[Feature request] Add number of ambiguous characters in seqkit stats

[apcamargo]

The number of ambiguous characters is often a good metric for genome quality. As far as I know, there's no way to add the number of counts of a given character to seqkit stats. I'm using -G "Nn" to fool seqkit to think that N is a gap character, but it would be nice to have that as an additional (optional) column.

[shenwei356]

Do ambiguous characters mean "N" for DNA/RNA and "N/X" for amino acids?

[apcamargo]

Yes. Sorry, the title is indeed confusing.

[shenwei356]

OK, so it means we just need to count "N" for DNA/RNA, and "N"+"X" for proteins.

[apcamargo]

Yep!

[shenwei356]

added, just count N/n/X/x for any kind of sequence.

• seqkit_darwin_arm64.tar.gz
• seqkit_linux_amd64.tar.gz

[kakuk9]

Hi Wei. Could you also add a feature that allows us to filter nt/aa sequences based on the number and/or proportion of ambiguous characters of individual sequence (i.e. N/X) for fasta files? Thanks :D

[shenwei356]

@kakuk9

$ echo -ne ">s\nactgn\n" | seqkit fx2tab -B NX
s       actgn           20.00

$ echo -ne ">s\nactgn\n" | seqkit fx2tab -B NX | awk -F'\t' '$4 >= 20' | seqkit tab2fx 
>s
actgn