Running SpeedSeq with Phred+64 Illumina reads

[brettChapman]

Hi

I've recently come across a forum (https://www.biostars.org/p/89088/) which mentions that bwa mem should be run with Phred+33 Illumina reads instead of Phred+64. My Illumina library is Phred+64, and I did not realise this was an issue until I accidentally stumbled across it.

This is there no mention of this in the bwa mem rep or SpeedSeq rep. Is it a significant problem for me?

I've also been having problems with SpeeSeq seg faulting with Lumpy, and I'm not sure if the problem is related (https://github.com/hall-lab/speedseq/issues/54).

Thanks.

[ekg]

Why not convert the files first? I think you can even stream them into bwa mem.

[brettChapman]

Ok. Thanks for the suggestion.

[cc2qe]

The LUMPY issues are definitely not related to base quality phred scores, but it could be something else about the library design. continuing discussion over at https://github.com/arq5x/lumpy-sv/issues/101