FreeBSD Manual Pages
AD2VCF(1) General Commands Manual AD2VCF(1) NAME AD2VCF - Extract allelic depth from a SAM stream and add to VCF files SYNOPSIS ad2vcf file.vcf minimum-MAPQ < file.sam samtools view [flags] file.{bam|cram} | ad2vcf file.vcf minimum-MAPQ PURPOSE ad2vcf is designed to augment VCF files lacking allelic depth (AD in FORMAT) with allelic depth extracted from a SAM stream for the same sample used to generate the VCF. It was specifically created for a TOPMed project to augment dbGaP BCF files lacking AD info in preparation for use with haplohseq, which re- quires either NGS data with AD info or micro-array data. DESCRIPTION ad2vcf scans a single-sample VCF file one call at a time and simultane- ously scans the SAM stream for the same sample, extracting the alleles from the read sequence for each call position in the VCF. Output is compressed using xz and saved in file-ad.vcf.xz. Both files must be sorted by chromosome and position. Otherwise, it would not be possible to perform this task in a single pass. Since a single read in the SAM stream could overlap multiple VCF calls, SAM alignments are cached until the next call in the VCF stream is beyond the end of the read. Usually this results in no more than a few thou- sand SAM alignments being held in memory at once, but in very rare cases, we've seen hundreds of thousands for a particular sample. Processing a single-sample human VCF and CRAM file from SRA takes about 15 to 30 minutes on a typical processor, with CRAM decoding being a ma- jor bottleneck due to it's complex compression. For this reason, ad2vcf was intentionally designed as a filter, so that it can utilize a second core while samtools view saturates the first core decoding the CRAM file. If is advisable to filter out questionable alignments before feeding data to ad2vcf. For example, samtools can remove unmapped, secondary, qcfail, dup, and supplementary alignments using --excl-flags 0xF0C. This will greatly reduce the number of buffered alignments in rare cases, preventing runaway memory use to buffer alignments that are probably not useful. SEE ALSO vcf-split, haplohseq BUGS Please report bugs to the author and send patches in unified diff for- mat. (man diff for more information) AUTHOR J. Bacon AD2VCF(1)
NAME | SYNOPSIS | PURPOSE | DESCRIPTION | SEE ALSO | BUGS | AUTHOR
Want to link to this manual page? Use this URL:
<https://man.freebsd.org/cgi/man.cgi?query=ad2vcf&sektion=1&manpath=FreeBSD+Ports+15.0>