GA pipeline chastity/purity filters

[dvh]

Anyone have experience of changing the default settings of the filtering parameters in the GA pipeline used to remove low quality reads?

From the manual "The default filter is equivalent to:
--smt-filter failed-chastity --smt-relation le --smt-threshold 1 --chastity-threshold 0.6 --pure-bases 25 "

We have somewhat more data per sample than we need. We want to do SNP calling, and I'd like to have the highest quality data only as the SNP calling is sensitive to error rate.

regards

David

[gridbird]

where to find GA pipeline? is it free to use? I want to remove low quality reads from SRA.

[csc]

I would hesitate to change the value from 0.6 unless I had a good look at the normalized intensity values of each channel. Illumina claims that 0.6 is quite a conservative cutoff. I think changing the number of cycles that chastity evaluates should be relatively safe though. How long are your reads? You could easily bump the value from 25 to maybe 35 - even higher perhaps. Try running the pipeline on a few different intervals and see how many reads you are left with and what the corresponding quality scores look like until you are happy with it.