Entering edit mode
4.6 years ago
dark.lord
▴
30
Hey guys,
I noticed that with featureCounts the -Q (quality) makes a big difference. I tried with no -Q (default = 0) and -Q 10, and the results were pretty different:
Q 10
Geneid Chr Start End Strand Length sorted_mapped_3_rRNA.bam
PROKKA_00001 Bin3_16S 1 1533 + 1533 610692
PROKKA_00002 Bin2_16S 1 1533 + 1533 182
PROKKA_00003 Bin1_16S 1 1533 + 1533 4111
Q 0
Geneid Chr Start End Strand Length sorted_mapped_3_rRNA.bam
PROKKA_00001 Bin3_16S 1 1533 + 1533 832450
PROKKA_00002 Bin2_16S 1 1533 + 1533 234918
PROKKA_00003 Bin1_16S 1 1533 + 1533 215641
My question is: how do I properly select a reasonable mapping quality score?
Best and TIA,
Stefano
Mapping Q scores are not implemented in the same way by different aligners. Take a look at this blog post for more. If something multi-maps (Q = 0) is probably the best take away.
Thanks @genomax. I think I will go with Q = 0, also because I made trimming and bbmap'd with minid=1, so reads in the SAM file should have already a good quality.
Read the blog and makes me wonder, why are there so few instances of standardisation in bioinformatics? :(