Flybase BLAST vs. UCSC BLAT

Ask questions about sequence improvement / finishing D. mojavensis projects here.
Post Reply
cjones
Posts: 99
Joined: Sun Feb 04, 2007 10:19 pm
Location: Moravian College, Bethlehem PA
Contact:

Flybase BLAST vs. UCSC BLAT

Post by cjones » Tue Feb 12, 2013 2:01 am

One student pulled two mate pair reads and searched in both BLAT and BLAST to determine if the corresponding subclone belongs in his project or not. FlyBase BLAST placed both reads in the 3L scaffold we're using (#13337), but the BLAT engine on gander gave only a different fosmid for one read and no hits at all for the other. That different fosmid *does* overlap his, according to the scaffold map, but why doesn't his project merit a hit? And why does the other end not match anything according to BLAT?

His project is screwy for other reasons, perhaps related: every time he's opened it, he's gotten a warning from Consed: http://home.moravian.edu/users/bio/mecj ... s/adam.png -- is this reparable? He's been working on his project, apparently without (undue) problems, but I can't recall seeing this message before.
Chris Jones
Assoc. Prof. of Biology
Moravian College
Bethlehem PA

wleung
Posts: 182
Joined: Sun Feb 04, 2007 7:41 pm
Location: Washington University in St. Louis

Re: Flybase BLAST vs. UCSC BLAT

Post by wleung » Tue Feb 12, 2013 5:19 am

There are a few possible reasons why the fosmid end reads cannot be mapped by BLAT. First, the D. ananassae genome assembly on the gander server is incomplete. Consequently, you should use the BLAT service at the official UCSC Genome Browser when performing BLAT searches against the whole genome assembly. The BLAT service on the gander server is setup only for the set of fosmids that have been improved.

Second, BLAT is less sensitive than BLAST (e.g. uses a larger word size). If the reads are low quality, BLAST might be able to detect regions of similarity that is undetectable by BLAT.

Note that because of the large number of misassemblies in the D. ananassae assembly, the fosmids track on the genome browser only shows the approximate placement of the fosmids.


> His project is screwy for other reasons, perhaps related: every time he's opened it, he's gotten a warning from Consed:
> http://home.moravian.edu/users/bio/mecj ... s/adam.png

For some of the projects that have been worked on by GEP students in previous years and a large number of sequencing reactions have been called, a message might appear indicating that there are a large number of reads that do not follow the Consed read name convention. You can ignore the warning because the custom reactions use a different naming convention than the original set of reads from the whole genome assembly.
Last edited by wleung on Tue Feb 12, 2013 11:36 pm, edited 1 time in total.

cshaffer
Posts: 211
Joined: Sun Feb 04, 2007 10:29 pm
Location: Washington University in St Louis
Contact:

Re: Flybase BLAST vs. UCSC BLAT

Post by cshaffer » Tue Feb 12, 2013 11:07 pm

That message is consed warning the user that when consed tried to recognize which reads were of which type it looked like there are WAY to many "walking reads" (i.e. reads that come from a unique primer. Consed is giving the warning becuase it is probably more likely that there is an error in the names of the files and consed is making errors when it is parsing the reads. However, since we run 3 reactions per primer and some of the very difficult regions have had multiple primers from multiple labs it is possible that there indeed 164 walking reactions in the project.

You will need to actually look at the read list in the main window to determine if the consed warning is a false alarm (i.e. sure enough you see 164 or so reads with the names like: "XXXXX12XBAD-1036A03_1.b1".) or if there is some kind of parsing error (i.e. almost all reads are from the whole genome project and have names like: "03664275L07.b1")

If you believe there is a parsing error get in touch as there is something wrong with the installation on that particular computer.

image now available and is attached
Attachments
adam.png
adam.png (9.08 KiB) Viewed 2858 times

Post Reply