[maq-help] Sol2sanger giving higher error rate

[Girish B <bg...@he...>]

Hi !

I am currently aligning some illumina data against the human genome using maq 0.6.8 . The illumina runs are from the 1000 genomes project. They are given as paired ends ( ftp://ftp.ncbi.nih.gov/pub/TraceDB/ShortRead/SRA000271/fastq/ )and I wrote a  code in python to only pick 50,000 of the reads of each paired end. I first aligned without using sol2sanger and my error rate was 0.044 ( average over 7 lanes) . I then did the same thing , however this time, I used sol2sanger on my fastq files and aligned and the error rate was 0.015 ( Average over 7 lanes ).

For the original solexa file :
http://paste.pocoo.org/show/83498/

and for after converting to sanger fastq using sol2sanger :
http://paste.pocoo.org/show/83499/


I would have thought that the error rate for after conversion to sanger fastq would give me a lower error rate but instead it is the opposite. And 4% error rate is fairly "big" . I did go through the usual procedure as given by maq user ref. Can anyone advise?


Regards,

Girish B