Appending /1 or /2 to read identifiers (illumina 1.8+ format)

[Kennels]

Hello,

I've just started working with Illumina 1.8+ format fastq files, and the read identifiers don't contain the /1 or /2 suffixes for PE reads any more, and this information is contained in column 2.

I'm not scripting savvy, and i'm just using a very clumsy perl script to add this information:

perl -e ' while(<>) { @cols=split("\t", $_); $cols[0] = "$cols[0]/1"; print "$cols[0]\t$cols[1]\t$cols[2]\t$cols[3]\t$cols[4]\t$cols[5]\t$cols[6]\t$cols[7]\t$cols[8]\t$cols[9]\t$cols[10]\t$cols[11]"; } exit '

I'm wondering if anyone can share a good sed or awk command or script to do this more efficiently/quickly as I'm working with a few Gb to tens of Gb+ sized files. I've been trying but can't get the syntax right.

thanks,

[simonandrews]

From what I can see the new fastq files are space delimited, not tab. Since only the header line should contain any spaces you could probably get away with something as simple as:

perl -pe 's/ /\/1 /'

which should be pretty quick.

[pmiguel]

But, if you really want a sed solution, the equivalent would be:

sed -e 's% \([0-9]\)%/\1 %'

or just:

sed -e 's% %/%'

if you want the exact results that Simon's perl one-liner would give you. But the reads I deal with are nearly always indexed, so they look like:

@H-148:620D1KACXX:8:1101:1853:1942 2:Y:0:GCCAAT

So, I don't know what havoc the extra fields will play. My first sed expression results in:

@H-148:620D1KACXX:8:1101:1853:1942/2 :Y:0:GCCAAT

whereas the second gives:

@H-148:620D1KACXX:8:1101:1853:1942/2:Y:0:GCCAAT

Oh, one other I have to include because it highlights a power that sed has, but perl does not:

sed -e 's% %/%' -e 's/:/ /7'

it gives the (possibly more desirable)

@H-148:620D1KACXX:8:1101:1853:1942/2 Y:0:GCCAAT

result. Get it? The final "7" denotes that sed should do the substitution the seventh time it sees ":" on the line! Not a feature in perl s/// commands! Although, no doubt, someone will be able to dig up a pragma that will confer this ability upon perl.

--
Phillip

[Kennels]

Thanks for the help.

that sed function is really cool. I really need to sit down and learn these things properly...