A classifier, an aligner or a colleague gave you a list of read or contig names, one per line, and you need the corresponding sequences.

A00709:43:HYG25DSXX:1:1101:3640:1000
A00709:43:HYG25DSXX:1:1101:6189:1000
A00709:43:HYG25DSXX:1:1101:10818:1000

Simple extraction

seqfu list prints the records whose name appears in the list. Lines starting with # are ignored, and names can keep a leading > or @:

seqfu list ids.txt reads_R1.fq.gz > selected_R1.fq

Add --strict to fail if some names are not found, or --report for a summary on the standard error. With --outdir and several --lists, one output file per list is written in a single pass.

Paired-end extraction

seqfu by-id selects both mates together, so the two output files stay in sync:

seqfu by-id -x -f ids.txt \
  -1 reads_R1.fq.gz -2 reads_R2.fq.gz \
  -o selected_R1.fq -O selected_R2.fq

seqfu count selected_R1.fq selected_R2.fq

-x requires the whole identifier to match; without it, each entry of the list is searched as a pattern. Use --pair-mode both to require that both mates match.

The opposite: remove the listed reads

Invert the selection with -v:

seqfu by-id -x -v -f ids.txt reads_R1.fq.gz > remaining_R1.fq

If you have the unwanted records as a sequence file rather than as a list, seqfu subtract prints the records of the first file that are absent from the second:

seqfu subtract reads_R1.fq.gz selected_R1.fq > remaining_R1.fq