Translate¶
The translate command creates new VCFs, adding a controlled vocabulary of new FORMAT tags. It will only work with VCF files from the supported variant callers.
Translate adds new FORMAT Tags : (A) Jacquard reads input VCF files (B) deriving new tags based on information in each variant record and (C) writes new VCFs appending the new tags to the original record.
Usage¶
$ jacquard translate <input_dir> <output_dir> [OPTIONS]
positional arguments:
| input_dir | Directory containing VCF files (and VarScan high confidence
files)
|
| output_dir | Directory containing VCF files. Will create if doesn’t exist
and will overwrite files in output directory if –force
|
optional arguments:
| –allow_inconsistent_sample_sets | Set this flag if not every patient is
represented by the same set of
caller-VCFs. (For example if you ran
VarScan on only a subset of cases.)
|
| –varscan_hc_filter_file_regex= | Regex pattern that identifies optional
optional VarScan high-confidence filter
files.
The VCF, high-confidence file pairs
should share the same prefix. For
example, given files:
patientA.snp.vcf
patientA.indel.vcf
patientA.snp.fpfilter.pass
patientA.indel.fpfilter.pass
you could enable this option as
varscan_hc_filter_file_regex=
‘.fpfilter.pass$’
|
Description¶
The translate command accepts a directory of VCF files and creates a new directory of “translated” VCF files, which include several Jacquard-specific FORMAT tags and their corresponding metaheaders.
You can either gather all input VCFs into a single directory and run translate once or partition VCFs into separate directories (for example, by variant caller) and run translate once for each input directory. When partitioning into separate input directories, all file names must be unique.
- Currently, translate adds Jacquard-specific FORMAT tags for:
- Allele Frequency
- Depth
- Genotype
- Somatic Status
- Passed: Indicates whether the variant record as a whole passed the VC filters; this tag is used later on when merging translated VCFs
- Reported: This tag is used when merging translated VCFs.
See VCF metaheader excepts below for more details on how values are derived:
Strelka Translated Tags¶
| Tag name | Description
|
| JQ_SK_AF | Jacquard allele frequency for Strelka: Decimal allele
frequency rounded to 4 digits (based on
alt_depth/total_depth. Uses [TIR tier 2]/DP2 if available,
otherwise uses (ACGT tier2 depth) / DP2)
|
| JQ_SK_DP | Jacquard depth for Strelka (uses DP2 if available, otherwise
uses ACGT tier2 depth)
|
| JQ_VS_GT | Jacquard genotype (based on SGT).
Example for snv: ALT=C, INFO SGT=AA->AC is translated as
normal=0/0, tumor=0/1.
Example for indel: INFO SGT=ref->het is translated as
normal=0/0, tumor=0/1.
|
| JQ_VS_HC_SOM | Jacquard somatic status for Strelka:
0=non-somatic,1=somatic (based on PASS in FILTER column)
|
MuTect Translated Tags¶
| Tag name | Description
|
| JQ_SK_AF | Jacquard allele frequency for MuTect: Decimal allele
frequency rounded to 4 digits (based on FA).
|
| JQ_SK_DP | Jacquard depth for MuTect (based on DP)
|
| JQ_MT_GT | Jacquard genotype (based on GT)
|
| JQ_MT_HC_SOM | Jacquard somatic status for MuTect:
0=non-somatic,1=somatic (based on SS FORMAT tag)
|
VarScan Translated Tags¶
| Tag name | Description
|
| JQ_VS_AF | Jacquard allele frequency for VarScan: Decimal allele
frequency rounded to
4 digits (based on FREQ)
|
| JQ_VS_DP | Jacquard depth for VarScan (based on DP)
|
| JQ_VS_GT | Jacquard genotype (based on GT)
|
| JQ_VS_HC_SOM | Jacquard somatic status for VarScan: 0=non-somatic,
1=somatic (based on SOMATIC info tag where sample column is
TUMOR and variant record passed VarScan filter).
|
Jacquard can incorporate VarScan high-confidence files¶
To run translate with VarScan calls, Jacquard requires the VarScan VCF files (snp and/or indel). For each VarScan VCF, Jacquard can optionally accept VarScan somatic high-confidence files; these are supplemental non-VCF files that list variant records which passed a more stringent set of VarScan filters.
When high-confidence files are present, the translate command adds a FILTER field value for low-confidence variant records (i.e. records which may have initially passed filters, but are absent in the high-confidence files).
To use VarScan’s somatic high-confidence files, they must be placed alongside corresponding VarScan VCFs and must have the same file name prefix as their corresponding VCF file. The high-confidence filename suffix can be specified using the command line argument.
- Example VarScan files:
- case_A.varscan.indel.vcfcase_A.varscan.indel.Somatic.hc.filter.passcase_A.varscan.snp.vcfcase_A.varscan.snp.Somatic.hc.filter.passcase_B.varscan.indel.vcfcase_B.varscan.indel.Somatic.hc.filter.pass...