-
Notifications
You must be signed in to change notification settings - Fork 11
Options
-h | --help show this help message and exit
Required
-i | -in FILE Sample information [ID, BAM-PATH, VCF-PATH, GENDER]
The sample information file is required by SV2.
The sample information file must be either tab or space delimited and must contain four columns. Each line contains one sample to be genotyped by SV2.
Multiple samples can be run in parallel with the -c | -cpu argument.
SV2 can take multiple files containing SV predictions as input. BED and VCF files are supported.
-b | -bed ... BED file(s) of SVs
One or more BED files can be passed to SV2, separating by a space.
$ sv2 -i in.txt -b del.bed dup.bed
BED files are either space or tab delimited, formatted as CHROM START END SVTYPE.
Details on the required BED format
-v | -vcf ... VCF file(s) of SVs
One or more VCF file can be passed to SV2, separating by a space.
VCF files are tab delimited, END= and SVTYPE= are required in the INFO column.
Details on the required VCF format
-c | -cpu INT Parallelize sample-wise: 1 per CPU [1]
Given more than one sample in the sample information input, SV2 can perform preprocessing and feature extraction in parallel. Each subprocess operates on one sample, this is limited by the number of cores on a CPU.
By default SV2 will run each sample serially. Note that SV2 does not parallelize across chromosomes.
-g | -genome STR Reference genome build [hg19, hg38]. Default: hg19
Accepted reference genome builds for SV2 are hg19 (GRCh37) or hg38 (GRCh38). Accepted command line argument strings are either hg19 or hg38.
-pcrfree GC content normalization for PCRfree libraries
SV2 performs a GC content normalization for coverage estimates adapted from CNVator. Supply this flag if the samples in the sample information list were sequenced with PCRfree chemistries.
By default this flag is off and SV2 assumes samples were sequenced with PCR protocols.
-M bwa mem -M compatibility. Split-reads flagged as secondary instead of supplementary
SV2 can accommodate legacy alignments with chimeric reads flagged as secondary. Pass the -M flag if samples in the sample information file were aligned with bwa mem -M.
By default SV2 assumes chimeric reads are flagged as supplementary (-M is off).
-s | -seed INT Random seed for genome shuffling in preprocessing [42]
During preprocessing, SV2 randomly selects reads from each chromosome to generate basic alignment statistics. The random seed is set at 42.
-o | -out STR Output name
Prefix for the output files in sv2_genotypes/
-merge Merge SV after genotyping
SV2 can merge breakpoints that are reciprocally overlapping by 80% (by default). This step is done reciprocally until no more SVs can be merged. The SV position with the maximum ALT genotype likelihood is retained.
By default SV2 does not merge breakpoints.
-min-ovr FLOAT Minimum reciprocal overlap for merging SVs [0.8]
Users can define the minimum reciprocal overlap required for merging SVs after genotyping. The -merge flag is not required if -min-ovr option is passed.
-pre PATH Preprocessing output directory. Skips preprocessing
Users can skip preprocessing by passing the path of the sv2_preprocessing/ directory to the -pre argument. Doing this will instruct SV2 to load the values in sv2_preprocessing/ skipping this step. This is useful if users wish to genotype a different set of variants in previously processed samples.
-feats PATH Feature output directory. Skips feature extraction
Passing the path of the sv2_features/ to the -feats argument will skip this step. This is useful for users that wish to generate a genotype matrix containing multiple samples. An example of skipping feature extraction.
-load-clf PATH Add custom classifiers. `-load-clf <clf.JSON>`
SV2 can incorporate new classifiers for genotyping. Packaged with SV2 is a guide on training new classifiers. The output of this guide is a JSON file containing paths to the new classifier.
Pass the JSON file to the -load-clf argument to add more classifiers to SV2. More details are located in the Training section of the User Guide.
-clf STR Specify classifers for genotyping [default]
After loading a new classifier, specify the name of the classifier in the -clf argument to genotype variants with that classifier. The original classifier from SV2 is named default, and this is the default classifier.
Before genotyping, users have to supply the full path to FASTA files for SV2. At least one FASTA file is required for SV2 to run. Configuration needs only to be executed once or updated if the FASTA paths change.
-hg19 PATH hg19 FASTA file
-hg19 takes the full path to a faidx indexed FASTA file for the hg19 (GRCh37) reference build.
-hg38 PATH hg38 FASTA file
-hg38 takes the full path to a faidx indexed FASTA file for the hg38 (GRCh38) reference build.