Tuning of the HISAT2 parameters --mp and --sp

Question

[JJ]

Hi all,

I played a little with the --mp and --sp parameters of the HISAT2 aligner.

--mp <int>,<int>   max and min penalties for mismatch; lower qual = lower penalty <6,2>
--sp <int>,<int>   max and min penalties for soft-clipping; lower qual = lower penalty <2,1>

The default settings appear quite harsh and if I relax them (as stated in this paper's supplemental) I obviously get a higher alignment rate (12% more - so I have 99% alignment rate). However, I am not sure if relaxing them that much is a good idea. I am not sure how to access this on real data either. So I wanted to ask for your input/experience. Any help is very much appreciated.

[wang-yanfang]

Dear JJ,

I am using Hisat2 align my bacteria rna_seq, and by using the defalt settings I only get 7% alignment rate, and I am not sure how loose the parameters. Can you give some advice?

Best regards, Yanfang University of Groningen

[JJ]

hi wang-yanfang,

I think think your alignment rate is too low with only 7%. There is something wrong - either with the library or the reference/data processing? Mine was still at >80 %

Look at this paper: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4631051/ It's actually for the STAR aligner, but the mapping rate statement is a general statement. Everything below <50% indicates an issue.

Hope this helps!