Request: Option to filter reads in alignment

[Former user]

I had a mysterious case where I got poor assembly results from KMA.

When digging into it, I discovered that I got too many spurious hits to the template, e.g. a 150 bp read with a 30 bp stretch of conserved bases which aligned to the template and threw off basecalling.

Is there any way to filter which reads is allowed to map? I tried adding -mq 100, but the hit is still present (with a score of < 30). Ideally, I would like to be able to specify something like "a read should have at least 0.7 * length of read in score to be able to align".

[ptlcc]

I can make an option to set a minimum coverage of the read.

Best,
Philip

[Jakob Nissen]

Hey - I’m the one who made this issue, just forgot to log in.

I think the problem with setting the -mq option is that it still showed up in the fragments file. The consensus sequence did change after setting -mq. But a coverage of read option would be very nice as well. Ideally, I’d want to discard any reads where e.g. less than 75% of it maps to the template.

[ptlcc]

I have added the option: “-mrc“.