Supported options

Pisces will accept command-line arguments.

Example command line call:

CallSomaticVariants.exe -B \path\to\my_S1.bam -g [D:\Genomes\Homo_sapiens\UCSC\hg19\Sequence\WholeGenomeFASTA] -f 0.01 -fo True -b 20 -q 100 -c 0 -s 0.5 -a 20 -F 20 -gVCF True -i \path\to\my_interval_file.txt -CallMNVs true -MaxMNVLength 15 -MaxGapBetweenMNV 5

option	default	type	safe?	notes
-a	20	int	yes	variant will only be considered if its variant qscore is >= this value
-b	20	int	yes	a base in a read will only be used if its basecall qscore >= this value. Note that this sets a lower bound for the frequency at which Pisces can call variants. Because if the basecall qscore filer is 20 = 1/100, then the noise level in the input data passing this filter has an upper bound at 1%. Put another way, if the noise is 1% and the var freq is 1% it is impossible to resolve signal from noise. (so if you want to call variants with f=0.001, -b must be 30 or more)
-B [pisces5]	none	string	yes	path to each bam file as comma-separated list with no spaces. Yes, the bam directory as an option has been deprecated.
-B [pisces4]	none	string	yes	path to a bam file or folder with bams in it.
-c	10	int	yes	variant will only be considered if its coverage depth is >= this value. If writing gvcf, "LowDP" filter will be applied and the loci will be a no-call (GT ./.). If outputting a vcf file, no vcf line will be written.
-CallMNVs or -PhaseSNVs	false	bool	yes	call MNVs. For example if C->T always happens at position 4 and G->C always happens at position 5 on the same read, they will be called as MNVs, instead of two separate SNVs. Note, just because two SNVs are adjacent in the vcf, does not mean the will be reported as phased. they might be trans and not cis. Inter-codon phasing is critical for determination of functional consequence.
-chr[Pisces4]	none	string	yes	restrict var calling to a chr of interest. Will process all chr if left unset. Example use: -chr chr3 or -chr [chr3,chr4]
-f	0.01	float	yes	variant will only be considered if its frequency is >= this value. Note that the given min frequency will affect the # of significant figures of the VF reported in the output vcf. if f is 0.01, 3 significant digits after the decimal point are provided for VF in the VCF. Do not set f to zero. Please read the "-b" argument to understand the interplay between -f and -b options. In the Isis wrapper (in Isis v2.6.15) if -f is set too low, -b will be automatically adjusted.
-fo	false	bool	yes	filter out variants found on only one strand (stronger than regular SB filter)
-F	30	int	yes	variant will only be considered PASS if its variant qscore is >= this value. Otherwise, will be filtered with "q[N]" (N being 30 by default) in the filter column of the vcf. In Isis -f and -a are always set to the same value when calling Pisces, so this filter is rarely applied by Pisces. (Isis applies its own Q score filter in postprocessing.)
-g	none	string	yes	path to genome file or files, comma separated with no spaces. -g [path1,path2]. In Pisces 5, multiple genomes should be listed in order 1-1 with input bam files. In Pisces 4, multiple genomes should be 1-1 with the genomes in the genome list from the calling workflow's Illumina.SecondaryAnalysis.GenomeFolders .
-gVCF	false	bool	yes	output gvcf or just vcf files (will do one or the other) Note, a pisces genome.vcf will still be restricted to the calling intervals, if supplied.
-i [Pisces5]	none	string	yes	use interval files to restrict the calling range. In Pisces 5, -i [interval file1, interval files 2] should be explicitly listed in order 1-1 with input bam files. See the input file section for interval file format.
-i [Pisces4]	false	bool	yes	use interval files to restrict the calling range. In Pisces 4: If gvfs are selected, (a) if -i true, results will be reported for all loci in the intervals or (b) if -i false, for all positions with coverage. See the input file section for interval file format.
-MaxGap BetweenMNV	10	int	yes	Do we want to allow strings of reference bases between MNVs? And if yes, how many? For example ACGG-> GCGT has a gap of two bases. CG are reference calls.
-MaxMNV Length -MaxPhased SNVLength	15	int	yes	How long MNVs do we want to call? For example ACGG-> GCGT has a length of 4. The longer the length, the longer PISCES will take to run. Also, the longer the length, the more potential MNVs are possible (so the same genomic event may get split into several different overlapping MNVs, and may no longer look like the SNV you think you are looking for). Its best to use the minimum value to suit your need.
-OutFolder	none	string	yes	all output will be directed to this folder. Nothing on the input directory will be changed.
-q	100	int	yes	The maximum variant quality Q score
-r [Pisces4]	none	string	yes	run folder, where Pisces will go first to find the sample sheet. It uses the sample sheet to figure out which interval files and genome files go with which bam. In Pisces 5 the "-r" option is removed, because reliance on the sample sheet has been removed.
-Report NoCalls	false	bool	yes-but talk to me	New feature, use with caution. Will report the fraction of total bases at the given loci that were not callable (number of Ns at that loci + number of bases with basecall quality less than the minimum filter setting) in the genotype field (NC tag). This gives a measure of how sub-sampled the bases were when the variant was called. Since basecalling can be biased and machine and cycle dependent, this can have a big impact on reported frequencies, especially for Amplicon data.
-t	10	int	yes	By default, Pisces threads by sample. it will use up to the given number of threads. If you only have two samples, and you use -t 10, you will get the same performance as if -t 2.