{NORMAL}
{NORMAL_REQUIRED}
[--help]
Prints this help message.
# either one of these two
# why would config file need number of alleles?
# try whether we can remove it
[(-c|--configFile) <configFile>]
The input configuration file. Tells you meta parameter and which
sequences go into which demes.
[(-n|--numPerDeme) <numPerDeme>]
The number of individuals per deme. e.g. 5,3,2 means the first 5
haplotypes are in deme 0, the next 3 are in deme 1, and the next 2 are
in deme 2.
[--vcfFile <vcfFile>]
The sequence file. Just a bunch of sequences. In vcf format.
[--compositeLikelihood <compositeLikelihood>]
The composite likelihood type (LOL, PAC, SuperPAC, PCL, PCLOL, or FILE).
(-p|--paramFile) <paramFile>
The input parameter file.
# maybe break instead of ignoring
# ignoring is ok
[--vcfDiAllelic]
If this switch is set, the vcf file is represented with two alleles
internally. Sites with more then two alleles are ignored.
# maybe break instead of ignoring
# ignoring is ok
[--vcfTriAllelic]
If this switch is set, the vcf file is represented with three alleles
internally. Sites with more then three alleles are ignored.
# hmm, I guess we need this
[--vcfFilterPassString <vcfFilterPassString>]
The String in the FILTER column of a vcf-file that marks a snp as having
passed the filters. ALL OTHER SITES ARE IGNORED
# also good to have
[--vcfOffset <vcfOffset>]
Offset(s) to shift the positions given in the vcf file(s).
[--demoFile <demoFile>]
The demography parameter file.
[--bounds <bounds>]
A string of semi-colon separated list of pairs of doubles, each pair
spearated by a comma. This list gives the bounds for the parameters
during the EM-estimation.
# how to deal with when not needed?
[--ratesFile <ratesFile>]
The file with the exponential growth rates. Has to kind of match the
demogrpahy file.
# good
[--lociPerHmmStep <lociPerHmmStep>]
If this flag is used with k > 1, then each HMM-step consists of k loci,
and we assume there are no recombinations within blocks of k loci. Not
compatible with locus skipping. (default: 1)
# i would say make logUniform default, but it needs parameters
# see what the options are, and then maybe make a default
# make some of these required.
[--intervalType <intervalType>]
The method for creating the intervals.
[single,simple,meanrate,mixtureuniform,old]
[--intervalParams <intervalParams>]
The parameters for generating the intervals. [single=<none>,
simple=<none>, meanrate=<numIntervals>, mixtureuniform=<numIntervals>,
old=<numIntervals>] (default: )
[--printIntervals]
If this switch is set, the intervals used are printed.
# highlight EXCLUDED
[--bedFile <bedFile>]
Include *.bed file(s) that lists all regions of the chromosome that
should be excluded from analysis
[--metaStartFile <metaStartFile>]
A file containing the start points to be used in tab-separated format.
[--metaNumStartPoints <metaNumStartPoints>]
The number of starting points (in each dimension if we have a grid).
[--metaNumIterations <metaNumIterations>]
The number of meta optimization steps for the gentic algorithm.
(default: 1)
[--metaParallelEmSteps <metaParallelEmSteps>]
Number of EM steps to be exectued in parallel during gentic algorithm.
(default: 1)
[--metaKeepBest <metaKeepBest>]
Number of best points to keep after a meta step.
[--metaNumPoints <metaNumPoints>]
Number of points to be used in each meta iteration.
[--metaGridStart]
If flag is set we start with a grid of points, otherwise the start
points are chosen randomly.
[-v|--verbose]
Verbose for the EM procedure.
# make sure it's only required when actual randomness involved
--seed <seed>
The seed to initialize the randomness.
[--numPermutations <numPermutations>]
Number of permutations to use for PAC-like methods.
[--numCsdsPerPerm <numCsdsPerPerm>]
Number of CSDs used for each permutation.
[--diffPermsPerChunk]
If this switch is set, use different set of permutations for each
independent chunk.
[--permutationsFile <permutationsFile>]
Specfiy file(s) containing a list of permutations to be used to analyze
each chunk respectively.
[--startPoint <startPoint>]
The starting point for the EM.
[--numberIterationsEM <numberIterationsEM>]
EM terminates after the given number of iterations.
[--numberIterationsMstep <numberIterationsMstep>]
M-step search terminates after given number of iterations
[--parallel <parallel>]
Specifies number of parallel threads
[--printEmPath]
If this switch is set, we print the parameter values at each EM step.
# make this default
[--coordinateWiseMStep]
Do a coordinatewise M Step optimization. The number of iterations and
the reletive error thresholds are applied coordinatewise
# here is hidden randomness, not nice
[--coordinateOrder <coordinateOrder>]
Order to update parameters for coordinateWiseMStep. If not specified,
use random order.
{HIDDEN}
# leave it
[(-s|--sequenceFile) <sequenceFile>]
The sequence file. Just a bunch of sequences.
# leave it
[--unCompressedSequence]
If this switch is set, the sequence file is uncompressed and contains
full haplotypes.
# these have to do with non-VCF
[--alleles <alleles>]
A comma separated string of alleles, e.g. 'A,C,G,T'.
# these have to do with non-VCF
[--missingAlleles <missingAlleles>]
A comma separated string of the alleles that are interpreted as missing,
e.g. 'N,T,S'.
# dont even know whether this works
[--stationaryForPartiallyMissing]
If this switch is set, the partially missing sites are not ignored, but
the stationary distribution is used for them. Kinda only works with the
multilocus hmm version.
# make this standard and a hidden modifier to not do this
[--renormalizeMutationMatrix]
If this switch is set, we renormalize the mutation matrix so rows sum to
1
# not really sure what this one does
[--makeUniformMutationMatrix]
If this switch is set, we use a uniform mutation matrix instead of the
given one, and adjust the mutation rate accordingly.
# make a good default (0.005)
[--minimalMigration <minimalMigration>]
If this factor is given, you can always migrate anywhere with a minial
rate of min(theta,rho)*<factor>. (Except the rates you estimate)
# leave it
[--useLocusSkipping]
If this switch is set, we use the locus-skipping speedup (requires
uniform PIM mutations)
# leave it
[--ancientDemeStates]
If this switch is set, the diCal hidden states are a triplet (epoch,
ancientDeme, haplotype).
# make migratingEthan default
--trunkStyle <trunkStyle>
The style of trunk. [simple,oldCake,meanCake,multiCake]
# make a good default, probably 'average'
[--cakeStyle <cakeStyle>]
Where in the interval to get the cake factor
[beginning,middle,end,average] (default: null)
# what even
[--printLoci <printLoci>]
If this flag is specified with an integer k, additionally decodes at
every k loci, so that we can print at them.
[--printSnps]
Print at the SNPs as well.
# w00t?
[--addTrunkIntervals <addTrunkIntervals>]
Number of trunk intervals to add. (default: 0)
[--oldCakeRenormalizeSettings]
Use old cake intervals and don't renormalize
[--oldCompressedDataFormat]
In old data format, the compressed sequence file didn't know how many
total loci there were, and this info was gotten from the config file.
# make a good default, probably 0.2
[--nmFraction <nmFraction>]
The initial stepsize of the nelder mead optimizer is this fraction of
the starting point.
# just leave it
[--condOnTransitionType]
If this switch is set, we condition on the type of transition
(recombination or no recombination).
[--marginalKL]
If this switch is set, instead of using EM objective function, we use KL
divergence to average marginal posterior decoding.
# too complicated
[--metaStretchProportion <metaStretchProportion>]
Probability that a new point is chosen in a ridge-like manner, as
opposed to each dimension independently. (default: 0.25)
[--metaDisperseFactor <metaDisperseFactor>]
SD for new Gaussian points is multiplied by this factor. (default: 1)
[--estimateRecomScaling]
If this switch is set, the EM will estimate a scaling factor for the
recombination rate, so inputRecomRate * scaling = recomRate.
[--estimateTheta]
If this switch is set, the EM will estimate theta for each interval
[--thetaIntervals <thetaIntervals>]
factory type by which to build the theta intervals, recommend
customFixed
[--thetaIntervalParams <thetaIntervalParams>]
params for the thetaIntervals factory.
# these didn't work anyways
[--relativeErrorM <relativeErrorM>]
M-step optimization terminates if relative improvement is less then the
given value.
[--relativeErrorE <relativeErrorE>]
EM optimization terminates if relative likelihood improvement is less
then the given value.
[--useParamRelErr]
If this switch is set, we use the relative error of the parameters
(instead of log-likelihood) as stopping criteria of E-step.
[--useParamRelErrM]
If this switch is set, we use the relative error of parameters (instead
of log-likelihood) as stopping criteria for M-step.