Cannot repair BWA paired reads that have different name

Question

[Alex]

The error that I am receiving is:

[bwa_sai2sam_pe_core] paired reads have different names: "SRR567550.1.1", "SRR567550.1.2"

I think that the problem is the '.1' and '.2' and have attempted to resolve this using repair.sh from BBmap, as in bwa sampe: paired ends with different names. Unfortunately my fastq files are both ~8 GB and java runs out of memory even when allocating the maximum memory that I can and setting compression to 9 (max).

One idea that I have had is to just take the first column using awk, but this will result in information loss.

I am unsure how to resolve this. Is there a quick/efficient way to strip the '.1' and '.2' while preserving information?

Answer 1

[GenoMax]

Download the original fastq format files for this sample from EBI-ENA here and avoid this issue altogether.

If you used fastq-dump to dump the reads out then you should have used -F option to get the original read headers without this extra stuff from SRA.

[Alex]

Thank you, I didn't know about either of these options.