.Ab1 Trace File Quality Threshold

Question

[Nicolas Rosewick]

Hi,

I've several .ab1 files from a sanger sequencing experiment. What is a good threshold for the quality of the sequencing ? And is it a way to export them into a fastq file. It would simplier to assemble them into a longer contig.

Thanks,

N.

Answer 1

[ALchEmiXt]

In addition to the answer by Casey you could use "old class" assemblers that handle qualityscores directly like (academic free) PHRED/PHRAP/CONSED and their derivatives. Alternatively use some DNA sequence analysis solution as provided (commercially) by for instance DNASTAR or CLC bio.... But most likely there will be free substitutes for those as well. We do all our Sanger reads this way.

[Neilfws]

+1. Sanger ABI reads predate fastq by some years and there are lots of free tools available which may be "old", but still work perfectly well.

[Casey Bergman]

Another old-school assembler that's easy to use and works well on fasta + quals is PCAP: http://seq.cs.iastate.edu/

Answer 2

[Casey Bergman]

For a single Sanger read, you should be using at least Q>20 and preferable Q>30, although there is no hard and fast rule. Read more about what Q-scores mean here: http://en.wikipedia.org/wiki/Phred_quality_score

Galaxy has a fasta+qual -> fastq converter under NGS: QC and manipulation -> Roche-454 data -> Combine FASTA and QUAL into FASTQ. You could use this on the web or pull this tool out of the galaxy distribution and run it locally.

Answer 3

[lh3]

See io_lib from Staden.

As to trimming, you should consider the algorithm used by phrap/phred, which is later used in bwa. Setting a simple per-base threshold is not the best way to trim reads. The bwa manual gives the algorithm in the -q option of the aln command. It is quite simple once you understand the math. There is a seqanswers thread explains what it does, but I cannot find it now.