Recognition of Fastq format

[mountainogre]

Hi there,
We are re-analyzing several Solexa Fastq files generated during the past couple of years. They were not all created on the same version of the Illumina Genome Analyzer pipeline. Does anyone know of a straightforward (and perhaps automatic?) way to recognize the precise fastq format so we can analyze these data appropriately?
Many thanks.

[svl]

The format of the fastq files should be equal, I guess..? Do you mean the different quality value encoding? If so, here is a description:

http://en.wikipedia.org/wiki/FASTQ_format#Encoding

You can use several converters, listed a bit lower on the page.

[maubp]

You have to look at the range of ASCII quality scores present, and this will allow you to exclude some of the formatting variants.

For example, since the Solexa scores can be negative you can get ASCII values from to 59 (Solexa -5) to 63 (Solexa -1), while PHRED scores must be zero or more (thus ASCI 64 or higher for Illumina 1.3+ FASTQ).

However, if you have only ASCII values of 64 or higher, you can't be 100% sure of the version - but it won't make that much difference since Solexa and PHRED scores are equal above quality ten (after rounding).

You should also be able to spot very recent Illumina 1.5+ FASTQ files by the funny quality distribution. They use 'B' (PHRED 2) as a special marker, an apparently don't use PHRED 0, 1, or 3 anymore. See:

http://news.open-bio.org/news/2010/04/illumina-q2-trim-fastq/

http://seqanswers.com/forums/showpost.php?p=17491&postcount=3

[fkrueger]

A very quick and easy way is to run everything through fastqc, it will mention at the top which kind of fastQ format you are looking at.

[mountainogre]

Thanks for all the helpful suggestions.