FastX Quality Score format

[kbushley]

I've just been revisiting differences in quality scores between different fastq format (re-reading Cock et. al. 2010 Nucleic Acids Research - what a MESS Illumina has made of the fastq format).

In any case, perhaps I am missing something but am trying to figure out how FastX-Toolkit (I'm using version fastx_toolkit-0.0.13.tar.bz2) is both reading in, scoring, and filtering on quality scores (ie., is it using solexa_fastq, sanger_fastq, illuminav1.3+_fastq quality scores or can it handle all). When plotting out the graph of quality scores, you get a range from -15 to 45 which corresponds to a range of 60 but doesn't seem to correspond to any of these published fastq formats ?*&!. Since this is 2010 and we've been using Illumina 1.3+ for some time now and I've read here on SEQ-ANSWERS that the first two quality scores (0,1) are not included so the range for Illumina 1.3+ quality scores is really 60, I'm assuming that FastX is both reading in and filtering on Illumina pipeline v1.3+ fastq format quality scores. I'd feel better though is someone else could confirm!

Thanks!

[splaisan]

Below the answer I got for a similar problem from Assaf Gordon In my case this was the solution, you may have to change the assci offset accordingly
Cheers,
Stephane

Hi Stéphane,

There is an undocumented argument "-Q" that determines the input quality ASCII offset (33 or 64 or other).

The program (every program in the fastx-toolkit) knows  automatically to detect ASCII or numeric values.
If it sees ASCII characters, it subtract 64 for the ASCII value.

So for example, the character 'a' (ASCII code 97) is converted to value 33 (97-64).

If you need a different ASCII offset (e.g. 33 instead of 64), you can use the (undocumented) argument "-Q", as so:

$ fastx_quality_stats -Q 33 -i INPUT.txt -o OUTPUT.txt

you'll need version 0.0.10 or later for the "-Q" argument.

-Gordon.