Alternatively, with BBMap:

reformat.sh in=r1.fq in2=r2.fq out1=renamed1.fq out2=renamed2.fq addslash

Yes, I believe that abyss requires the pair information to be present (either 1/2, forward/reverse or A/B) and the files may be separate or interleaved. You can add the pair information back with Pairfq. Here is an example (requires curl and perl):

curl -sL git.io/pairfq_lite | perl - addinfo -i R1.fq -o R1_info.fq -p 1
curl -sL git.io/pairfq_lite | perl - addinfo -i R2.fq -o R2_info.fq -p 2

That should go pretty fast and the input can be fasta or fastq (compressed is fine also I believe).

Solutions posted so far would work great. I just remembered an old blogpost where there was a onliner to convert new illumna naming scheme to old using this one liner:

cat new-style_.fastq | awk '{if (NR % 4 == 1) {split($1, arr, ":"); printf "%s_%s:%s:%s:%s:%s#0/%s (%s)\n", arr[1], arr[3], arr[4], arr[5], arr[6], arr[7], substr($2, 1, 1), $0} else if (NR % 4 == 3){print "+"} else {print $0} }' > old-style.fastq

https://contig.wordpress.com/2011/09/01/newbler-input-iii-a-quick-fix-for-the-new-illumina-fastq-header/

It's kinda nice since you really are not relying on any other tools, just bash and good ol' awk. I think this will only work if you do have the new header (something like 1:N:0 and 2:N:0), it may not if you have no info about pairs in your header.

Thank you lads. I went ahead and assumed that the suffixes are important. Thank you for confirming that.

I used BBMap's reformatter script for this.

I have a related question though. I run abyss-2fastq on some data I was analysing a week ago and it added /1/2 not at the very end but towards the end before the last few characters e.g. below. Is this recognisable by Abyss?

@HISEQ:149:C76YNACXX:3:1101:1159:2191/1 1:N:0:GCCAAT
--------------------------------------^
ATAATTAAAGCAGGAATAGTAAAAAAACGTCCCTTAAAACGTATCAAGAAATCCGACCCAGACTGGGATTACGCAACCTGCGACGGCCCGTTGTGCCTGCG
+
BBBFFFFFFFFFFIBFIFIIIIIIIIIIIIIIIFIFFIIIFFFIBFIIIIIIFFFIFFFFFFFFFFFBBFFFFBFFFFFFBBFFFFFFF<BFFBFBBFFFF
@HISEQ:149:C76YNACXX:3:1101:1159:2191/2 2:N:0:GCCAAT
--------------------------------------^
AACCTTGCGACGACCTGAAGGACGGACCGTCGCAGGCACAACGGGCCGTCGCAGGTTGCGTAATCCCAGTCTGGGTCGGATTTCTTGATACGTTTTAAGGG
+
BBBFFFFFFFFFFFIIIIFFIFIIIFFFFIFFFFFFFFFFFFFBFBBFFF7<77B<BB<BBBBFFFFBBBFBFFF<BBF7BBBBFFB<BBFBBFFF<<BBF

Hello again,

I received a reply from Ben Vandervalk who is one of the authors of Abyss and it goes as follows:

pe="r1.fastq r2.fastq" should suffice.

ABySS requires that either:

(i) the read names for both reads are identical, OR (ii) the read names have an identical prefix, followed by "/1" and "/2", respectively.

- Ben