Home

variant

This repository houses code dealing with variant calling or variant disambiguation. Currently, the repository contains just one program: AlleleCaller.jar

AlleleCaller

Quick start

A copy of the JAR file can be found in this repository. Run the program like so:

java -jar AlleleCaller.jar --bam input.bam --vcf mouse-snps-all.vcf --gt1 C57BL --gt2 CAST

where

• --bam: The BAM file of aligned reads.
• --vcf: The VCF file corresponding to the appropriate reference (e.g, mm9). This file must be indexed. Members of the Guttman Lab can find an indexed VCF file for mm9 SNPs at /mnt/new-storage/Genomes/mus_musculus/mm9/snps/mouse-snps-all.vcf. Names of the different mouse strains are listed in line 11 of mouse-snps-all.vcf.
• --gt1: The name of the first strain, matching one of the names in the VCF file.
• --gt2: The name of the second strain, matching one of the names in the VCF file.

The above command will produce four BAM files in your working directory:

• input.C57BL.bam: reads containing SNPs consistent with C57Bl
• input.CAST.bam: reads containing SNPs consistent with CAST
• input.ambiguous.bam: reads originating from a region identical in both C57Bl and CAST
• input.conflicting.bam: reads containing SNPs inconsistent with both C57Bl and CAST

Warning

AlleleCaller is missing some basic functionality. This functionality will be added as needed by our lab.

AlleleCaller does not support paired-end alignments yet. AlleleCaller will accept a paired-end BAM file, but it will process it as though it contained single-read alignments. It is possible (probable) that two mated records will be sorted into two different output BAM files -- for example, if one has a SNP and the other does not.

AlleleCaller only supports SNPs. It does not recognize indels. Furthermore, AlleleCaller skips positions with heterozygous SNPs (SNPs represented as 0/1, 1/0, 0/2, 2/0, and so on -- contrasted with the homozygous 0/0, 1/1, 2/2).