Filter multi-fasta by length

[hchang10]

Hi everyone,

I'd like to filter a multi-fasta file by length, for example, keep length <300bp and filter out longer ones. I knew there is tool in Galaxy doing this, but it takes so long to upload the file. Can anyone share a script or tool to me?

Thank you very much.

[GenoMax]

See this thread for script examples: http://seqanswers.com/forums/showthread.php?t=13966

[dariober]

Hi everyone,

I'd like to filter a multi-fasta file by length, for example, keep length <300bp and filter out longer ones. I knew there is tool in Galaxy doing this, but it takes so long to upload the file. Can anyone share a script or tool to me?

Thank you very much.

Hi- This strategy should be quite fast (untested):

samtools faidx myfasta.fa
awk '{if($2 < 300) print $1 "\t0\t" $2 "\t" $1}' myfasta.fa.fai > selectSeq.bed
fastaFromBed -fi myfasta.fa -bed selectSeq.bed -name -fo  selectSeq.fa

fastaFromBed is part of bedtools

Dario

[hchang10]

Awesome!!!

Dario, I appreciate it!!!!

Thank you so much!!!!

[peromhc]

Also, a BioPython solution. In this case filtering my length 200nt

#!/usr/bin/python
from Bio import SeqIO
import sys
import os

#usage: python long.seq.py in.fasta out.fasta 200

input_seq_iterator = SeqIO.parse(open(sys.argv[1], "rU"), "fasta")
short_seq_iterator = (record for record in input_seq_iterator \
                      if len(record.seq) > int(sys.argv[3]))

output_handle = open(sys.argv[2], "w")
SeqIO.write(short_seq_iterator, output_handle, "fasta")
output_handle.close()