Hi,

I have a question for you guys regarding quality filtering.

I have a data set of double MID tagged 454 amplicons, from which I wish to select high quality sequences above Q20. The 454 quality filtering system seems to work differently from that given for the Illumina sequencing i.e. 454 filtering takes high quality segments, while Illumina (FASTQ) can select high quality full reads based on certain parameters.

OK, so I know that the total length of my amplicon, including primers and barcodes is around 260bp. If I then set the 454 quality filtering tool to extract contiguous high quality sequence of >260, it gives me back around 45% of my raw data as hitting this criterion i.e. All 260bp are above Q20. I don¹t necessarily need this high stringency as most bases may not be informative.

But if I convert my 454 data to FASTQ format and then run the Illumina filtering system which also allows me to set the number of bases allowed to deviate from the Q20 criteria, I get back over 90% of my data (allowing 10bp to deviate from Q20).

I then need to go ahead and convert back to 454 format.

Can you tell me if this is OK? Will I loose /confuse information somewhere along these conversions?

It seems that if I do this, my barcodes are removed, as amplicons do not sort properly when I parse them through my barcode filtering program.

Does anyone know of a program to filter 454 data based on average sequence quality score, which doesn¹t involve Linux and the Roche off instrument program (I have no experience in Linux! )

Thanks!