different Illumina convention in fastq files?

[mchaisso]

We have a few different Illumina data sets here, and I'm wondering if there has been a change in the characters used in fastq files, or that we have just been given bad data.

we have some sequences for which the translation works fine:
10*log(1+10**(c - 64/10.0))/log(10)

@SLXA-EAS1_89:3:1:715:750/1
GTCTTGAAAGCTATGATGTCAAGATTAATTTAATC
+SLXA-EAS1_89:3:1:715:750/1
bbbbbbbbbbbbbbbbb^bbbbbbbbbbbbbbbba
@SLXA-EAS1_89:3:1:747:756/1
GTGTATTGCTCAATCTTCGAACGGGGGGAGGATTG
+SLXA-EAS1_89:3:1:747:756/1
bbbbb^bbbbbbbbbbbbbbbbbbbbbb\bbObbb
@SLXA-EAS1_89:3:1:859:343/1
GTTAATAGATTTAATTGCCACCGCAATACCAGCAA
+SLXA-EAS1_89:3:1:859:343/1
ccccccccccccccccbcccccccbbbbbbbYbb^

However we have other sequences where the quality is 10 or below using the same translation. Is one using Sanger center convention and the other Illumina?

@ILunknown_unknown_1_1_107_394
TATTCCCCGCCCCCCCGTCGTGCCCGGTCTTTGTCC
+
IIIIIIGI/<I8IIIIII.II//I);II)%<>III&
@ILunknown_unknown_1_1_118_376
TTGGGAAGCGCACCCGGCCCGTGTTGGCTTTCGCCT
+
I%I!I(,@+I"%",!#&)"5$'&$&%$.*%$"#"""
@ILunknown_unknown_1_1_342_174
GCGGCGGGTTATCGGGGCACTCCCCCCCTCCCGCAC
+
IIIIII2I<#?II+@CI1/CI0III5%(5'I%0-%$
@ILunknown_unknown_1_1_121_440
TGGCGTCCAGGCCGGCCTGGCTGCCGTCCGGCCCCC
+
II50I5I.&-*+9-&>2$&&3#&#"#%%"&"#&#&&
@ILunknown_unknown_1_1_109_530
TTTGCGTTAGGTAAAATGCTAGAAGCAGGTGAGCAG
+
IIIIIIIIIIIIIIIIIIIIIII&II0IIII%II0A

thanks,
-mark

[acnoll]

The top is illumina/solexa format (what you would see in s_[1-8]_sequence.txt) and the bottom is the fastq Sanger format.

To convert solexa values
perl -e 'my $qual = 10 * log(1 + 10 ** ((ord($myChar) - 64) / 10.0))/log(10); print int($qual);'

To convert fastq values
perl -e 'my $qual = ord($myChar) -33; print $qual;'

hope that helps