Relax genome preparation filtering criteria

Historically, we used to filter out any positions from VCF files where the alternative allele was not defined as as a single base (probably a relic from the days when there was one VCF file for a single strain). For the current mouse genomes project VCF file this seems overly harsh though since there may be different strains that are homozygous for different bases but at the same position.

Here is an example:

chr  //  pos  //  REF  //   ALT  //  GT strain1 // GT strain2  // GT strain3
 1      135446     G        A,T          0/0           2/2           1/1

Here all three strains would be homozygous compared to the reference, strain1 would have the same sequence as the reference, i.e. G/G, strain2 would be T/T and strain3 would be A/A. Can we please include these multiple variants as valid positions for the genome preparation.

FelixKrueger / SNPsplit

Relax genome preparation filtering criteria #8