phrap quality scores

Todd Scheetz todd-scheetz at uiowa.edu
Sun Jan 9 14:07:22 EST 2000


Rachel,

I'm assuming you've got the phred generated sequence and quality files.
>From the limited testing I've done recently, with limited amounts of
overlapping sequence to keep the complexity down, here's the way it
appears to work.

1. If sequence in only one direction, the highest/best score is used.
2. If sequence in both directions, the best qualities from each
direction are added together.

However, this comes with a few caveats. Phrap brings the sequences
together in segments, so the highest quality value mention above is
really the quality value of that base within the segment chosen as the
representative. Also, I haven't quite figured out the quality given to
discordant bases. I haven't bothered digging into the source yet,
though. If I find out anything more, I'll let you know (if you'd like).

-Todd

"Rachel M. Kadel-Garcia" wrote:
> 
> In article <387698C5.EEA17FB at cellworks.mbt.washington.edu>,
> Stephen R. Lasky wrote:
> >"Rachel M. Kadel-Garcia" wrote:
> >>
> >> Does anyone here have an understanding of phrap quality scores
> >> beyond what's in the docs?  I'm trying to figure out how to
> >> assign similar quality scores to assemblies not done with phrap,
> >> and I'm resorting to spelunking in the source, but it's rough
> >> going.
> >>
> >>                                                 Rachel
> >
> >
> >Phrap, I believe, is not the program assigning the quality scores.
> >Phred is.  Phrap uses the phred quality scores to do its alignments.
> 
> Phred assigns quality scores to bases in individual reads.  Phrap uses
> those quality scores and various things about the alignments to generate
> consensus quality scores.
> 
> >Do you have the Phred docs?
> >
> >I presume that you don't have the chromats, otherwise you could run them
> >through phred and get the quality scores.
> 
> I have the phred docs and the chromat files, but that won't get me
> consensus quality scores.
> 
>                                                 Rachel





More information about the Bio-soft mailing list