fastq with reads missing 3rd line

[a_mt]

Hi all,
I have a set of fastq files. Some of the fastq files have reads which are missing the 3rd line (which begins with +).

@HWI-ST750:151:C1C6AACXX:5:2316:17997:100881 1:N:0:
AGCGGTNCGCAATATTTTAGTAGCTCGTTACAGTCCGGTGCGTTTTTGGTTTTTTGAAAGTGCGTCTTCAGAGCGCTTTTGGTTTTCAAAAGCGCTCTGAAGTT+
1:+B+0#2<DDDDIIIIIIFIIIIIIIIIIIIIEEIID?DDDBBADIII@DDDDD@@AAAAAAA??A?ADAAAA?????A>?8<>?>AAAAA8>>?><AAAA:4
@HWI-ST750:151:C1C6AACXX:5:2316:19129:100793 1:N:0:
AGCGCTNCTTGATATCAATCAACTGCTAGACAAATCCAATAGTAAATTGGGTAAACCAAATCTCGATATCGACAGCAAAGTATCACAATATGCCTATAACTACA+
;1=DD?#2ACDDDIDEEIIIEIEIEIEEIIIIDEIEIIIDEDDCEDDDEID?BBDIDIIIIEECDIDA@DD=A?D@DAAAA@DDDA>AAABE>AAAA>A>A>AA
@HWI-ST750:151:C1C6AACXX:5:2316:19695:100854 1:N:0:
AGCGCTNCACCGCGGTAAGCTTTAGCAGATCTCACTTTGTCTAGCGTTTGAACCATGTTTTCAAGGATATTGGCTCTAAGTTGTGGGTATTTTTCGATCACTTC+
@<1DDD#2<DFDDGI@CEEGHIIIIIIIEGIIHCGHIIIHGGIGIIAFFHFHAHHIG?CCHFHEEBBC@CDCCCCACCCC5>CCBBB'>ACDECCBBDB7?CC>

And also sequence line contains the + at the end. I guess 3rd line has been concatenated to the end of 2nd line.
Any thoughts on how to proceed with this kind of data?? Any scripts to change it into proper format ??

[jiaco]

If this problem is consistent throughout the file

sed "s/+$/\\`echo -e '\n\r'`+/g" bad.fastq > good.fastq

should do the trick.

EDIT: never remember adding a newline with sed to be complicated like that, but just tested on a mac and this was required. Maybe on linux it is simpler but I do not have a system here to test.

[a_mt]

Thanks for the snippet jiaco.
Even I had tried this before but this aslso messes up with quality score line which are ending with +

And also there are reads having empty lines in between. My question is what is the source of this kind of output?? Is this some sort of of sequencing error ??

@@1D4A#2AFHHFIHIIIIIIIIIIIIIIIIIBHHIIIIIIIIIIIIIIIIIIIIIIIDEHIIIHFFHFEEBDEEECCCBBBCC?CB?CCCBBBBB@BBBBBBB/1

@HWI-ST750:151:C1C6AACXX:5:2316:9996:50328/1
GGCCCCNATACATTTACTGATTCATCCTCAGCGGACTCTGATATGACATCCACTAAAAAATATGTCAGACCACCACCAATGTTAACCTCACCTAATGACTTTCC+
=71?A@#23CDCD@E@ED?FEFCEI<ECFEA>CDDD6?BDEEC9<DBEEIC<BEEIE3@8?;=>?BA>A:(;;@;=???3:>>D####################/1

@HWI-ST750:151:C1C6AACXX:5:2316:9999:44022/1
GGCCACNATCTCGATAATTATAAGATATCTTTAGCACAGGCAAATTGGAACGCAAGCGAAGTTTCGAAAAAGCTAGTAAATATTCAAACAGATGGGTCTATTTC+
???D;B#2ADDDDIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIEIIIIIIIIIIID?CDEDDD@AAA?AAAADEAEEDDDEDBA?AAAAA?A>?ADDD3/1

[jiaco]

You could expand the expression to match

/^[ACGT].*+$/

to avoid quality lines, but I have no idea where you got the file, let alone how it got corrupted.

EDIT: saw your new example just now, there is an issue with this file. Maybe someone else has seen it before.
But I would not try to fix this mess. You need to re-acquire the data.

[a_mt]

Yes, sequence files were given to me by our sequence provider, which I demultiplexed. But after demultiplexing this is the result. May be there is an issue with this. Anyways I will contact them. Thanks for the suggestion.

[sklages]

How do the original files look like? Format?
How did you multiplex? What program?

Sven