Dindel on PGM IonTorrent Data

[adrian]

Dear users,
has anyone used or tested Dindel on PGM IonTorrent data. Although Dindel manual says it is meant for Illumina, is there any new development for optimizing Dindel for PGM data.

Samtools generic Indel prediction missing several known indel due to homopolymer stretches.



thank you.
Adrian

[adaptivegenome]

DINDEL is designed for Illumina data because it uses an error rate derived from that platform. The DINDEL paper describes how the error was calculated and even makes note of accounting for homopolymers, which are error-prone on all platforms, even Illumina. To best use DINDEL I would recommend computing an error-rate for Ion Torrent data. Since DINDEL is also implemented as part of GATK, I would also recommend contacting Eric Banks to see if he or someone in the GSA group is working on an Ion Torrent error model as well. You might be also to save some effort if they are doing it already.