Illumina is over-creative with their own
formats.
.seq.txt: base-called sequences
.prb.txt: qualities for each
nucleotide/each cycle (is some kind of semi-raw data?)
Illumina FASTQ is not real FASTQ, but looks
similarly (to cause confusion). The quality score is an own Illumina score, computed by
subtracting 64 from the ASCII code of the character. For example the ‘h’ in the
example is ord(‘h’) – 64 = 104 – 64 = 40.
See http://www.asciitable.com/ for ASCII
codes (Dec is the ‘ord’ in decimal numbers).
The Illumina score can be transformed to
Phred score. Given a character ch,
the following
expression should give the Phred quality q (please double-check:) q = 10 * log(1 + 10 ^ (ord(ch) - 64) / 10.0)) /
log(10)
Qseq: Not looking like FASTQ, but the
scores are proper Phred scores. Their encoding is however still ASCII – 64.
Illumina “single line format” and SCARF:
base-called sequences and quality scores in a bit more handy format.