tagdust output is not fastq!

[joseph]

Hi
can you please help me with the output of tagdust? It doesn't look like the proper fastq format.
Thanks
Joseph Dhahbi

tagdust contaminants.fasta myfile.fastq -fdr 0.05 -o myfile_tagdust.fastq

head -10 myfile.fastq
@HS1:202:C0UY5ACXX:4:1101:1485:2099 1:N:0:ACAGTG
NAAGATAGTTATGAAACAGAAGATGAAAGTTCCTGGGATAATGTTGAGTTAGGAGACTACACTACACAGGCCATAGAAGATGAAACCTATAGTGATATTA
+HS1:202:C0UY5ACXX:4:1101:1485:2099 1:N:0:ACAGTG
#1:A1?>DFF?FC>GICEFGEHBHHIHGG?EAEFHIIBFHEGIFHBCHBFHEGEHIIDF>BBHDHGIGGCGHGDCEC>EFF?B@@@EDDDCD>CDDD<>C
@HS1:202:C0UY5ACXX:4:1101:1458:2150 1:N:0:ACAGTG
CTAGAATAGGATTGCGCTGTTATCCCTAGGGTAACTTGTTCCGTTGGGCAAGTTATTGGATCAATTGAGTATAGTAGTTCGCTTTGACTGGTGAAGTCTT
+HS1:202:C0UY5ACXX:4:1101:1458:2150 1:N:0:ACAGTG
@<@FFADFGGFFFHJIIIIFHEEHGDD@EBA1?DGHIJAGGIJEHGG(93;BHGGCHIFA:==CEH>CC?>B@>A>??)6;A;BBA@AAC<@C>CAC>A;
@HS1:202:C0UY5ACXX:4:1101:1748:2155 1:Y:0:ACAGTG
CTCCGGAGGGACCCTCCAGCTGTGATGAAGGCCCAGAGCCCCATGGAAACAGTGACTGCATAGATGGGCAGCAGGCTGGCCTCATACCAGGGGTTCAGGA

head -10 myfile_tagdust.fastq
@HS1:202:C0UY5ACXX:4:1101:1485:2099 1:N:0:ACAGTG
NAAGATAGTTATGAAACAGAAGATGAAAGTTCCTGGGATAATGTTGAGTTAGGAGACTAC
ACTACACAGGCCATAGAAGATGAAACCTATAGTGATATTA
+HS1:202:C0UY5ACXX:4:1101:1485:2099 1:N:0:ACAGTG
#1:A1?>DFF?FC>GICEFGEHBHHIHGG?EAEFHIIBFHEGIFHBCHBFHEGEHIIDF>
BBHDHGIGGCGHGDCEC>EFF?B@@@EDDDCD>CDDD<>C
@HS1:202:C0UY5ACXX:4:1101:1458:2150 1:N:0:ACAGTG
CTAGAATAGGATTGCGCTGTTATCCCTAGGGTAACTTGTTCCGTTGGGCAAGTTATTGGA
TCAATTGAGTATAGTAGTTCGCTTTGACTGGTGAAGTCTT
+HS1:202:C0UY5ACXX:4:1101:1458:2150 1:N:0:ACAGTG

[maubp]

It might be valid, but appears to be using line wrapping, which is not recommended. Some tools assume 4 lines per record (for speed) and line wrapping breaks them. Other tools have more robust (but slightly slower) FASTQ parsers which will cope.

You could try running this line wrapped FASTQ through EMBOSS seqret or Biopython or something similar which will accept line wrapped FASTQ input, but produce typical 4 line per record unwrapped FASTQ as output.

[SES]

I took a look at main.c in tagdust and saw this at line 52:

	if(!param->linewrap){
		linewrap = 0;
	}else{
		linewrap = 1;
	}

which suggested there was an option to control this behavior. Sure enough, after you compile the program you can see that there is an option to print the sequences on one line (the -s option, specifically).

$ ./tagdust
TagDust version 1.13, Copyright (C) 2009 Timo Lassmann <[email protected]>

Usage: tagdust [options]  lib.fa read1.fa read2.fa ...
	

	Options:
	-f, -fdr	False discovery rate (default: 0.01)
	-o <file>	print clean tags to file.
	-a <file>	print artifactual tags to file.
	-trim5 <X>	trim 'X' residues from the start of all reads.
	-trim3 <X>	trim 'X' residues from the end of all reads.
	-fasta		output format is fasta.
	-s, -singleline	sequences are written in a single line.
	-q, -quiet	quite mode

	Identifies tags as artifactual sequences if they match to library sequences.
	Library sequences must be in fasta; tag sequences in either fasta or fastq format.