Quality of Sequencing Run

[win804]

Hi Everyone,
I would like to ask, is there anyway we can determine the quality of a sequencing run? In short, is there any quality metric for this purpose? for example: the number of mapped reads, number of reads with adapter inserted, etc?

I would be glad if someone can point out any reference or guideline for this.

Thanks a lot.

[NextGenSeq]

Probably the quality score values from the run are the best way to determine quality.

http://seqanswers.com/forums/showthread.php?t=5268&highlight=quality

[ScottC]

Although it depends on the purpose of your evaluation... for instance, Illumina don't have a quality score metric for determining whether a run is within machine specifications.

[win804]

Thanks for the reply. I know about the quality score of a read sequence, but I am more concern about the quality of the overall run. Is there any quality metric for that? For example: how many % of the reads have to pass filter, etc. How do people normally judge whether their run is fine or not?

[ScottC]

If you're judging a 'technically successful' instrument run, as defined by Illumina, rather than how useful the data is from a user's point of view (they can be quite different), then you should be looking at the number of clusters passing filter (>80%), the error rate of the PhiX174 control (<1% for read lengths below 75b, 1.5% for 75b reds and 2% for 100b reads). That's using V1.6 software (CASAVA, OLB, SCS/RTA etc), V4 cluster generation and sequencing kits. They do actually say on their 'specifications' sheet on their website that there should be greater than 70% Q30+ bases, but when we have queried this, they tell us that this is not an 'official' specification!

http://www.illumina.com/Images/products/gaiix_specs.gif

You may also find this tool useful for generating some summaries:

http://www.bioinformatics.bbsrc.ac.uk/projects/fastqc/

Cheers.

[win804]

ScottC, thank you very much for your answer. That really helps.

[ScottC]

By the way, if the runs don't meet those specs, then Illumina should replace the reagents, too.

[tim]

fastx toolkit

You should definitely take a look at the fastx toolkit. There is code there to generate summary statistics for each cycle: quality and nucleotide (including blanks). It generates some very nice graphs which enable you to get a very quick idea of the quality of each cycle, how the quality deteriorates with increasing cycles and whether there is any adapter left in the sequences (which would make mapping impossible for the affected reads). I think some of the tools in the toolkit are also available on Galaxy.

Tim.

[bioinfosm]

Tim,

I used fastx and liked it, but fastqc toolkit may provide even more detail on the QC for a solexa sequencing run.

sm

[tim]

Thanks for the suggestion of fastqc. Very nice.