option xsize has no effect with certain commands and options (fastx_filter and chimeras)

[frederic-mahe]

According to the vsearch -h command, the option --xsize removes the abundance annotation from fastq/fasta headers. But it has no effect when used with --fastx_filter (with vsearch 2.5.0):

vsearch --fastx_filter <(printf "@s;size=1;\nA\n+\nI\n") --xsize --fastqout - 2> /dev/null

returns:

@s;size=1;
A
+
I

[gregcaporaso]

I'm also noticing this with the --chimeras output (but not the --nonchimeras output) for both --uchime_ref and --uchime_denovo.

$ vsearch --uchime_ref    seqs-w-sizes.fasta --uchimeout out.tsv \
 --nonchimeras nonchim.fna --chimeras chime.fna --dn 1.4 --mindiffs 3 --mindiv 0.8 \
 --minh 0.28 --xn 8.0 --xsize --qmask none --db 76-otus/dna-sequences.fasta --quiet

$ head -1 chime.fna
>chime2;size=1

$ head -1 nonchim.fna
>eb3564a35320b53cef22a77288838c7446357327

$ vsearch --uchime_denovo    seqs-w-sizes.fasta --uchimeout out.tsv \
 --nonchimeras nonchim.fna --chimeras chime.fna --dn 1.4 --mindiffs 3 \
 --mindiv 0.8 --minh 0.28 --xn 8.0 --xsize --qmask none --quiet

$ head -1 chime.fna
>chime1;size=1

$ head -1 nonchim.fna
>e84fcf85a6a4065231dcf343bb862f1cb32abae6

It also looks like the abundance information is retained in the --uchimeout. It would be helpful for my application if --xsize impacted that as well. For example, a line in out.tsv from my uchime_ref command above:

0.0384	fd7a7221db15d82dbea087a1ecb66ce871978e7b;size=4	268769	516290	268769	84.2	82.7	75.5	66.9	82.7	17	54	10	8	5	1.4	N

[torognes]

I plan to fix this on Monday 18 Dec.

[torognes]

The reported problems should be fixed in release 2.6.2. I suspect the problems may apply to other commands than fastx_filter, uchime_denovo and uchime_ref, so I will keep the issue open until I have checked that out. The code involved also needs some refactoring.

[gregcaporaso]

Thanks for the quick response @torognes!

[torognes]

Fixed in version 2.7.0 just released.