the 24 biomarkers i n Ta ble 1 in com bination with (ii) any of the other 23 biomarkers in Table 1. Suitable panels are described below and panels of pa rticula r interest include those listed in Ta bles 2 to 16. Preferred pa nels have from 2 to 15 biomarkers, as using >15 of them adds little to sensitivity and specificity.

The Ta ble 1 biomarkers can be used in combination with one or more of: (a) known biomarkers for prostate ca ncer, which may or may not be a uto-antibodies or a ntigens; a nd/or (b) other information about the subject from whom a sample was taken e.g. age, genotype (genetic va riations ca n affect a uto-a nti body profi les [11]), weight, other clinically-relevant data or phenotypic information.; and/or (c) other diagnostic tests or clinical indicators for prostate cancer. Such combinations can enhance the sensitivity and/or specificity of diagnosis. Thus the invention provides a method for ana lysing a subject sample, comprising a step of determining:

(a) the level(s) of y Table 1 bioma rker(s), wherein the levels of the biomarkers provide a diagnostic indicator of whether the subject has prostate cancer; and also one or both of:

(b) if a sample from the subject contains a known biomarker selected from the group consisting of PSA antigen, PCA3 antigen and/or mRNA, DD3 antigen and/or mRNA, AMACR antigen and/or mRNA, EPCA antigen and/or mRNA, EPCA-2 a ntigen and/or mRNA, and sarcosine (and optionally, any other known biomarkers e.g. see above); wherein detection of the known biomarker provides a second diagnostic indicator of whether the su bject has prostate cancer;

(c) the subject's age, and combining the different diagnostic indicators to provide an aggregate diagnostic indicator of whether the subject has prostate cancer.

The sam ples used in (a) a nd (b) may be the same or different. I n one embodiment the method uses (a) and (b). I n another em bodiment the method uses (a) and (c). I n a nother em bodiment the method uses (a), (b) and (c). The biomarkers listed in Table 18 can also be utilised. Thus the invention also provides a method for ana lysing a subject sample, comprising a step of determining: (a) the level(s) of y Table 1 bioma rker(s), wherein the levels of the biomarkers provide a diagnostic indicator of whether the subject has prostate cancer; and also one, two or three of:

(b) if a sample from the subject contains a known biomarker selected from the group consisting of PSA antigen, PCA3 antigen and/or mRNA, DD3 antigen and/or mRNA, AMACR antigen and/or mRNA, EPCA antigen and/or mRNA, EPCA-2 a ntigen and/or mRNA, and sarcosine (and optionally, any other known biomarkers e.g. see above); wherein detection of the known biomarker provides a second diagnostic indicator of whether the subject has prostate cancer; and/or

(c) the level(s) of at least one Table 18 biomarker(s), wherein the levels of the Ta ble 18 biomarker(s) provides a further diagnostic indicator of whether the subject has prostate cancer; and/or

(d) the subject's age, a nd com bi ni ng the different diagnostic indicators to provide a n aggregate diagnostic indicator of whether the subject has prostate cancer. The samples used in (a) and (b) and (c) may be the same or different. I n one embodiment the method uses (a ) a nd (b) . I n a nother e m bodi me nt the method uses (a ) a nd (c) . I n a nothe r embodiment the method uses (a) a nd (d). I n another em bodiment the method uses (a), (b) and (c). I n a nother e m bodi ment the method uses (a ), ( b) a nd (d) . I n a nother e m bodi ment the method uses (a), (c) and (d). In another embodiment the method uses (a), (b), (c) and (d). The value of y is 1 or more e.g. 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 {e.g. up to 24). When y >1 the invention uses a panel of different Table 1 biomarkers.

The invention also provides, in a method for diagnosing if a subject has prostate cancer, an improvement consisting of determining in a sample from the subject the level(s) of y biomarker(s) of Table 1, wherein the level(s) of the biomarker(s) provide a diagnostic indicator of whether the subject has prostate cancer.

The i nvention a lso provides a method for diagnosi ng a subject as havi ng prostate ca ncer, com prising steps of: (i) determining the levels of y biomarkers of Table 1 in a sample from the subject; and (ii) comparing the determination from step (i) to data obtained from samples from subjects without prostate cancer and/or from subjects with prostate cancer, wherein the comparison provides a diagnostic indicator of whether the subject has prostate cancer. The comparison in step (ii) can use a classifier algorithm as discussed in more detail below. The invention also provides a method for monitoring development of prostate cancer in a subject, comprising steps of: (i) determining the levels of zj biomarker(s) of Table 1 in a first sample from the subject taken at a first time; and (ii) determining the levels of z₂ biomarker(s) of Table 1 in a second sample from the subject taken at a second time, wherein: (a) the second time is later than the first time; (b) one or more of the z₂ biomarker(s) were present in the first sample; and (c) a change in the level(s) of the biomarker(s) in the second sample compared with the first sample indicates that prostate cancer is in remission or is progressing. Thus the method monitors the biomarker(s) over time, with changi ng levels indicating whether the disease is getting better or worse.

The disease development can be either an improvement or a worsening, and this method may be used in various ways e.g. to monitor the natural progress of a disease, or to monitor the efficacy of a therapy being administered to the subject. Thus a subject may receive a therapeutic agent before the first time, at the first time, or between the first time and the second time. Increased levels of antibodies against a particular antigen may be due to "epitope spreading", in which additional antibodies or antibody classes are raised to antigens against which an antibody response has already been mounted [12].

The value of Zj is 1 or more e.g. 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 {e.g. up to 24). The value of z₂ is 1 or more e.g. 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 {e.g. up to 24). The values of zj and z₂ may be the same or different. If they are different, it is usual that zj>z₂ as the later analysis (z₂) can focus on biomarkers which were already detected in the earlier analysis; in other embodiments, however, z₂ can be larger than zi e.g. if previous data have indicated that an expanded panel should be used; in other embodiments

e.g. so that, for convenience, the same panel can be used for both analyses. When ZJ>1 or z₂>l, the biomarkers are different biomarkers.

The invention also provides a method for monitoring development of prostate cancer in a subject, comprising steps of: (i) determining the level of at least I I/J Table 1 biomarkers in a first sample taken at a first time from the subject; and (ii) determining the level of at least w₂ Table 1 biomarkers in a second sample taken at a second time from the subject, wherein: (a) the second time is later than the first time; (b) at least one biomarker is common to both the I I/J and w₂ biomarkers; (c) the level of at least one biomarker common to both the I I/J and w₂ biomarkers is different in the first and second samples, thereby indicating that the prostate cancer is progressing or regressing. Thus the method monitors the range of biomarkers over time, with a broadening in the number of detected biomarkers indicating that the disease is getting worse. As mentioned above, this method may be used to monitor disease development in various ways.

The value of v j is 1 or more e.g. 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 (e.g. up to 24). The value of w₂ is 2 or more e.g. 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 {e.g. up to 24). The values of I I/J and w₂ may be the same or different. If they are different, it is usual that w₂>w_lt as the later analysis should focus on a biomarker panel that is at least as wide as the number already detected in the earlier analysis. There will usually be an overlap between the wi and w₂ biomarkers (including situations where they are the same, such that the same biomarkers are measured at two time points) but it is also possible for wi and w₂ to have no biomarkers in common.

Where the methods involve a first time and a second time, these times may differ by at least 1 day, 1 week, 1 month or 1 year. Samples may be taken regularly. The methods may involve measuring biomarkers in more than 2 samples taken at more than 2 time points i.e. there may be a 3rd sample, a 4th sample, a 5th sample, etc. The invention also provides a diagnostic device for use in diagnosis of prostate cancer, wherein the device permits determination of the level(s) of y Table 1 biomarkers. The value of y is defined above. The device may also permit determination of whether a sample contains one or more of the known PC biomarkers mentioned above e.g. PSA and/or PCA3.

The invention also provides a kit comprising (i) a diagnostic device of the invention and (ii) instructions for using the device to detect y of the Table 1 biomarkers. The value of y is defined above. The kit is useful in the diagnosis of prostate cancer.

The invention also provides a kit comprising reagents for measuring the levels of x different Table 1 biomarkers. The kit may also include reagents for determining whether a sample contains one or more of the known PC biomarkers mentioned above e.g. PSA and/or PCA3. The value of x is defined above. The kit is useful in the diagnosis of prostate cancer.

The invention also provides a kit comprising components for preparing a diagnostic device of the invention. For instance, the kit may comprise individual detection reagents for x different biomarkers, such that an array of those x biomarkers can be prepared.

The invention also provides a product comprising (i) one or more detection reagents which permit measurement of x different Table 1 biomarkers, and (ii) a sample from a subject.

The invention also provides a software product comprising (i) code that accesses data attributed to a sample, the data comprising measurement of y Table 1 biomarkers, and (ii) code that executes an algorithm for assessing the data to represent a level of y of the biomarkers in the sample. The software product may also comprise (iii) code that executes an algorithm for assessing the result of step (ii) to provide a diagnostic indicator of whether the subject has prostate cancer. As discussed below, suitable algorithms for use in part (iii) include support vector machine algorithms, artificial neural networks, tree-based methods, genetic programming, etc. The algorithm can preferably classify the data of part (ii) to distinguish between PC subjects and non-PC subjects based on measured biomarker levels in samples taken from such subjects. The invention also provides methods for training such algorithms.

The invention also provides a computer which is loaded with and/or is running a software product of the invention.

The invention a lso extends to methods for com municating the results of a method of the invention. This method may involve com municating assay results and/or diagnostic results. Such comm unication may be to, for exa mple, technicians, physicia ns or patients. I n some embodiments, detection methods of the invention will be performed in one country and the results will be communicated to a recipient in a different country.

The i nvention a lso provides a n isolated a nti body (preferably a hu ma n a nti body) which recognises one of the antigens listed in Table 1. The invention also provides an isolated nucleic acid encoding the heavy and/or light chain of the antibody. The invention also provides a vector comprising this nucleic acid, and a host cell comprising this vector. The invention also provides a method for expressing the antibody comprising culturing the host cell under conditions which permit production of the antibody. The invention also provides derivatives of the human antibody e.g. F(ab')₂ and F(ab) fragments, Fv fragments, single-chain antibodies such as single chain Fv molecules (scFv), minibodies, dAbs, etc.

The invention also provides the use of a Table 1 biomarker as a biomarker for PC.

The invention also provides the use of x different Table 1 biomarkers as biomarkers for prostate cancer. The value of x is defined above. These may include (i) any specific one of the 24 biomarkers in Table 1 in combination with (ii) any of the other 23 biomarkers in Table 1.

The invention also provides the use as combined biomarkers for prostate cancer of (a) at least y Table 1 biomarker(s) and (b) PSA, PCA3, DD3, AMACR, EPCA, EPCA-2 and/or sarcosine (and optionally, any other known biomarkers e.g. see above). The value of y is defined above. When y>l the invention uses a panel of biomarkers of the invention. Biomarkers of the invention

Auto-antibodies against 105 different human antigens have been identified and these can be used as PC biomarkers. Details of the 105 antigens are given in Table 17. Within the 105 antigens, 24 human antigens are particularly useful for distinguishing between samples from subjects with PC and from subjects without PC. Details of these 24 antigens are given in Table 1. Further auto-antibody biomarkers can be used in addition to these 24 (e.g. any of the other biomarkers listed in Table 17). The sequence listing provides an example of a natural coding sequence for each of these antigens. These specific coding sequences are not limiting on the invention, however, and auto-antibody biomarkers may recognise variants of polypeptides encoded by these natural sequences (e.g. allelic variants, polymorphic forms, mutants, splice variants, or gene fusions), provided that the variant has an epitope recognised by the autoantibody. Details on allelic variants of or mutations in human genes are available from various sources, such as the ALFRED database [13] or, in relation to disease associations, the OMIM [14] and HGMD [15] databases. Details of splice variants of human genes are available from various sources, such as ASD [16].

As mentioned above, detection of a single Table 1 biomarker can provide useful diagnostic information, but each biomarker might not individually provide information which is useful i.e. auto-antibodies against a Table 1 antigen may be present in some, but not all, subjects with prostate cancer. An inability of a single biomarker to provide universal diagnostic results for all subjects does not mean that this biomarker has no diagnostic utility, however, or else PSA also would not be useful; rather, any such inability means that the test results (as in all diagnostic tests) have to be properly understood and interpreted.

To address the possibility that a single biomarker might not provide universal diagnostic results, and to increase the overall confidence that an assay is giving sensitive and specific results across a disease population, it is advantageous to analyse a plurality of the Table 1 biomarkers (i.e. a panel). For instance, a negative signal for a particular Table 1 antigen is not necessarily indicative of the absence of PC (just as a low PSA concentration is not), confidence that a subject does not have PC increases as the number of negative results increases. For example, if all 24 biomarkers are tested and are negative then the result provides a higher degree of confidence than if only 1 biomarker is tested and is negative. Thus biomarker panels are most useful for enhancing the distinction seen between diseased and non-diseased samples. As mentioned above, though, preferred panels have from 2 to 15 biomarkers as the burden of measuring a higher number of markers is usually not rewarded by better sensitivity or specificity. Preferred panels are given below. Where a biomarker or panel provides a strong distinction between PC and non-PC subjects then a method for ana lysing a subject sample can function as a method for diagnosing if a subject has prostate cancer. As with many diagnostic tests, however, and as is already known for the PSA test, a method may not always provide a definitive diagnosis and so a method for analysing a subject sample can sometimes function only as a method for aiding in the diagnosis of prostate ca ncer, or as a method for contri buti ng to a diagnosis of prostate cancer, where the method's result may imply that the subject has prostate cancer (e.g. the disease is more likely than not) a nd/or may confirm other diagnostic indicators (e.g. passed on clinica l symptoms). Dealing with these considerations of certainty/uncertainty is well known in the diagnostic field. The subject

The invention is used for diagnosing disease in a subject. The subject will be male. The subject will usually be at least 20 years old (e.g. >25, >30, >35, >40, >45, >50, >55, >60, >65, >70). They will usua lly be at least 50 years old as the risk of PC increases in these men, and for these subjects it may be appropriate to offer a screening service for Table 1 bioma rkers. The subject may be pre-symptomatic for PC or may already be displaying clinica l symptoms. For pre-symptomatic subjects the invention is useful for predicting that sym ptoms may develop in the future if no preventative action is taken. For subjects already displaying clinical symptoms, the invention may be used to confi rm or resolve a nother diagnosis. The subject may a l ready have begun treatment for PC. I n some embodiments the subject may a lready be known to be predisposed to development of PC e.g. due to family or genetic links. I n other embodiments, the subject may have no such predisposition, and may develop the disease as a result of environmental factors e.g. as a result of exposure to pa rticular chemica ls (such as toxins or pharmaceuticals), as a result of diet [17], as a result of infection, etc. Because the invention can be implemented relative easily and chea ply it is not restricted to being used in patients who a re already suspected of having PC. Rather, it can be used to screen the general population or a high risk population e.g. men at least 20 years old, as listed above.

The subject will typically be a human being. I n some embodiments, however, the invention is useful in non-human organisms e.g. mouse, rat, rabbit, guinea pig, cat, dog, horse, pig, cow, or non-human primate (monkeys or apes, such as macaques or chimpanzees). I n non-human em bodiments, a ny detection a ntigens used with the i nvention wi ll typica lly be based on the releva nt non-huma n ortholog of the h uman a ntigens disclosed herein. I n some embodiments animals can be used experimentally to monitor the impact of a therapeutic on a particular biomarker.

The sample

The invention analyses samples from subjects. Many types of sample can include auto- antibodies and/or antigens suitable for detection by the invention, but the sample will typically be a body fluid. Suitable body fluids include, but are not limited to, blood, serum, plasma, saliva, prostate tissue, prostate fluid (i.e. fluid which immediately surrounds the prostate in vivo), prostatic secretions, lymphatic fluid, a wound secretion, urine, faeces, mucus, sweat, tears and/or cerebrospinal fluid. The sample is typically serum or plasma. In some embodiments, a method of the invention involves an initial step of obtaining the sample from the subject. In other embodiments, however, the sample is obtained separately from and prior to performing a method of the invention. After a sample has been obtained then methods of the invention are generally performed in vitro.

Detection of biomarkers may be performed directly on a sample taken from a subject, or the sample may be treated between being taken from a subject and being analysed. For example, a blood sample may be treated to remove cells, leaving antibody-containing plasma for analysis, or to remove cells and various clotting factors, leaving antibody-containing serum for analysis. Faeces samples usually require physical treatment prior to protein detection e.g. suspension, homogenisation and centrifugation. For some body fluids, though, such separation treatments are not usually required (e.g. tears or saliva) but other treatments may be used. For example, various types of sample may be subjected to treatments such as dilution, aliquoting, sub-sampling, heating, freezing, irradiation, etc. between being taken from the body and being analysed e.g. serum is usually diluted prior to analysis. Also, addition of processing reagents is typical for various sample types e.g. addition of anticoagulants to blood samples.

Biomarker detection

The invention involves determining the level of Table 1 biomarker(s) in a sample. Immunochemical techniques for detecting antibodies against specific antigens are well known in the art, as are techniques for detecting specific antigens themselves. Detection of an antibody will typically involve contacting a sample with a detection antigen, wherein a binding reaction between the sample and the detection antigen indicates the presence of the antibody of interest. Detection of an antigen will typically involve contacting a sample with a detection antibody, wherein a binding reaction between the sample and the detection antibody indicates the presence of the antigen of interest. Detection of an antigen can also be determined by non-immunologica l methods, depending on the nature of the antigen e.g. if the antigen is an enzyme then its enzymatic activity can be assayed, or if the antigen is a receptor then its binding activity can be assayed, etc. For example, the MAPK9 ki nase ca n be assayed usi ng methods such as those disclosed in reference 18. A detection antigen for a biomarker antibody can be a natura l a ntigen recognised by the auto-antibody (e.g. a mature human protein disclosed in Table 1), or it may be an antigen comprising an epitope which is recognized by the auto-antibody. It may be a recombinant protein or synthetic peptide. Where a detection antigen is a polypeptide its amino acid sequence can vary from the natural sequences disclosed above, provided that it has the ability to specifica lly bind to an auto-antibody of the invention (i.e. the binding is not non-specific and so the detection antigen will not arbitra rily bind to antibodies in a sample). It may even have little in common with the natura l sequence (e.g. a mimotope, a n aptamer, etc. ). Typically, though, a detection antigen will comprise an amino acid sequence (i) having at least 90% (e.g. >91%, >92%, >93%, >94%, >95%, >96%, >97%, >98%, >99%) sequence identity to the relevant SEQ I D NO disclosed herein across the length of the detection antigen, and/or (ii) comprising at least one epitope from the releva nt SEQ. I D N O disclosed herei n. Thus the detection a ntigen may be one of the variants discussed above.

Epitopes are the parts of an antigen that are recognised by and bind to the antigen binding sites of a ntibodies and are also known as "antigenic determinants". An epitope-containing fragment may contain a linea r epitope from within a SEQ I D NO and so may comprise a fragment of at least n consecutive amino acids of the SEQ I D NO :, wherein n may be 7 or more (e.g. 8, 10, 12, 14, 16, 18, 20, 25, 30, 35, 40, 50, 60, 70, 80, 90, 100, 150, 200, 250 or more). B-cell epitopes can be identified empirica lly (e.g. usi ng P EPSCAN [19,20] or simila r methods), or they can be predicted e.g. usi ng the Jameson-Wolf a ntigenic i ndex [21], ADEPT [22], hydrophilicity [23], antigenic index [24], MAPITOPE [25], SEPPA [26], matrix-based approaches [27], the amino acid pair antigenicity scale [28], or any other suitable method e.g. see ref.29. Predicted epitopes ca n readily be tested for actual immunochemica l reactivity with samples.

Detection antigens can be purified from human sources but it is more typical to use recombinant antigens (particula rly where the detection antigen uses sequences which are not present in the natural antigen e.g. for attachment). Various systems are available for recombinant expression, and the choice of system may depend on the auto-antibody to be detected. For example, prokaryotic expression (e.g. using E.coli) is useful for detecting many auto-antibodies, but if a n auto-antibody recognises a glycoprotein then eukaryotic expression may be required. Similarly, if an auto-antibody recognises a specific discontinuous epitope then a recombinant expression system which provides correct protein folding may be required.

The detection antigen may be a fusion polypeptide with a first region and a second region, wherein the first region can react with an auto-antibody in a sample and the second region can react with a substrate to immobilise the fusion polypeptide thereon.

A detection antibody for a biomarker antigen can be a monoclonal antibody or a polyclonal antibody. Typically it will be a monoclonal antibody. The detection antibody should have the ability to specifically bind to a Table 1 antigen (i.e. the binding is not non-specific and so the detection antibody will not arbitrarily bind to other antigens in a sample). Various assay formats can be used for detecting biomarkers in samples. For example, the invention may use one or more of western blot, immunoprecipitation, silver staining, mass spectrometry (e.g. MALDI-MS), conductivity-based methods, dot blot, slot blot, colorimetric methods, fluorescence-based detection methods, or any form of immunoassay, etc. The binding of antibodies to antigens can be detected by any means, including enzyme-linked assays such as ELISA, radioimmunoassays (RIA), immunoradiometric assays (IRMA), immunoenzymatic assays (IEMA), DELFIA™ assays (dissociation-enhanced lanthanide fluorescent immunoassay), surface plasmon resonance or other evanescent light techniques (e.g. using planar waveguide technology), label-free electrochemical sensors, etc. Sandwich assays are typical for immunological methods. In embodiments where multiple biomarkers are to be detected an array-based assay format is preferable, in which a sample that potentially contains the biomarkers is simultaneously contacted with multiple detection reagents (antibodies and/or antigens) in a single reaction compartment. Antigen and antibody arrays are well known in the art e.g. see references 30-36, including arrays for detecting auto-antibodies. Such arrays may be prepared by various techniques, such as those disclosed in references 37-41, which are particularly useful for preparing microarrays of correctly-folded polypeptides to facilitate binding interactions with auto-antibodies. It has been estimated that most B-cell epitopes are discontinuous and such epitopes are known to be important in diseases with an autoimmune component. For example, in autoimmune thyroid diseases, auto-antibodies arise to discontinuous epitopes on the immunodominant region on the surface of thyroid peroxidase and in Goodpasture disease auto-antibodies arise to two major conformational epitopes. Protein arrays which have been developed to present correctly-folded polypeptides displaying native structures and discontinuous epitopes are therefore particularly well suited to studies of diseases where auto-antibody responses occur [34]. Methods and apparatuses for detecting binding reactions on protein arrays are now standard in the art. Preferred detection methods are fluorescence-based detection methods. To detect biomarkers which have bound to immobilised proteins a sandwich assay is typical e.g. in which the primary antibody is an auto-antibody from the sample and the secondary antibody is a labelled anti-sample antibody (e.g. an anti-human antibody).

Where a biomarker is an auto-antibody the invention will generally detect IgG antibodies, but detection of auto-antibodies with other subtypes is also possible e.g. by using a detection reagent which recognises the appropriate class of auto-antibody (IgA, IgM, IgE or IgD rather than Ig). The assay format may be able to distinguish between different antibody subtypes and/or isotypes. Different subtypes [42] and isotypes [43] can influence auto-antibody repertoires. For instance, a sandwich assay can distinguish between different subtypes by using differentially-labelled secondary antibodies e.g. different labels for anti-lgG and anti-lgM.

As mentioned above, the invention provides a diagnostic device which permits determination of whether a sample contains Table 1 biomarkers. Such devices will typically comprise one or more antigen(s) and/or antibodies immobilised on a solid substrate (e.g. on glass, plastic, nylon, etc.). Immobilisation may be by covalent or non-covalent bonding (e.g. non-covalent bonding of a fusion polypeptide, as discussed above, to an immobilised functional group such as an avidin [39] or a bleomycin-family antibiotic [41]). Antigen arrays are a preferred format, with detection antigens being individually addressable. The immobilised antigens will be able to react with auto-antibodies which recognise a Table 1 antigen.

In some embodiments, the solid substrate may comprise a strip, a slide, a bead, a well of a microtitre plate, a conductive surface suitable for performing mass spectrometry analysis [44], a semiconductive surface [45,46], a surface plasmon resonance support, a planar waveguide technology support, a microfluidic devices, or any other device or technology suitable for detection of antibody-antigen binding.

Where the invention provides or uses an antigen array for detecting a panel of auto-antibodies as disclosed herein, in some embodiments the array may include only antigens for detecting these auto-antibodies. In other embodiments, however, the array may include polypeptides in addition to those useful for detecting the auto-antibodies. For example, an array may include one or more control polypeptides. Suitable positive control polypeptides include an anti-human immunoglobulin antibody, such as an anti-lgM antibody, an anti-lgG antibody, an anti-lgA antibody, an anti-lgE antibody or combinations thereof. Other suitable positive control polypeptides which can bind to sample antibodies include protein A or protein G, typically in recombinant form. Suitable negative control polypeptides include, but are not limited to, β-galactosidase, serum albumins (e.g. BSA or HSA), protein tags, bacterial proteins, yeast proteins, citrullinated polypeptides, etc. Negative control features on an array can also be polypeptide-free e.g. buffer alone, DNA, etc. An array's control features are used during performance of a method of the invention to check that the method has performed as expected e.g. to ensure that expected proteins are present (e.g. a positive signal from serum proteins in a serum sample) and that unexpected substances are not present (e.g. a positive signal from an a rray spot of buffer alone would be unexpected).

I n an antigen a rray of the invention, at least 10% (e.g. >20%, >30%, >40%, >50%, >60%, >70%, >80%, >90%, >95%, or more) of the total number of different proteins present on the array may be for detecting auto-antibodies as disclosed herein.

An antigen array of the invention may include one or more replicates of a detection antigen and/or control feature e.g. duplicates, triplicates or quadruplicates. Replicates provide redunda ncy, provide intra-array controls, and facilitate inter-a rray com parisons.

An antigen array of the invention may include detection antigens for more than just the 24 different auto-a ntibodies described here, but preferably it can detect antibodies against fewer than 10000 antigens (e.g. <5000, <4000, <3000, <2000, <1000, <500, <250, <100, etc. ).

An array is advantageous because it allows simultaneous detection of multiple biomarkers in a sample. Such simultaneous detection is not mandatory, however, and a panel of bioma rkers can also be evaluated in series. Thus, for instance, a sample could be split into sub-samples and the sub-samples could be assayed in series. I n this embodiment it may not be necessary to complete ana lysis of the whole pa nel e.g. the diagnostic indicators obtained on a subset of the panel may indicate that a patient has PC without requiring analysis of any further members of the panel. Such incom plete analysis of the panel is encompassed by the invention because of the intention or potential of the method to analyse the complete panel. As mentioned above, some embodiments of the invention can include a contribution from known tests for PC, such as PSA and/or PCA3 tests. Any known tests can be used e.g. tota l PSA score, PSA velocity, the PROGENSA™ assay for urinary PCA3 mRNA, etc. Typically, PSA levels less than 4ng/ml in blood are considered as normal, 4-10ng/ml is suspicious, a nd >10ng/ml is high.

Thus an array of the invention (or any other assay format) may also provide an assay for one or more of these additiona l markers e.g. an array may include a PSA spot.

Data interpretation

The invention involves a step of determining the level of Table 1 biomarker(s). I n some embodiments of the invention this determination for a pa rticular marker can be a sim ple yes/no determination, whereas other embodiments may require a quantitative or semiquantitative determination, still other embodiments may involve a relative determination (e.g. a ratio relative to another marker, or a measurement relative to the same marker in a control sample), and other embodiments may involve a threshold determination (e.g. a yes/no determination whether a level is above or below a threshold). Usually biomarkers will be measured to provide quantitative or semi-quantitative results (whether as relative concentration, absolute concentration, titre, etc.) as this gives more data for use with classifier algorithms.

Usually the raw data obtained from an assay for determining the presence, absence, or level (absolute or relative) require some sort of manipulation prior to their use. For instance, the nature of most detection techniques means that some signal will sometimes be seen even if no antigen/antibody is actually present and so this noise may be removed before the results are interpreted. Similarly, there may be a background level of the antigen/antibody in the general population which needs to be compensated for. Data may need scaling or standardising to facilitate inter-experiments comparisons. These and similar issues, and techniques for dealing with them, are well known in the immunodiagnostic area.

Various techniques are available to compensate for background signal in a particular experiment. For example, replicate measurements will usually be performed (e.g. using multiple features of the same detection antigen on a single array) to determine intra-assay variation, and average values from the replicates can be compared (e.g. the median value of binding to quadruplicate array features). Furthermore, standard markers can be used to determine inter-assay variation and to permit calibration and/or normalisation e.g. an array can include one or more standards for indicating whether measured signals should be proportionally increased or decreased. For example, an assay might include a step of analysing the level of one or more control marker(s) in a sample e.g. levels of an antigen or antibody unrelated to PC. Signal may be adjusted according to distribution in a single experiment. For instance, signals in a single array experiment may be expressed as a percentage of interquartile differences e.g. as [observed signal - 25th percentile] / [75th percentile - 25th percentile]. This percentage may then be normalised e.g. using a standard quantile normalization matrix, such as disclosed in reference 47, in which all percentage values on a single array are ranked and replaced by the average of percentages for antigens with the same rank on all arrays. Overall, this process gives data distributions with identical median and quartile values. Data transformations of this type are standard in the art for permitting valid inter-array comparisons despite variation between different experiments. The level of a biomarker relative to a single baseline level may be defined as a fold difference. Normally it is desirable to use techniques that can indicate a change of at least 1.5-fold e.g. >1.75-fold, >2-fold, >2.5-fold, >5-fold, etc.

As well as compensating for variation which is inherent between different experiments, it can also be important to compensate for background levels of a biomarker which are present in the genera l population . Again, suita ble tech niques a re wel l known . For exa m ple, levels of a particular antigen or auto-a ntibody in a sa m ple will usua lly be measured quantitatively or semi-quantitatively to permit comparison to the background level of that biomarker. Various controls ca n be used to provide a suita ble baseline for compa rison, a nd choosing suita ble controls is routine in the diagnostic field. Further details of suitable controls are given below.

The measured level(s) of Table 1 biomarker(s), after any compensation/normalisation/eic, can be transformed into a diagnostic result in various ways. This transformation may involve an algorithm which provides a diagnostic result as a function of the measured level(s). Where a panel is used then each individual biomarker may make a different contribution to the overall diagnostic result and so two biomarkers may be weighted differently.

The creation of algorithms for converting measured levels or raw data into scores or results is well known in the art. For example, linear or non-linear classifier algorithms can be used. These a lgorith ms ca n be tra i ned usi ng data from a ny pa rticu la r tech niq ue for measuring the marker(s). Suita ble training data will have been obtained by measuring the biomarkers in "case" and "control" samples i.e. samples from subjects known to suffer from PC and from subjects known not to suffer from PC. Most useful ly the control samples will also include samples from subjects with a related disease which is to be distinguished from the disease of interest e.g. it is useful to train the algorithm with data from BPH subjects and/or with data from subjects with cancer(s) other than PC. The classifier algorithm is modified until it can distinguish between the case and control samples e.g. by adding or removing markers from the analysis, by changes in weighting, etc. Thus a method of the invention may include a step of analysing biomarke r leve ls i n a su bject's sa m ple by usi ng a classifie r a lgo rith m which distinguishes between PC subjects and non-PC subjects based on measured biomarker levels in samples taken from such subjects.

Various suitable classifier algorithms are available e.g. linear discriminant analysis, na^'ive Bayes classifiers, perceptrons, neural networks, support vector machines (SVM) [48] and genetic programming (GP) [49]. GP is particularly useful as it generally selects relatively small numbers of biomarkers and overcomes the problem of trapping in a local maximum which is inherent in many other classification methods. SVM-based approaches have previously been used for PC diagnosis by classifying images of prostate tissue [50,51], mass spectrometry proteomic data [52], patient data [53], or gene expression levels [8]. They have also been used for analysing auto-antibodies in general cancers [54]. The inventors have confirmed that both SVM and GP approaches can be trained to distinguish the auto-antibody/antigen biomarker profiles of healthy subjects from PC subjects with similar sensitivity and specificity i.e. the biomarkers are not dependent on a single method of analysis. Moreover, these approaches can potentially distinguish PC subjects from subjects with (i) other forms of cancer and (ii) BPH. The 24 biomarkers in Table 1 can be used to train such algorithms to reliably make such distinctions.

It will be appreciated that, although there may be some biomarkers in Table 1 which always give a negative absolute signal when contacted with negative control samples (and thus any positive signal is immediately indicative of PC), it is more common that a biomarker will give at least a low absolute signal (and thus that a disease-indicating positive signal requires detection of auto-antibody levels above that background level). Thus references herein detecting a biomarker may not be references to absolute detection but rather (as is standard in the art) to a level above the levels seen in an appropriate negative control. Similarly, for markers where an absence or decrease is associated with disease then the skilled person will understand that measured levels should be below the levels seen in an appropriate negative control. Such controls may be assayed in parallel to a test sample but it can be more convenient to use an absolute control level based on empirical data, or to analyse data using an algorithm which can (e.g. by previous training) use biomarker levels to distinguish samples from disease patients vs. non-disease patients.

The level of a particular biomarker in a sample from a PC-diseased subject may be above or below the level seen in a negative control sample. Antibodies that react with self-antigens occur naturally in healthy individuals and it is believed that these are necessary for survival of T- and B-cells in the peripheral immune system [55]. In a control population of healthy individuals there may thus be significant levels of circulating auto-antibodies against some of the antigens disclosed in Table 1 and these may occur at a significant frequency in the population. The level and frequency of these biomarkers may be altered in a disease cohort, compared with the control cohort. An analysis of the level and frequency of these biomarkers in the case and control populations may identify differences which provide diagnostic information. The level of auto-antibodies directed against a specific antigen may increase or decrease in a PC sample, compared with a healthy sample.

In general, therefore, a method of the invention will involve determining whether a sample contains a biomarker level which is associated with PC. Thus a method of the invention can include a step of comparing biomarker levels in a subject's sample to levels in (i) a sample from a patient with PC and/or (ii) a sample from a patient without PC. The comparison provides a diagnostic indicator of whether the subject has PC. An aberrant level of one or more biomarker(s), as compa red to known or standard expression levels of those biomarker(s) in a sample from a patient without PC, indicates that the subject has PC.

The level of a biomarker should be significantly different from that seen in a negative control. Adva nced statistica l tools ca n be used to dete rm i ne whethe r two levels a re the same or different. For example, an in vitro diagnosis will rarely be based on comparing a single determination. Rather, an a ppropriate number of determinations will be made with an appropriate level of accuracy to give a desired statistical certainty with an acceptable sensitivity and/or specificity. Antigen a nd/or anti body levels ca n be measured q ua ntitatively to permit proper com parison, and enough determinations will be made to ensure that any difference in levels can be assigned a statistical significance to a level of p<0.05 or better. The number of determinations will vary according to various criteria (e.g. the degree of variation in the baseline, the degree of up-regulation in disease states, the degree of noise, etc. ) but, again, this falls within the normal design capabilities of a person of ordinary skill in this field. For exa mple, interqua rtile differences of norma lised data ca n be assessed, a nd the threshold for a positive signal (i.e. indicating the presence of a pa rticula r auto-antibody) can be defined as requiring that antibodies in a sample react with a diagnostic antigen at least 2.5-fold more strongly that the interquartile difference above the 75th percentile. Other criteria are familia r to those skilled in the art and, depending on the assays being used, they may be more appropriate than qua nti le norma lisation . Other methods to normalise data include data tra nsformation strategies known in the art e.g. scaling, log norma lisation, median normalisation, etc.

The underlying aim of these data interpretation techniques is to distinguish between the presence of a Table 1 biomarker a nd of a n arbitrary control biomarker, and also to distinguish between the response of sa m ple from a PC su bject from a control su bject. Methods of the invention may have sensitivity of at least 70% (e.g. >70%, >75%, >80%, >85%, >90%, >95%, >96%, >97%, >98%, >99%). Methods of the invention may have specificity of at least 70% (e.g. >70%, >75%, >80%, >85%, >90%, >95%, >96%, >97%, >98%, >99%). Advantageously, methods of the invention may have both specificity and sensitivity of at least 70% (e.g. >70%, >75%, >80%, >85%, >90%, >95%, >96%, >97%, >98%, >99%). As shown i n Ta bles 2-16, the i nve ntion ca n consistently provide specificity and sensitivity which are both above 90%. Data obtained from methods of the invention, and/or diagnostic information based on those data, may be stored in a computer medium (e.g. in RAM, in non-volatile computer memory, on CD-ROM) and/or may be transmitted between computers e.g. over the internet.

If a method of the invention indicates that a subject has prostate cancer, further steps may then follow. For instance, the subject may undergo confirmatory diagnostic procedures, such as those involving physical inspection of the subject, and/or may be treated with therapeutic agent(s) suitable for treating prostate cancer.

Monitoring the efficacy of therapy

As mentioned above, some methods of the invention involve testing samples from the same subject at two or more different points in time. In general, where the above text refers to the presence or absence of biomarker(s), the invention also includes an increasing or decreasing level of the biomarker(s) over time. An increasing level of an auto-antibody biomarker includes a spread of antibodies in which additional antibodies or antibody classes are raised against a single antigen. Methods which determine changes in biomarker(s) over time can be used, for instance, to monitor the efficacy of a therapy being administered to the subject (e.g. in theranostics). The therapy may be administered before the first sample is taken, at the same time as the first sample is taken, or after the first sample is taken.

The invention can be used to monitor a subject who is receiving PC therapy. Current therapies for PC include chemotherapy and/or hormone therapy. Hormone therapy seeks to block access of dihydrotestosterone (DHT) to prostate cells or to block the effects of DHT within prostate cells. Anti-androgens are medications such as flutamide, bicalutamide, nilutamide, and cyproterone acetate which directly block the actions of testosterone and DHT within prostate cancer cells. They may be given in combination with drugs such as ketoconazole and aminoglutethimide which block the production of adrenal androgens. In related embodiments of the invention, the results of monitoring a therapy are used for future therapy prediction. For example, if treatment with a particular therapy is effective in reducing or eliminating disease symptoms in a subject, and is also shown to decrease levels of a particular biomarker in that subject, detection of that biomarker in another subject may indicate that this other subject will respond to the same therapy. Conversely, if a particular therapy was not effective in reducing or eliminating disease symptoms in a subject who had a particular biomarker or biomarker profile, detection of that biomarker or profile in another subject may indicate that this other subject will also fail to respond to the same therapy. In other embodiments, the presence of a particular biomarker can be used as the basis of proposing or initiating a particular therapy (patient stratification). For instance, if it is known that levels of a particular auto-antibody can be reduced by administering a particular therapy then that auto-antibody's detection may suggest that the therapy should begin. Thus the 5 invention is useful in a theranostic setting.

Normally at least one sample will be taken from a subject before a therapy begins.

Immunotherapy

Where the development of auto-antibodies to a newly-exposed auto-antigen is causative for a disease, early priming of the immune response can prepare the body to remove antigenic) exposing cells when they arise, thereby removing the cause of disease before auto-antibodies develop dangerously. For example, one antigen known to be recognised by auto-antibodies is p53, and this protein is considered to be both a vaccine target and a therapeutic target for the modulation of cancer [56- 58]. The antigens listed in Tables 1 and 17 are thus therapeutic targets for treating PC.

15 Thus the invention provides a method for raising an antibody response in a subject, comprising eliciting to the subject an immunogen which elicits antibodies which recognise an antigen listed in Table 1. The method is suitable for immunoprophylaxis of prostate cancer.

The invention also provides an immunogen for use in medicine, wherein the immunogen can elicit antibodies which recognise an antigen listed in Table 1. Similarly, the invention also 20 provides the use of an immunogen in the manufacture of a medicament for immunoprophylaxis of prostate cancer, wherein the immunogen can elicit antibodies which recognise an antigen listed in Table 1.

As discussed above for detection antigens, the immunogen may be the antigen itself or may comprise an amino acid sequence having identity and/or comprising an epitope from the 25 antigen. Thus the immunogen may comprise an amino acid sequence (i) having at least 90% {e.g. >91%, >92%, >93%, >94%, >95%, >96%, >97%, >98%, >99%) sequence identity to the relevant SEQ ID NO disclosed herein, and/or (ii) comprising at least one epitope from the relevant SEQ. ID NO disclosed herein. Other immunogens may also be used, provided that they can elicit antibodies which recognise the antigen of interest.

30 As an alternative to immunising a subject with a polypeptide immunogen, it is possible to administer a nucleic acid (e.g. DNA or RNA) immunogen encoding the polypeptide, for in situ expression in the subject, thereby leading to the development of an antibody response. The immunogen may be delivered in conjunction (e.g. in admixture) with an immunological adjuvant. Such adjuvants include, but are not limited to, insoluble a luminium salts, water-in-oil emulsions, oil-in-water emulsions such as M F59 and AS03, saponins, ISCOMs, 3-O-deacylated M PL, immunostimulatory oligonucleotides (e.g. including one or more CpG motifs), bacterial ADP-ribosylating toxins and detoxified derivatives thereof, cytokines, chitosan, biodegrada ble microparticles, liposomes, imidazoquinolones, phosphazenes (e.g. PCPP), aminoalkyl glucosaminide phosphates, gamma inulins, etc. Combinations of such adjuvants can also be used. The adjuvant(s) may be selected to elicit an immune response involving CD4 or CD8 T cells. The adjuvant(s) may be selected to bias an immune response towards a THl phenotype or a TH2 phenotype.

The immunogen may be delivered by any suitable route. For example, it may be delivered by parenteral injection (e.g. subcutaneously, intraperitonea l^, intravenously, intramuscularly), or mucosally, such as by oral (e.g. tablet, spray), topical, transdermal, transcutaneous, intranasal, ocula r, a ural, pulmonary or other mucosal administration. The immunogen may be administered in a liquid or solid form. For example, the immunogen may be formulated for topical administration (e. g. as an ointment, cream or powder), for ora l administration (e.g. as a tablet or capsule, as a spray, or as a syrup), for pulmonary administration (e.g. as an inhaler, using a fine powder or a spray), as a suppository or pessary, as drops, or as an injectable solution or suspension. Imaging and staining

The antigens listed in Tables 1 and 17 can be useful for imaging. A labelled a ntibody against the a ntigen ca n be injected in vivo and the distribution of the antigen can then be detected. This method may identify the source of the antigen (e.g. an area in the body where there is a high concentration of the antigen), potentia lly offering early identification of PC. I maging techniques can also be used to monitor the progress or remission of disease, or the impact of a therapy.

The antigens listed in Tables 1 and 17 can be useful for analysing tissue samples by staining e.g. using standard immunocytochemistry. A labelled antibody against a Table 1/17 antigen can be contacted with a tissue sample to visualise the location of the antigen. A single sample could be stained with different antibodies against multiple different antigens, a nd these different antibodies may be differentia lly la belled to enable them to be distinguished. As an alternative, a plurality of different samples can each be stained with a single antibody. Thus the invention provides a labelled antibody which recognises an antigen listed in Tables 1 and 17. The antibody may be a human antibody, as discussed above. Any suitable label can be used e.g. quantum dots, spin labels, fluorescent labels, dyes, etc.

Alternative biomarkers

The invention has been described above by reference to auto-antibody and antigen biomarkers, with assays of auto-antibodies against an a ntigen being used in preference to assays of the antigen itself. In addition to these biomarkers, however, the invention can be used with other biological manifestations of the Table 1 antigens. For example, the level of mRNA transcripts encoding a Table 1 antigen can be measured, pa rticularly in tissues where that gene is not normally transcribed (such as in the potential disease tissue). Similarly, the chromosomal copy number of a gene encoding a Table 1 antigen can be measured e.g. to check for a gene duplication event. The level of a regulator of a Table 1 antigen can be measured e.g. to look at a microRNA regulator of a gene encoding the antigen. Furthermore, things which are regulated by or respond to a Table 1 antigen can be assessed e.g. if an antigen is a regulator of a metabolic pathway then disturbances in that pathway can be measured. Further possibilities will be apparent to the skilled reader.

Preferred panels

Preferred embodiments of the invention are based on a panel of biomarkers. Panels of particular interest consist of or comprise the combinations of biomarkers listed in Tables 3 to 16 (which show ten panels of 2, 3, 4, ... , 14 and 15 biomarkers).

The ten different panels listed in each of Tables 3 to 16 can be expanded by adding further biomarker(s) to create a larger panel. The further bioma rkers can usefully be selected from known biomarkers (such as PSA, PCA3, DD3, AMACR, EPCA, EPCA-2, sarcosine, etc.; see above), from Table 17, or from Table 1. In general the addition does not decrease the sensitivity or specificity of the panel shown in the Tables. Such panels include, but are not limited to:

• A panel comprising or consisting of 2 different biomarkers, namely: (i) a biomarker selected from Table 2 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 2 different biomarkers, namely: (i) a biomarker selected from Table 2 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of 3 different biomarkers, namely: (i) a group of 2 biomarkers selected from Table 3 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 3 different biomarkers, namely: (i) a group of 2 biomarkers selected from Table 3 and (ii) a further biomarker selected from Table 1. • A panel comprising or consisting of 4 different biomarkers, namely: (i) a group of 3 biomarkers selected from Table 4 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 4 different biomarkers, namely: (i) a group of 3 biomarkers selected from Table 4 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of 5 different biomarkers, namely: (i) a group of 4 biomarkers selected from Table 5 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 5 different biomarkers, namely: (i) a group of 4 biomarkers selected from Table 5 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of 6 different biomarkers, namely: (i) a group of 5 biomarkers selected from Table 6 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 6 different biomarkers, namely: (i) a group of 5 biomarkers selected from Table 6 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of 7 different biomarkers, namely: (i) a group of 6 biomarkers selected from Table 7 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 7 different biomarkers, namely: (i) a group of 6 biomarkers selected from Table 7 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of 8 different biomarkers, namely: (i) a group of 7 biomarkers selected from Table 8 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 8 different biomarkers, namely: (i) a group of 7 biomarkers selected from Table 8 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of 9 different biomarkers, namely: (i) a group of 8 biomarkers selected from Table 9 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 9 different biomarkers, namely: (i) a group of 8 biomarkers selected from Table 9 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of 10 different biomarkers, namely: (i) a group of 9 biomarkers selected from Table 10 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 10 different biomarkers, namely: (i) a group of 9 biomarkers selected from Table 10 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of 11 different biomarkers, namely: (i) a group of 10 biomarkers selected from Table 11 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 11 different biomarkers, namely: (i) a group of 10 biomarkers selected from Table 11 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of 12 different biomarkers, namely: (i) a group of 11 biomarkers selected from Table 12 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 12 different biomarkers, namely: (i) a group of 11 biomarkers selected from Table 12 and (ii) a further biomarker selected from Table 1. • A panel comprising or consisting of 13 different biomarkers, namely: (i) a group of 12 biomarkers selected from Table 13 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 13 different biomarkers, namely: (i) a group of 12 biomarkers selected from Table 13 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of 14 different biomarkers, namely: (i) a group of 13 biomarkers selected from Table 14 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 14 different biomarkers, namely: (i) a group of 13 biomarkers selected from Table 14 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of 15 different biomarkers, namely: (i) a group of 14 biomarkers selected from Table 15 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 15 different biomarkers, namely: (i) a group of 14 biomarkers selected from Table 15 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of a group of 15 different biomarkers selected from Table 16.

Preferred panels have between 2 and 15 biomarkers in total. General

The term "com prisi ng" encom passes "i ncl udi ng" as wel l as "consisting" e.g. a com position "comprising" X may consist exclusively of X or may include something additiona l e.g. X + Y.

References to a n antibody's a bi lity to "bind" a n a ntigen mea n that the antibody and a ntigen interact strongly enough to withsta nd sta nda rd washi ng procedures in the assay in question. Thus non-specific binding will be minimised or eliminated.

References to a "level" of a bioma rker mean the a mount of a n ana lyte measured in a sa m ple and this encom passes relative a nd a bsol ute concentrations of the a na lyte, analyte titres, relationships to a threshold, rankings, percentiles, etc.

An assay's "sensitivity" is the proportion of true positives which are correctly identified i.e. the proportion of PC subjects who test positive by a method of the i nvention. This ca n a pply to i ndivid ua l bioma rke rs, pa nels of bioma rkers, si ngle assays or assays which co m bi ne data integrated from multiple sources e.g. PSA score and DRE. It can relate to the ability of a method to identify samples containing a specific analyte (e.g. antibodies) or to the ability of a method to correctly identify samples from subjects with PC.

An assay's "specificity" is the proportion of true negatives which are correctly identified i.e. the proportion of subjects without PC who test negative by a method of the invention. This can apply to individual biomarkers, panels of bioma rkers, single assays or assays which combine data integrated from multiple sources e.g. PSA score and DRE. It can relate to the ability of a method to identify samples containing a specific analyte (e.g. antibodies) or to the ability of a method to correctly identify samples from subjects with PC.

Unless specifically stated, a method comprising a step of mixing two or more components does not require any specific order of mixing. Thus components can be mixed in any order. Where there are three components then two components can be combined with each other, and then the combination may be com bined with the third component, etc.

References to a percentage sequence identity between two amino acid sequences means that, when aligned, that percentage of amino acids are the same in comparing the two sequences. This alignment and the percent homology or sequence identity can be determined using software programs known in the art, for example those described in section 7.7.18 of ref. 59. A preferred alignment is determined by the Smith-Waterma n homology sea rch a lgorithm using a n affine gap sea rch with a ga p open pena lty of 12 a nd a gap extension pena lty of 2, BLOSU M matrix of 62. The Smith-Waterman homology search algorithm is disclosed in ref. 60.

BRIEF DESCRIPTION OF DRAWINGS

Figure 1 shows the combined sensitivity plus specificity (S+S) score for single markers (Figure 1A) or panels of markers (Figures IB to 10), ordered from highest S+S score to lowest. From A to 0 the nu m ber of ma rke rs increases by one per pa nel. The 10 ma rkers or pa nels with the highest S+S score are at the left of the graphs and are listed in Tables 2 to 16. The y-axis shows S+S score and the x-axis shows the num ber of pa nels (xlO⁵) tested except for Figure 1A where the x-axis refers to individua l proteins.

Figure 2 shows the development of the S+S score (y-axis) as pa nel size (x-axis) increases. Each analysis cycle corresponds to the development of an n-mer panel where n=l, 2, 3...15.

Figure 3 shows RFU distribution plots for three bioma rkers (ABCF3, DOM 3Z and YARS). The rfu distribution for three biomarkers (3A: ABCF3; 3B: DOM3Z; 3C: YARS) is plotted in one shade for healthy samples and a different shade for diseased samples. Stronger biomarkers have little overlap in the profiles.

MODES FOR CARRYING OUT THE INVENTION Array preparation

Three separate protein arrays were developed. Full-length open reading frames for target genes encoding 925 proteins covering a wide range of protein classes were cloned in-frame with a sequence encoding a C-terminal E. coli BCCP-myc tag [30, 40] in a baculovirus transfer vector and sequence-verified. Recombinant baculoviruses were generated, am plified and expressed in Sf9 cells using standard methods adapted for 24-well deep well plates. Recombina nt protein expression was analyzed for protein integrity and biotinylation by Western blotting. Cells harbouring recombinant protein were lysed and lysates were spotted in q uad ru p licate usi ng a Q.Array2 M icroa rraye r eq ui p ped with 300 μιη so l id pi ns on to streptavidin-coated glass slides. Spotted proteins project into an aqueous environment a nd orient away from the surface of the slide, exposing them for binding by auto-antibodies.

I n addition to the proteins on each array, four control proteins for the BCCP-myc tag (BCCP, BCCP-myc, β-galactosidase-BCCP-myc and β-galactosidase-BCCP) were a rrayed, a long with Cy3/Cy5-la beled biotin-BSA, dilution series of biotinylated-lgG and biotinylated IgM, a biotinylated-myc peptide dilution series and buffer-only spots.

Biomarker confirmation

Serum samples were obtained from two groups of subjects:

1. "disease": serum samples from PC-diagnosed subjects (n=73).

2. "healthy and confounding disease": serum samples from age-matched healthy donors (n=37) and serum samples from individuals diagnosed with BPH (n=23).

Serum samples from both groups were individually analysed using each of the three types of arrays. Serum samples were incubated with each of the three array types separately. Serum samples were clarified by centrifugation at 10-13K rpm for 2 minutes at 4°C to re move particulates, including lipids. The samples were then diluted 200-fold in 0.1% v/v Triton/0.1% v/v BSA in IX PBS (Triton-BSA buffer) and then applied to the arrays. Diluted serum (4 mL) sample was added to each array housed in a separate compartment of a plastic dish. All arrays were incubated for 2 hours at room temperature (RT, 20°C) with gentle orbital shaking (~50 rpm). Arrays were removed ca reful ly from the dish a nd any excess probing solution was removed by blotting the sides of the array onto lint-free tissue. Probed arrays were washed three times in fresh Triton-BSA buffer at RT for 20 minutes with gentle orbital shaking. The washed slides were then blotted onto lint-free tissue to remove excess wash buffer and were incubated in a secondary staining solution (prepared just prior to use) at RT for 2 hours, with gentle orbital shaking and protected from light using aluminium foil. The secondary staining solution was a labelled anti-human IgG antibody. Slides were washed three times in Triton-BSA buffer for 5 minutes at RT with gentle orbital shaking, rinsed briefly (5-10 seconds) in distilled water, and centrifuged for 2 minutes at 240g in a container suitable for centrifugation. To help wick away excess liquid on the arrays, a lint-free tissue was placed at the bottom of the arrays during centrifugation.

The probed and dried arrays were then scanned using a microarray scanner capable of using an excitation wavelength suitable for the detection of the secondary staining solution, to detect auto-antibodies bound by the array and to determine magnitude of auto-antibody binding. The microarray scans produced images for each array that were used to determine the intensity of fluorescence bound to each protein spot which were used to normalize and score array data.

Raw median signal intensity (also referred to as the relative fluorescent unit, RFU) of each protein feature (also referred to as a spot or antigen) on the array was subtracted from the local median background intensity. Alternative analyses use other measures of spot intensity such as the mean fluorescence, total fluorescence, as known in the art.

The resulting net fluorescent intensities of all protein features on each array were then normalized to reduce the influence of technical bias (e.g. laser power variation, surface variation, binding to BCCP, etc.) by a multiscaling procedure. Other methods for data normalization suitable for the data include, amongst others, quantile normalization [47], multiplication of net fluorescent intensities by a normalisation factor consisting of the product of the 1st quartile of all intensities of a sample and the mean of the 1st quartiles of all samples and the "VSN" method [61]. Such normalization methods are known in the art of microarray analysis. The normalized fluorescent intensities were then averaged for each protein feature.

The multiscaling method was applied to all 3700 quadruplicate signals from 133 protein arrays. Data were arbitrarily split in test and training sets and the data from the training set was then used with GP to identify classifiers which would successfully distinguish case from control samples. Classifiers were then assessed for performance by referring to the combined sensitivity and specificity (S+S score) using the test set. Data were repeatedly split into test and training sets and analysis cycles repeated until a stable set of classifiers was identified. The number of biomarkers in each panel was limited to n where n = 1-15. The performance of the derived panels was then ranked by combined S+S. The top 6000 panels for each n-mer panel were taken and the frequency of appearance of each protein in these panels was used to rank the predictive power of each protein for that specific n-mer. The top 10 markers for each n- mer, as judged by frequency of appearance were then combined into a single list and ranked by overall number of appearances. For each n-mer the 10 panels which provide the highest combined sensitivity and specificity score (S+S) are presented in Tables 2-16. The biomarkers frequently appearing in the top 10 panels for all the presented n-mers were combined to produce the set of 24 markers in Table 1. The top panels in Tables 2-16 each have a S+S score higher than the value of 1.2 (i.e. above the typical value for PSA assays [1]).

Overall, Tables 2-16 produced the set of 105 biomarkers presented in Table 17, a subset of 24 of which are presented in Table 1. Each of these 24 biomarkers has significant predictive power across multiple n-mers. For example, as a single marker, PITRM1 is ranked number 2 overall with a combined S+S of 1.319 (Table 2), but when n>2 PITRM1 does not appear in the top 10 panels. In contrast, ABCF3 increasingly dominates where n>4 but appears less frequently where the panels contain <4 markers. Similarly, SDCCAG10 appears in just one of the panels where n=2 but is also a dominant biomarker where n>4. Thus the contribution that a particular biomarker provides to the discriminatory power of a panel can depend on the number of markers in that panel as well as on their identity.

The presence of antibodies to the Table 1 antigens was confirmed to be significantly different between the two groups. A back propagation algorithm was used to confirm biomarkers that can distinguish between the two groups. The data analysis was validated by two permutation assays. These assays confirmed that the chosen biomarkers are related to the disease status of the sera. The core biomarker set was successfully validated by depleting the set of 925 proteins of the 24 identified bioma rkers a nd repeati ng the a na lysis. With the data from these biomarkers removed, it was no longer possible to derive a panel which could distinguish between healthy and diseased serum samples with comparable performance.

To confirm that the biomarkers can be used to distinguish between PC and non-PC data in general, rather than being limited to a particular algorithm, the panels were also tested using discriminant analysis in Matlab. Two separate algorithms (Matlab's "classify" and "SVM") were used with a panel and results were as follows:

Thus both classification methods performed with remarkably similar sensitivities and specificities, confirming that the panel possessed an inherent ability to discriminate between case and controls and was not reliant on a single method of analysis.

It wi ll be understood that the invention has been described by way of exa m ple only a nd modifications may be made whilst remaining within the scope and spirit of the invention. TABLE 1: Biomarkers useful with the invention

Table 1 lists biomarkers useful with the invention. The measured biomarker can be (i) presence of auto-antibody which binds to an antigen listed in Table 1 and/or (ii) the presence of an antigen listed in Table 1, but is preferably the former.

Columns

(i) The "Symbol" column gives the gene symbol which has been approved by the HGNC. The symbol thus identifies a unique human gene. This symbol can be related via Table 17 to the gene's Official Full Name provided by NCBI.

(ii) This number is the SEQ ID NO: for the coding sequence for the auto-antigen biomarker, as shown in Table 17.

(iii) The HUGO Gene Nomenclature Committee aims to give unique and meaningful names to every human gene. The HGNC number thus identifies a unique human gene. TABLE 2

Columns (Tables 2 to 16)

(i) This is the symbol for the relevant biomarker (or, for Tables 3-16, biomarkers in the panel).

(ii) S+S is the sum of the sensitivity and specificity columns. These final two columns show the sensitivity and specificity of a test based solely on the relevant biomarker (or, for Tables 3-16, panel) shown in the left-hand column when applied to the samples used in the examples.

TABLE 3

TABLE 4

TLK1 SDCCAG10 NDUFAB1 1.471 0.686 0.785

PRKCBP1 D0M3Z SDCCAG10 1.47 0.719 0.751

VEGFB ABCF3 TPI1 1.469 0.737 0.732

GOLGA5 SDCCAG10 PDCD6 1.469 0.786 0.683

TLK1 SDCCAG10 H2AFY 1.469 0.751 0.719

MAPKAPK5 ABCF3 DDX55 1.469 0.811 0.657

GOLGA5 SDCCAG10 AARS 1.469 0.769 0.7

TABLE 5

TABLE 6

TABLE 7

Panel S+S Sensitivity Specificity

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 PRC1 1.75 0.888 0.862

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 RH0T2 1.75 0.9 0.85

SDCCAG10 ABCF3 D0M3Z CTAG2 CSNK2A1 CABC1 1.749 0.903 0.847

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 CTAG2 1.748 0.92 0.828 SDCCAG10 ABCF3 D0M3Z MAPKAPK5 CTAG2 CAMK2A 1.746 0.942 0.804

SDCCAG10 ABCF3 D0M3Z CTAG2 CSNK2A1 DDIT3 1.744 0.898 0.846

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 CTAG2 PACE-1 1.743 0.9 0.843

SDCCAG10 ABCF3 D0M3Z CTAG2 CSNK2A1 TLK1 1.743 0.863 0.88

SDCCAG10 ABCF3 D0M3Z CTAG2 CSNK2A1 ETS2 1.742 0.896 0.846

SDCCAG10 ABCF3 D0M3Z CTAG2 CSNK2A1 NR4A1 1.74 0.897 0.843

TABLE 8

TABLE 9

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 CTAG2 1.83 0.947 0.883 MADH5 IKBKB

SDCCAG10 ABCF3 DOM3Z MAPKAPK5 GOLGA5 CTAG2 1.83 0.939 0.89 MADH5 STK11

SDCCAG10 ABCF3 DOM3Z MAPKAPK5 GOLGA5 CTAG2 1.829 0.945 0.884 MADH5 CABC1

SDCCAG10 ABCF3 DOM3Z MAPKAPK5 GOLGA5 CTAG2 1.827 0.937 0.89 MADH5 PSKH1

SDCCAG10 ABCF3 DOM3Z MAPKAPK5 GOLGA5 CTAG2 1.825 0.93 0.895 MADH5 DYRK2

SDCCAG10 ABCF3 DOM3Z MAPKAPK5 GOLGA5 CTAG2 PDK4 1.825 0.941 0.883 ETS2

SDCCAG10 ABCF3 DOM3Z MAPKAPK5 GOLGA5 CTAG2 1.824 0.936 0.888 MADH5 TYR03

TABLE 10

TABLE 11

TABLE 12

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 CTAG2 1.868 0.955 0.913 MADH5 SAV1 SRPK1 STK25 PCTK1

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 CTAG2 1.868 0.953 0.915 MADH5 CABC1 PSME3 PPM1A PIK3R1

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 CTAG2 1.868 0.961 0.907 MADH5 PRKAG3 PACE-1 RPL30 IGHG1

TABLE 13

Panel s+s Sensitivity Specificity

SDCCAG10 ABCF3 DOM3Z MAPKAPK5 GOLGA5 CTAG2 1.882 0.955 0.927 MADH5 CABC1 PSME3 SGKL PXK ASNA1

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 CTAG2 1.881 0.952 0.929 MADH5 CABC1 PSME3 SGKL PXK ZMAT2

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 CTAG2 1.881 0.955 0.927 MADH5 CABC1 PSME3 SGKL PXK APEG1

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 CTAG2 1.881 0.954 0.927 MADH5 CABC1 PSME3 PPM1A PXK RFK

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 CTAG2 1.881 0.953 0.928 MADH5 CABC1 PSME3 SGKL PXK RPL30

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 CTAG2 1.881 0.954 0.927 MADH5 CABC1 PSME3 SGKL PXK GSTT1

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 CTAG2 1.881 0.954 0.927 MADH5 CABC1 PSME3 SGKL PXK YARS

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 CTAG2 1.881 0.955 0.925 MADH5 CABC1 PSME3 SGKL PXK VIM

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 CTAG2 1.881 0.955 0.926 MADH5 CABC1 PSME3 SGKL PXK STK25

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 CTAG2 1.88 0.953 0.927 MADH5 CABC1 PSME3 PPM1A PXK TBC1 D2

TABLE 14

Panel S+S Sensitivity Specificity

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 CTAG2 1.886 0.955 0.931 MADH5 CABC1 PSME3 SGKL PXK ASNA1 FUS

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 CTAG2 1.885 0.954 0.931 MADH5 CABC1 PSME3 SGKL PXK FUS RFK

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 CTAG2 1.885 0.952 0.933 MADH5 CABC1 PSME3 SGKL PXK YARS FUS SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 CTAG2 1.885 0.953 0.932 MADH5 CABC1 PSME3 SGKL PXK YARS PPP4R1

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 CTAG2 1.885 0.956 0.93 MADH5 CABC1 PSME3 SGKL PXK PDK1 STK11

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 CTAG2 1.885 0.952 0.933 MADH5 CABC1 PSME3 SGKL PXK GSTT1 PDK1

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 CTAG2 1.885 0.955 0.93 MADH5 CABC1 PSME3 SGKL PXK RPL30 MAZ

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 CTAG2 1.885 0.954 0.931 MADH5 CABC1 PSME3 SGKL PXK RPL30 RFK

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 CTAG2 1.884 0.954 0.93 MADH5 CABC1 PSME3 SGKL PXK VIM FUS

SDCCAG10 ABCF3 D0M3Z MAPKAPK5 G0LGA5 CTAG2 1.884 0.954 0.93 MADH5 CABC1 PSME3 SGKL PXK ZMAT2 PAK4

TABLE 15

TABLE 16

TABLE 17

No:⁽ⁱ⁾ Symbol Name _G| (iv) ID ^,V)

1 AARS alanyl-t NA synthetase 15079237 16

2 ABCF3 ATP-binding cassette sub-family F (GCN20) member 3 38197099 55324

3 APEG1 aortic preferentially expressed protein 1 33873504 10290

4 ASNA1 arsA arsenite transporter ATP-binding homolog 1 38114734 439

(bacterial)

5 BMX BMX non-receptor tyrosine kinase 34189854 660

6 BTRC beta-transducin repeat containing transcript variant 1 20380815 8945

7 C20orf97 chromosome 20 open reading frame 97 20071610 57761

8 CABC1 chaperone ABC1 activity of bcl complex like (S. pombe) 33873178 56997

9 CAMK2A calcium/calmodulin-dependent protein kinase II alpha 26251711 815

10 CCM2 chromosome 7 open reading frame 22 33870299 83605

11 CCRK cell cycle related kinase 33988018 23552

12 CDKN1A cyclin-dependent kinase inhibitor 1A (p21 Cipl) 12653024 1026 transcript variant 1

CDKN2D cyclin-dependent kinase inhibitor 2D (pl9 inhibits 38114834 1032

CDK4) transcript varian

CHEK2 CHK2 checkpoint homolog (S. pombe) transcript variant 38114706 11200

1

CSNK1 G2 casein kinase 1, gamma 2 33870264 1455

CSNK2A1 casein kinase 2 alpha 1 polypeptide transcript variant 2 33991298 1457

CSTB cystatin B (stefin B) 13097209 1476

CTAG2 cancer/testis antigen 2 transcript variant 2 38114801 30848

DBNL drebrin-like 21619482 28988

DCLK2 hypothetical protein MGC45428 21619201 166614

DDIT3 DNA-damage-inducible transcript 3 33872688 1649

DDX55 DEAD (Asp-Glu-Ala-Asp) box polypeptide 55 34190861 57696

DNM1L dynamin 1-like transcript variant 2 19352980 10059

D0M3Z dom-3 homolog Z (C. elegans) 33878616 1797

DYRK2 dual-specificity tyrosine-(Y)-phosphorylation regulated 33871530 8445 kinase 2

ETS2 v-ets erythroblastosis virus E26 oncogene homolog 2 16877577 2114

(avian)

FEN1 flap structure-specific endonuclease 1 33875300 2237

FGFR2 Homo sapiens Homo sapiens fibroblast growth factor 25058744 2263 receptor 2 (bacteria-expressed kinase keratinocyte

FLJ10377 hypothetical protein FU10377 33988197 55131

FUS fusion (involved in t(1216) in malignant liposarcoma) 33875401 2521

G0LGA5 golgi autoantigen golgin subfamily a 5 18606387 9950

GRK5 G protein-coupled receptor kinase 5 mRNA (cDNA clone 40352898 2869

MGC:71228 )

GSK3B glycogen synthase kinase 3 beta 12652980 2932

GSTT1 glutathione S-transferase theta 1 13937910 2952

H11 protein kinase Hll 33877008 26353

H2AFY H2A histone family member Y 15426457 9555

HRB2 HIV-1 rev binding protein 2 34783224 11103

IGHG1 immunoglobulin heavy constant gamma 1 (Glm 15779221 3500 marker)

IKBKB inhibitor of kappa light polypeptide gene enhancer in B- 33873496 3551 cells kinase beta

LIMS1 LIM and senescent cell antigen-like domains 1 13529136 3987

MADH5 MAD mothers against decapentaplegic homolog 5 34189276 4090

(Drosophila)

MAP2K7 mitogen-activated protein kinase kinase 7 34192881 5609

MAPK13 mitogen-activated protein kinase 13 37589022 5603

MAPK6 mitogen-activated protein kinase 6 34782811 5597

MAPK9 mitogen-activated protein kinase 9 transcript variant 1 21618469 5601

MAPKAPK5 mitogen-activated protein kinase-activated protein 28704100 8550 kinase 5 transcript variant 1 mRNA (cDNA clone

MGC:54058 )

MAZ MYC-associated zinc finger protein (purine-binding 27371183 4150 transcription factor)

METTL3 methyltransferase like 3 33876376 56339

MRPL55 mitochondrial ribosomal protein L55 31127212 128308

MST4 Mst3 and SOKl-related kinase (MASK) 109633024 51765 MY0Z2 myozenin 2 13528788 51778

NDUFAB1 NADH dehydrogenase (ubiquinone) 1 alpha/beta 37748351 4706 subcomplex 1 8kDa

NR4A1 nuclear receptor subfamily 4 group A member 1 16359382 3164 transcript variant 1

NTRK2 neurotrophic tyrosine kinase receptor type 2 21594336 4915

PACE-1 ezrin-binding partner PACE-1 transcript variant 1 15779206 57147

PAK4 p21(CDKNlA)-activated kinase 4 33877350 10298

PCTK1 PCTAIRE protein kinase 1 transcript variant 2 33875920 5127

PCTK2 PCTAIRE protein kinase 2 21542570 5128

PDCD6 programmed cell death 6 15214523 10016

PDK1 pyruvate dehydrogenase kinase isoenzyme 1 24660127 5163

PDK2 pyruvate dehydrogenase kinase isoenzyme 2 38197017 5164

PDK4 pyruvate dehydrogenase kinase isoenzyme 4 25955470 5166

PIK3R1 phosphoinositide-3-kinase regulatory subunit 21410089 5295 polypeptide 1 (p85 alpha) tr

PITRM1 pitrilysin metalloproteinase 1 12654626 10531

PPARG peroxisome proliferative activated receptor gamma 13905055 5468 transcript variant 3

PPM1A protein phosphatase 1A (formerly 2C) magnesium- 20070651 5494 dependent alpha isoform tr

PPP2R5C protein phosphatase 2 regulatory subunit B (B56) 16740598 5527 gamma isoform transcript

PPP4R1 protein phosphatase 4 regulatory subunit 1 38174527 9989

PRC1 protein regulator of cytokinesis 1 13111934 9055

PRKAG3 protein kinase AMP-activated gamma 3 non-catalytic 47132576 53632 subunit (PRKAG3)

PRKAR1A protein kinase cAMP-dependent regulatory type 1 alpha 23273779 5573

(tissue specific e

PRKCBP1 protein kinase C binding protein 1 21315038 23613

PRKRA protein kinase interferon-inducible double stranded 14495716 8575

RNA dependent activator

PSKH1 protein serine kinase HI 38511461 5681

PSME3 proteasome (prosome macropain) activator subunit 3 33876201 10197

(PA28 gamma Ki) transc

PTK9L PTK9L protein tyrosine kinase 9-like (A6-related protein) 16741224 11344

PTPN11 protein tyrosine phosphatase non-receptor type 11 14250500 5781

(Noonan syndrome 1)

PXK PX domain containing serine/threonine kinase 15680248 54899

RFK riboflavin kinase 13937919 55312

RH0T2 ras homolog gene family member T2 15928946 89941

RIPK1 receptor (TNFRSF)-interacting serine-threonine kinase 1 57242760 8737

(RIPK1)

RNF6 ring finger protein (C3H2C3 type) 6 transcript variant 1 34193514 6049

RPL30 ribosomal protein L30 34783378 6156

RPS6KL1 ribosomal protein S6 kinase-like 1 33873209 83694

SAV1 Salvador homolog 1 (Drosophila) 18088227 60485

SDCCAG10 serologically defined colon cancer antigen 10 15082404 10283

SGKL serum/glucocorticoid regulated kinase-like transcript 15929809 23678 variant 1

SRPK1 SFRS protein kinase 1 23468344 6732 89 SRPK2 SF S protein kinase 2 transcript variant 2 23270875 6733

90 STIP1 stress-induced-phosphoprotein 1 (Hsp70/Hsp90- 12804256 10963 organizing protein)

91 STK11 serine/threonine kinase 11 (Peutz-Jeghers syndrome) 33872385 6794

92 STK25 serine/threonine kinase 25 (STE20 homolog yeast) 33873686 10494

93 STK32A Homo sapiens hypothetical protein MGC22688 18203872 202374

94 STK33 serine/threonine kinase 33 22658391 65975

95 TACC1 transforming acidic coiled-coil containing protein 1 27552854 6867

96 TBC1 D2 TBC1 domain family, member 2, 20810372 55357

97 TLK1 tousled-like kinase 1 21618509 9874

98 TPI1 triosephosphate isomerase 1 13937949 7167

99 TRB2 tribbles homolog 2 33990940 28951

100 TYR03 Homo sapiens TYR03 protein tyrosine kinase 29477257 7301

101 VEGFB vascular endothelial growth factor B 38197016 7423

102 VIM vimentin 13111800 7431

103 YARS tyrosyl-tRNA synthetase 37588917 8565

104 ZHX2 zinc fingers and homeoboxes 2 27503824 22882

105 ZMAT2 zinc finger matrin type 2 34785080 153527

Columns (Tables 17 & 18)

(i) This number is the SEQ ID NO: for the coding sequence for the auto-antigen biomarker, as shown in the sequence listing.

(ii) The "Symbol" column is as described for Table 1.

(iii) This name is taken from the Official Full Name provided by NCBI. An antigen may have been referred to by one or more pseudonyms in the prior art. The invention relates to these antigens regardless of their nomenclature.

(iv) A "Gl" number, "Genlnfo Identifier", is a series of digits assigned consecutively to each sequence record processed by NCBI when sequences are added to its databases. The Gl number bears no resemblance to the accession number of the sequence record. When a sequence is updated (e.g. for correction, or to add more annotation or information) it receives a new Gl number. Thus the sequence associated with a given Gl number is never changed.

(v) The "ID" column shows the Entrez GenelD number for the antigen marker. An Entrez GenelD value is unique across all taxa.

TABLE 18

_G| (iv)

Symbol Name ID ^,V)

ACAT2 acetyl-Coenzyme A acetyltransferase 2 (acetoacetyl 38197144 39

Coenzyme A thiolase),

ADSL adenylosuccinate lyase, 12652984 158

AK3 adenylate kinase 3, 15489347 50808

AK3L1 adenylate kinase 3-like 1, transcript variant 3, 16740594 205

AK7 adenylate kinase 7, 23272320 122481

AKT1 v-akt murine thymoma viral oncogene homolog 1, 33875493 207

ALPK1 alpha-kinase 1, mRNA (cDNA clone MGC:71554 ) 38174241 80216

ANXA1 annexin Al, 12654862 301 A IH2 ariadne homolog 2 (Drosophila), 33875424 10425

ASNA1 arsA arsenite transporter, ATP-binding, homolog 1 38114734 439

(bacterial),

ASPSCR1 alveolar soft part sarcoma chromosome region, 17511731 79058 candidate 1,

BAG3 BCL2-associated athanogene 3, 13623600 9531

BCL10 B-cell CLL/lymphoma 10 31565460 8915

BCL2A1 BCL2-related protein Al, 16740835 597

BIRC2 baculoviral IAP repeat-containing 2, 22382083 329

BRD2 bromodomain containing 2, mRNA (cDNA clone 39645316 6046

MGC:74927 )

BRD3 bromodomain containing 3, 33878091 8019

Clorf57 chromosome 1 open reading frame 57, 13477260 84284

C3IP1 kelch-like protein C3IP1, 13112018 59349

CA1 carbonic anhydrase 1, 20380765 759

CALM1 calmodulin 1 (phosphorylase kinase, delta), 33869376 801

CALM2 calmodulin 2 (phosphorylase kinase, delta), 13097164 805

CALM3 calmodulin 3 (phosphorylase kinase, delta), 13544109 808

CALU calumenin, 33870609 813

CAMK1 calcium/calmodulin-dependent protein kinase 1 21536281 8536

(CAMK1), OriGene unique variant 1

CAMK2G calcium/calmodulin-dependent protein kinase (CaM 21707841 818 kinase) II gamma, transcrip

CAMKK2 calcium/calmodulin-dependent protein kinase 33991300 10645 kinase 2, beta, transcript varia

CAMKV hypothetical protein MGC8407, 33875513 79012

CARKL carbohydrate kinase-like, 18088235 23729

CASP3 caspase 3, apoptosis-related cysteine protease, 34190795 836 transcript variant alpha,

CBX5 chromobox homolog 5 (HP1 alpha homolog, 13905073 23468

Drosophila),

CDC2 cell division cycle 2, Gl to S and G2 to M, transcript 15778966 983 variant 1,

CDC20 CDC20 cell division cycle 20 homolog (S. cerevisiae), 33875656 991

CDC25B cell division cycle 25B, transcript variant 3, 33991200 994

CDC25C cell division cycle 25C, transcript variant 1, 33877967 995

CDC42 cell division cycle 42 (GTP binding protein, 25kDa), 33990903 998 transcript variant 1,

CDK3 cDNA clone MGC:54300 complete cds 28839544 1018

CDKN2B cyclin-dependent kinase inhibitor 2B (pl5, inhibits 15680230 1030

CDK4), transcript varian

CDKN2C cyclin-dependent kinase inhibitor 2C (pl8, inhibits 18921420 1031

CDK4), transcript varian

CDKN2D cyclin-dependent kinase inhibitor 2D (pl9, inhibits 38114834 1032

CDK4), transcript varian

CHN1 chimerin (chimaerin) 1, 15030253 1123

CKB creatine kinase, brain, 12654700 1152

CLK2 CDC-like kinase 2, transcript variant phclk2, 33873844 1196

CNN1 calponin 1, basic, smooth muscle, 34190276 1264

COL4A3BP Similar to collagen, type IV, alpha 3 (Goodpasture 33990709 10087 antigen) binding protein, clone MGC:1410 CREB1 cAMP responsive element binding protein 1, 14714955 1385 transcript variant B,

CSK c-src tyrosine kinase (CSK) 187475371 1445

CSNK1D casein kinase 1, delta, transcript variant 1, 13097701 1453

CSNK1G2 casein kinase 1, gamma 2, 33870264 1455

CSNK2A1 casein kinase 2, alpha 1 polypeptide, transcript 33991298 1457 variant 2,

CSNK2A2 casein kinase 2, alpha prime polypeptide, 38197019 1459

CSTB cystatin B (stefin B), 13097209 1476

DDB1 damage-specific DNA binding protein 1, 127kDa, 33874415 1642

DDIT3 DNA-damage-inducible transcript 3, 33872688 1649

DDR1 discoidin domain receptor family, member 1, 33870104 780 transcript variant 2, mRNA (cDNA clone MGC:3909 )

DNAJB1 DnaJ (Hsp40) homolog, subfamily B, member 1, 38197192 3337

DNCLI2 dynein, cytoplasmic, light intermediate polypeptide 19684162 1783

2,

D0M3Z dom-3 homolog Z (C. elegans), 33878616 1797

DR1 down-regulator of transcription 1, TBP-binding 38197217 1810

(negative cofactor 2),

DUSP12 dual specificity phosphatase 12, 13623373 11266

DYRK2 dual-specificity tyrosine-(Y)-phosphorylation 33871530 8445 regulated kinase 2,

E1B-AP5 ElB-55kDa-associated protein 5, 33987968 11100

EGR2 early growth response 2 (Krox-20 homolog, 23272557 1959

Drosophila),

ELK3 ELK3, ETS-domain protein (SRF accessory protein 2), 16924203 2004

EN01 enolase 1, (alpha), 33876448 2023

EN02 enolase 2, (gamma, neuronal), 33877116 2026

ERBB3 v-erb-b2 erythroblastic leukemia viral oncogene 52789418 2065 homolog 3 (avian), mRNA (cDNA clone MGC:88033 )

ESR2 estrogen receptor 2 (ER beta), 34193698 2100

ETS2 v-ets erythroblastosis virus E26 oncogene homolog 16877577 2114

2 (avian),

EZH2 enhancer of zeste homolog 2 (Drosophila), 34194096 2146 transcript variant 1,

FADD Fas (TNFRSF6)-associated via death domain, 33875320 8772

FEN1 flap structure-specific endonuclease 1, 33875300 2237

FGF1 fibroblast growth factor 1 (acidic), 21595686 2246

FGF13 fibroblast growth factor 13, transcript variant 1A, 15706433 2258

FGFR1 fibroblast growth factor receptor 1 (fms-related 22450877 2260 tyrosine kinase 2, Pfeiffer syndrome), transcript

variant 2,

FIP1L1 FIP1 like 1 (S. cerevisiae), 17389362 81608

FU12577 hypothetical protein FU12577, 33877556 81617

FOLH1 folate hydrolase (prostate-specific membrane 19343603 2346 antigen) 1,

FUS fusion (involved in t(12;16) in malignant 33875401 2521 liposarcoma),

GALK1 galactokinase 1, 12654656 2584

GCLC glutamate-cysteine ligase, catalytic subunit, 25058512 2729

GEM GTP binding protein overexpressed in skeletal 34193982 2669 muscle, transcript variant 2,

GMPS guanine monphosphate synthetase, 15082534 8833

GNAZ guanine nucleotide binding protein (G protein), 22382164 2781 alpha z polypeptide,

GNB1 guanine nucleotide binding protein (G protein), 33880237 2782 beta polypeptide 1,

GNG4 guanine nucleotide binding protein (G protein), 18490900 2786 gamma 4,

GNGT2 guanine nucleotide binding protein (G protein), 14250451 2793 gamma transducing activity p

G0T1 glutamic-oxaloacetic transaminase 1, soluble 38197170 2805

(aspartate aminotransferase 1),

G B10 growth factor receptor-bound protein 10, 18999461 2887

GRB2 growth factor receptor-bound protein 2, 33875666 2885

GSK3B glycogen synthase kinase 3 beta, 12652980 2932

GUK1 guanylate kinase 1, 33871471 2987

HOXB6 homeo box B6, transcript variant 2, 15779174 3216

HP T1 hypoxanthine phosphoribosyltransferase 1 (Lesch- 34784789 3251

Nyhan syndrome),

HSPA1A heat shock 70kDa protein 1A, 33876702 3303

HSPCA heat shock 90kDa protein 1, alpha, 18605740 3320

HSPE1 heat shock lOkDa protein 1 (chaperonin 10), 33871754 3336

ID1 inhibitor of DNA binding 1, dominant negative helix- 33875639 3397 loop-helix protein, tran

IGHG1 immunoglobulin heavy constant gamma 1 (Glm 15779221 3500 marker),

IHPK1 inositol hexaphosphate kinase 1, 15277916 9807

IL18 interleukin 18 (interferon-gamma-inducing factor), 13937810 3606

IMPDH1 IMP (inosine monophosphate) dehydrogenase 1, 21706906 3614

IRF5 interferon regulatory factor 5, transcript variant 2, 34782796 3663

ISG20 interferon stimulated gene 20kDa, 33871974 3669

KRT14 keratin 14 (epidermolysis bullosa simplex, Dowling- 38114838 3861

Meara, Koebner),

LDHB lactate dehydrogenase B, 12803116 3945

LIMK2 LIM domain kinase 2, 15341773 3985

LIN28 lin-28 homolog (C. elegans), 33872076 79727

MADH2 MAD, mothers against decapentaplegic homolog 2 15928761 4087

(Drosophila),

MADH5 MAD, mothers against decapentaplegic homolog 5 34189276 4090

(Drosophila),

MAP2K6 mitogen-activated protein kinase kinase 6, 15080539 5608 transcript variant 1,

MAP3K2 mitogen-activated protein kinase kinase kinase 2 85838510 10746

(MAP3K2)

MAP3K7 mitogen-activated protein kinase kinase kinase 7, 34189719 6885 transcript variant A,

MAP4K5 mitogen-activated protein kinase kinase kinase 23273902 11183 kinase 5,

MAPK1 mitogen-activated protein kinase 1, transcript 17389605 5594 variant 2,

MAPK11 mitogen-activated protein kinase 11, 20379774 5600 MAPK3 mitogen-activated protein kinase 3, 15559270 5595

MAPK8 mitogen-activated protein kinase 8 (MAPK8), 20986493 5599 transcript variant 2

MAPK9 mitogen-activated protein kinase 9, transcript 21618469 5601 variant 1,

MAPKAPK3 mitogen-activated protein kinase-activated protein 33876390 7867 kinase 3,

MARIO MAP/microtubule affinity-regulating kinase 3, 19353235 4140

MAZ MYC-associated zinc finger protein (purine-binding 27371183 4150 transcription factor),

MCM5 MCM5 minichromosome maintenance deficient 5, 12652780 4174 cell division cycle 46 (S. cere

MIF macrophage migration inhibitory factor 33875452 4282

(glycosylation-inhibiting factor),

MKNK1 MAP kinase-interacting serine/threonine kinase 1, 33877125 8569

MLH1 mutL homolog 1, colon cancer, nonpolyposis type 2 13905125 4292

(E. coli),

MMP2 matrix metalloproteinase 2 (gelatinase A, 72kDa 33876889 4313 gelatinase, 72kDa type IV co

MPP1 membrane protein, palmitoylated 1, 55kDa, 38197472 4354

MSC musculin (activated B-cell factor-1), 33991179 9242

MST4 Mst3 and SOKl-related kinase (MASK) 109633024 51765

MTCP1 mature T-cell proliferation 1, 12803540 4515

MX1 myxovirus (influenza virus) resistance 1, interferon- 21619146 4599 inducible protein p78 (

MYD88 myeloid differentiation primary response gene (88), 15488922 4615

NME5 non-metastatic cells 5, protein expressed in 34190528 8382

(nucleoside-diphosphate kinase),

NR3C1 nuclear receptor subfamily 3, group C, member 1 33874523 2908

(glucocorticoid receptor),

NR4A1 nuclear receptor subfamily 4, group A, member 1, 16359382 3164 transcript variant 1,

NUDT2 nudix (nucleoside diphosphate linked moiety X)- 34189644 318 type motif 2, transcript vari

OSR1 oxidative-stress responsive 1, 33869411 9943

PAK2 p21 (CDKNlA)-activated kinase 2, mRNA (cDNA 47482155 5062 clone MGC:97077 )

PAK4 p21(CDKNlA)-activated kinase 4, 33877350 10298

PANK3 pantothenate kinase 3, 15489194 79646

PCBP2 poly(rC) binding protein 2, transcript variant 2, 12654634 5094

PCTK2 PCTAIRE protein kinase 2, 21542570 5128

PDE4A phosphodiesterase 4A, cAMP-specific 18043808 5141

(phosphodiesterase E2 dunce homolog, Dro

PDK3 pyruvate dehydrogenase kinase, isoenzyme 3, 16198532 5165

PELO pelota homolog (Drosophila), 33870521 53918

PFKFB4 6-phosphofructo-2-kinase/fructose-2,6- 16307443 5210 biphosphatase 4,

PIM1 pim-1 oncogene, 18044377 5292

PIP5K2B phosphatidylinositol-4-phosphate 5-kinase, type II, 20071965 8396 beta, transcript variant 2,

PKLR pyruvate kinase, liver and RBC, transcript variant 1, 19343992 5313

PKM2 pyruvate kinase, muscle, transcript variant 1, 14043290 5315 PLD2 phospholipase D2, 15929159 5338

PLK1 polo (Drosophia)-like kinase, clone MGC:8502 33876611 5347

PMVK phosphomevalonate kinase, 13543886 10654

PPA G peroxisome proliferative activated receptor, 13905055 5468 gamma, transcript variant 3,

PPP2R2B protein phosphatase 2 (formerly 2A), regulatory 21619304 5521 subunit B (PR 52), beta isof

PPP2R2C protein phosphatase 2 (formerly 2A), regulatory 34192271 5522 subunit B (PR 52), gamma iso

PRC1 protein regulator of cytokinesis 1, 13111934 9055

PRKCI protein kinase C, iota, 34191041 5584

PRKCZ protein kinase C, zeta, 33873791 5590

PRKD2 protein kinase D2, 19263754 25865

PRPS2 phosphoribosyl pyrophosphate synthetase 2, 26251732 5634

PTK2 PTK2 protein tyrosine kinase 2, 34786073 5747

PTK9L PTK9L protein tyrosine kinase 9-like (A6-related 16741224 11344 protein),

PXK PX domain containing serine/threonine kinase, 15680248 54899

PYCR1 pyrroline-5-carboxylate reductase 1, transcript 37589044 5831 variant 1,

RAC2 ras-related C3 botulinum toxin substrate 2 (rho 33878888 5880 family, small GTP binding pr

RAN RAN, member RAS oncogene family, 33871120 5901

RBKS ribokinase, 16924286 64080

RDBP RD RNA binding protein, 34193418 7936

RFK riboflavin kinase, 13937919 55312

RIPK2 receptor-interacting serine-threonine kinase 2, 33871163 8767

RPL30 ribosomal protein L30, 34783378 6156

RPS6KA1 ribosomal protein S6 kinase, 90kDa, polypeptide 1, 15929012 6195

RPS6KA4 ribosomal protein S6 kinase, 90kDa, polypeptide 4, 28839795 8986 mRNA (cDNA clone MGC:57704 )

RRAS related RAS viral (r-ras) oncogene homolog, 16740850 6237

RRAS2 related RAS viral (r-ras) oncogene homolog 2, 15341856 22800

SGK serum/glucocorticoid regulated kinase, 12654838 6446

SGKL serum/glucocorticoid regulated kinase-like, 15929809 23678 transcript variant 1,

SNK serum-inducible kinase, 33988188 10769

S0CS4 suppressor of cytokine signaling 4, 18043068 9306

S0CS5 suppressor of cytokine signaling 5, 23273933 9655

STAU1 staufen, RNA binding protein (Drosophila), 29792189 6780 transcript variant T3,

STK17B serine/threonine kinase 17b (apoptosis-inducing), 16359142 9262

STK22B serine/threonine kinase 22B (spermiogenesis 34191166 23617 associated), mRNA (cDNA clone MGC:41904 )

STK25 serine/threonine kinase 25 (STE20 homolog, yeast), 33873686 10494

STK29 BR serine/threonine kinase 2 (BRSK2) 116089334 9024

STK3 serine/threonine kinase 3 (STE20 homolog, yeast), 34189966 6788

STK38L serine/threonine kinase 38 like, 20306368 23012

STMN1 stathmin 1/oncoprotein 18, 15680063 3925

SYK spleen tyrosine kinase, 33876370 6850

TBL1X transducin (beta)-like lX-linked, 21619189 6907 TD KH tudor and KH domain containing protein, 21595811 11022

TK1 thymidine kinase 1, soluble, 39644822 7083

TNFRSF6 tumor necrosis factor receptor superfamily, 15214691 355

member 6, transcript variant 1,

TOLLIP toll interacting protein, 13325203 54472

TPD52 tumor protein D52, 17390256 7163

TRAF3IP1 TNF receptor-associated factor 3 interacting protein 37590615 26146

1,

TRAF5 TNF receptor-associated factor 5, transcript variant 20810008 7188

2,

TRAP100 thyroid hormone receptor-associated protein (100 15030229 9862

kDa),

TRB2 tribbles homolog 2, 33990940 28951

UK114 translational inhibitor protein pl4.5, 16307462 10247

UMPK uridine monophosphate kinase, 33991104 7371

VDRIP vitamin D receptor interacting protein, 13528773 29079

REFERENCES

[I] Harvey et al. (2009) BMC Urol. 9:14.

[2] Shappell (2008) Rev Urol 10:44-69.

[3] Bussemakers et al. (1999) Cancer Res 59:5975-9.

[4] Hessels et al. (2003) Eur Urol 44:8-16.

[5] Groskopf et al. (2006) Clin Chem 52:1089-95.

[6] Marks et al. (2007) Urology 69:532-5.

[7] US-2008/0081339.

[8] US-2009/0226915.

[9] US-2009/0258792.

[10] Boyle & Brawley (2009) CA Cancer J Clin 59:220-4.

[II] Pappworth et al. (2009) Mol Immunol 46:1042-9.

[12] Vanderlugt & Miller (1996) Curr Opin Immunol. 8:831-6.

[13] Cheung et al. (2000) Nucleic Acids Res. 28(l):361-3. http://alfred.med.yale.edu/alfred/

[14] McKusick (1998) Mendelian Inheritance in Man. A Catalog of Human Genes and Genetic

Disorders. Baltimore: Johns Hopkins University Press, 1998 (12th edition). See also http://www. ncbi. nlm.nih. gov/omim/.

[15] Stenson et al. (2009) Genome Med 1:13.

[16] Stamm et al. (2006) Nucleic Acids Res 34: D46-D55.

[17] Sonn et al. (2005) Prostate Cancer Prostatic Dis 8:304-10.

[18] Forrer et al. (1998) Biol Chem 379:1101-11.

[19] Geysen et al. (1984) PNAS USA 81:3998-4002.

[20] Carter (1994) Methods Mol Biol 36:207-23.

[21] Jameson, BA et al. 1988, CABIOS 4(1):181-186.

[22] Maksyutov & Zagrebelnaya (1993) Comput AppI Biosci 9(3):291-7.

[23] Hopp (1993) Peptide Research 6:183-190.

[24] Welling et al. (1985) FEBS Lett. 188:215-218.

[25] Bublil et al. (2007) Proteins 68(l):294-304.

[26] Sun et al. (2009) Nucleic Acids Res 37:W612-6. [27] Raddrizzani & Hammer (2000) Brief Bioinform l(2):179-89.

[28] Chen et al. (2007) Amino Acids 33(3):423-8.

[29] Reimer (2009) Methods Mol Biol 524:335-44.

[30] Boutell et al. (2004) Proteomics 4:1950-8.

[31] Tassinari et al. (2008) Curr Opin Mol Ther 10:107-15.

[32] Stoevesandt et al. (2009) Expert Rev Proteomics 6:145-57.

[33] Tao et al. (2007) Comb Chem High Throughput Screen 10:706-18.

[34] Gnjatic et al. (2009) J Immunol Methods 341:50-8.

[35] Hartmann et al. (2009) Anal Bioanal Chem 393:1407-16.

[36] Fall & Niessner (2009) Methods Mol Biol 509:107-22.

[37] WO01/57198.

[38] WO02/27327.

[39] Blackburn & Hart (2005) Methods Mol Biol. 310:197-216

[40] WO03/064656.

[41] WO2004/046730.

[42] Stahl et al. (2006) Immunol Lett 102:50-9.

[43] Quintana (2008) PNAS USA 105:18889-94.

[44] Koopmann & Blackburn (2003) Rapid Commun Mass Spectrom.17:455-62.

[45] WO01/61040.

[46] Oleinikov et al. (2003) J Proteome Res. 2:313-9.

[47] Bolstad et al. (2003) Bioinformatics 19:185-93.

[48] Meyer et al. (2003) Neurocomputing 55:169-86.

[49] Koza (1992), Genetic Programming: On the Programming of Computers by Means Natural Selection, MIT Press.

[50] Makinani (2005) World Acad Sci Eng Tech 7:166-9.

[51] Zhan & Shen (2004) LNCS 3314/2005.

[52] Wagner et al. (2004) BMC Bioinformatics 5:26.

[53] CJnar et al. (2009) Expert Systems with Applications 36:6357-61.

[54] Keller et al. (2009) Informatik Spektrum 32:332-8.

[55] Elkon & Casali (2008) Nat Clin Pract Rheumatol. 4(9):491-8.

[56] Chada et al. (2003) Curr Opin Drug Discov Devel. 6(2):169-73.

[57] Chene (2003) Nature Reviews Cancer 3, 102-109.

[58] Wang & El-Deiry (2008) Curr Opin Oncol. 20(l):90-6.

[59] Current Protocols in Molecular Biology (F.M. Ausubel et al., eds., 1987) Supplement 30 [60] Smith & Waterman (1981) Adv. Appl. Math. 2: 482-489.

[61] Huber et al. (2002) Bioinformatics 18 suppl. 1 S96-S104.

Claims

I. A method for analysing a subject sample, comprising a step of determining the levels of x different biomarkers in the sample, wherein the levels of the biomarkers provide a diagnostic indicator of whether the subject has prostate cancer; wherein x is 1 or more and wherein the x different biomarkers are selected from auto-antibodies against (i) PXK; (ii) PRKAG3; (iii) PITRM1; (iv) DOM3Z; (v) DCLK2; (vi) GOLGA5; (vii) SGKL; (viii) CSNK2A1; (ix) GRK5; (x) MAPK9; (xi) PSME3; (xii) TLK1; (xiii) ABCF3; (xiv) CTAG2; (xv) DNM1L; (xvi) SAVl; (xvii) SDCCAG10; (xviii) PRKCBP1; (xix) CABC1; (xx) CCM2; (xxi) DBNL; (xxii) MAPKAPK5; (xxiii) MADH5; and/or (xxiv) MRPL55.

2. The method of claim 1, wherein x is 2 or more.

3. The method of claim 2, wherein x is 10 or more.

4. The method of any preceding claim, wherein x is 24 or fewer.

5. The method of claim 4, wherein x is 15 or fewer.

6. The method of any preceding claim, wherein the method also includes a step of determining if a sample from the subject contains PSA and/or PCA3.

7. The method of any preceding claim, wherein the sample is a body fluid.

8. The method of claim 7, wherein the sample is blood, serum or plasma.

9. The method of any preceding claim, wherein the subject is (i) pre-symptomatic for prostate cancer or (ii) already displaying clinical symptoms of prostate cancer.

10. The method of any preceding claim, wherein the presence of auto-antibodies is determined using an immunoassay.

II. The method of claim 10, wherein the immunoassay utilises an antigen comprising an amino acid sequence (i) having at least 90% sequence identity to an amino acid sequence encoded by a SEQ ID NO listed in Table 1, and/or (ii) comprising at least one epitope from an amino acid sequence encoded by a SEQ. ID NO listed in Table 1.

12. The method of claim 10 or claim 11, wherein the immunoassay utilises a fusion polypeptide with a first region and a second region, wherein the first region can react with an auto-antibody in a sample and the second region can react with a substrate to immobilise the fusion polypeptide thereon.

13. The method of any preceding claim, wherein the subject is a human male.

14. The method of any preceding claim, wherein the method involves comparing levels of the biomarkers in the subject sample to levels in (i) a sample from a patient with prostate cancer and/or (ii) a sample from a patient without prostate cancer.

15. The method of any preceding claim, wherein the method involves analysing levels of the biomarkers in the sample with a classifier algorithm which uses the measured levels of to distinguish between patients with prostate cancer and patients without prostate cancer.

16. The method of any one of claims 2 to 15, wherein the 2 or more different biomarkers are:

• A panel comprising or consisting of 2 different biomarkers, namely: (i) a biomarker selected from Table 2 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 2 different biomarkers, namely: (i) a biomarker selected from Table 2 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of 3 different biomarkers, namely: (i) a group of 2 biomarkers selected from Table 3 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 3 different biomarkers, namely: (i) a group of 2 biomarkers selected from Table 3 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of 4 different biomarkers, namely: (i) a group of 3 biomarkers selected from Table 4 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 4 different biomarkers, namely: (i) a group of 3 biomarkers selected from Table 4 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of 5 different biomarkers, namely: (i) a group of 4 biomarkers selected from Table 5 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 5 different biomarkers, namely: (i) a group of 4 biomarkers selected from Table 5 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of 6 different biomarkers, namely: (i) a group of 5 biomarkers selected from Table 6 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 6 different biomarkers, namely: (i) a group of 5 biomarkers selected from Table 6 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of 7 different biomarkers, namely: (i) a group of 6 biomarkers selected from Table 7 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 7 different biomarkers, namely: (i) a group of 6 biomarkers selected from Table 7 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of 8 different biomarkers, namely: (i) a group of 7 biomarkers selected from Table 8 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 8 different biomarkers, namely: (i) a group of 7 biomarkers selected from Table 8 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of 9 different biomarkers, namely: (i) a group of 8 biomarkers selected from Table 9 and (ii) a further biomarker selected from Table 17. • A panel comprising or consisting of 9 different biomarkers, namely: (i) a group of 8 biomarkers selected from Table 9 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of 10 different biomarkers, namely: (i) a group of 9 biomarkers selected from Table 10 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 10 different biomarkers, namely: (i) a group of 9 biomarkers selected from Table 10 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of 11 different biomarkers, namely: (i) a group of 10 biomarkers selected from Table 11 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 11 different biomarkers, namely: (i) a group of 10 biomarkers selected from Table 11 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of 12 different biomarkers, namely: (i) a group of 11 biomarkers selected from Table 12 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 12 different biomarkers, namely: (i) a group of 11 biomarkers selected from Table 12 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of 13 different biomarkers, namely: (i) a group of 12 biomarkers selected from Table 13 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 13 different biomarkers, namely: (i) a group of 12 biomarkers selected from Table 13 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of 14 different biomarkers, namely: (i) a group of 13 biomarkers selected from Table 14 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 14 different biomarkers, namely: (i) a group of 13 biomarkers selected from Table 14 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of 15 different biomarkers, namely: (i) a group of 14 biomarkers selected from Table 15 and (ii) a further biomarker selected from Table 17.

• A panel comprising or consisting of 15 different biomarkers, namely: (i) a group of 14 biomarkers selected from Table 15 and (ii) a further biomarker selected from Table 1.

• A panel comprising or consisting of a group of 15 different biomarkers selected from Table 16.

17. A diagnostic device for use in diagnosis of prostate cancer, wherein the device permits determination of the level(s) of 1 or more Table 1 biomarkers.

18. The device of claim 17, wherein the device comprises a plurality of antigens immobilised on a solid substrate as an array.

19. The device of claim 18, wherein the device contains antigens for detecting auto-antibodies against all of the antigens listed in Table 1.

20. The device of claim 19, wherein the device contains antigens for detecting auto-antibodies against all of the antigens listed in Table 17.

21. The device of any one of claims 18-20, wherein the array includes one or more control polypeptides.

22. The device of claim 21, comprising one or more an anti-human immunoglobulin antibody(s).

23. The device of any one of claims 17 to 22, including one or more replicates of an antigen.

24. The method of any one of claims 1 to 15, using the device of any one of claims 17 to 23.

25. In a method for diagnosing if a subject has prostate cancer, an improvement consisting of determining in a sample from the subject the level(s) of y biomarker(s) of Table 1, wherein y is 1 or more and the level(s) of the biomarker(s) provide a diagnostic indicator of whether the subject has prostate cancer.

26. A human antibody which recognises an antigen listed in Table 17 (preferably in Table 1).