Need to improve adapter filtering

[scholzmb]

Most tools of comparable speed TRIM off adapters, rather than removing reads with adapters (trimmomatic, etc). Improving this step would go a long way to allowing this tool to replace the two step process of trimmomatic/FastQC for many workflows I see.

[chienchi]

It is trimming not removing. The words in the usage mislead the adapter trimming function. And yes, we plan to rewrite the code with other progarmming language to make the tool with better performance. The fuzzy match on the adapters slows down the process. If user knows the adapters is always in the 5'end/3'end 15 bp, use the options below is much faster.

-5end <INT> Cut # bp from 5 end before quality trimming/filtering

-3end <INT> Cut # bp from 3 end before quality trimming/filtering

[scholzmb]

I am so updating all of my QC stuff.

Generally, adapter problem is that in a normal length distribution (say 200 BP fragment length, 2X100 BP reads), a non-zero fraction of reads will have some amount of adapter on them, and these need to be trimmed off. Can't cut. This may be more typical for RNAseq than genomic reads.