All Base Qualities Are Low

[fuad193]

Hi again

sorry for posting again about the same topic. I will try to explain my problem in more details:

I have recently received whole genomic reads with Phred+33 quality from CompleteGenomics (FileFormat v1.5), but the problem is that all bases in all the reads files are range from 33 to 61 ASCII which mean that Phred is range from 0 to 28 only.

In another old exome seq data which comes from our partner's lab the ASCII ranges from 33 to 126 and I already analysed it with no problems.

1) Why the reads quality from CompleteGenomics looks different (i.e range only from 33 to 61 ASCII) ?

2) Do I need to rescale the quality before run it at BWA, Bowtie2, or whatever aligner ?

Thanks

[tonybolger]

In another old exome seq data which comes from our partner's lab the ASCII ranges from 33 to 126 and I already analysed it with no problems.

I would question this data first, to be honest. I don't know of any technology which gives reads with Q scores in the order of 90 - AFAIK that would be 1 error per billion bases. Assembled data can have Q scores that high, but not raw data.

1) Why the reads quality from CompleteGenomics looks different (i.e range only from 33 to 61 ASCII) ?

It's likely correct - scores of Q20 - Q30 (error rates of around 1:100 to 1:1000) are normal for raw reads, if a little below the maximum you could hope for.

2) Do I need to rescale the quality before run it at BWA, Bowtie2, or whatever aligner ?

No, the Q scores you have are reasonable and easily sufficient for alignment.