[BiO BB] Getting better assemblies from phrap

Amir Karger akarger at CGR.Harvard.edu
Wed Sep 5 13:23:48 EDT 2007

I've got a client who did a bunch of shotgun sequencing of a 100kb BAC
insert, and is now trying  to build an assembly with phredPhrap. He got
350 contigs. One was 12k, several were a few kb. We tried running with
-forcelevel 5, but that didn't help at all.


Are there other parameters that magically make phredPhrap work better?
I'm sure it depends on the quality of the reads. Many of them have at
least a region of good quality. There's a very long tail on a lot of
that, but the phred documentation says not to use the -trim option that
would remove those (and indeed phrap builds contigs even though the long
bad-quality tails don't match each other). Is there some way to tell
whether I have good enough quality? Is it not unusual to get just 12k,
and he'll just need to use primers suggested by consed to fill in gaps?


I'd appreciate any suggestions.


-Amir Karger

Research Computing Group

Life Sciences Division

Harvard University

akarger at cgr.harvard.edu

More information about the BBB mailing list