Hi Jamie, Thank you for writing back and clarifying the situation. We suspected that your original question was not completely addressed!
The short answer is that the version of Blat we provide as a download is different than the browser version (hgBlat). Output features and options differ. When or if these two will be in sync in the future is not decided. There is also a webBlat submission tool, but as you have noticed it has a limit. It also has a two known bugs that are undergoing revision. 1) The first bug is that all the coordinates on the 2nd part of the display are off by two lines (2*60=120). 2) The second bug is that there is a problem with negative strands. Details are not clarified at this time. There are some ideas to create a new file output format that would provide the type of alignment formatting you have so clearly described. But when this project would be started/completed has yet to be decided. I hope this helps you to understand the results you were getting and to know that our developers agree that there is no work-around (except for developing your own cgi/html). Thank you for your patience, Jennifer Jackson UCSC Genome Bioinformatics Group Jamie Teer wrote: > Sorry, my initial request was a bit muddled! I generally use the UCSC BLAT > server to realign a small number of next-gen sequencing short reads. These > reads overlap, and any mismatches are displayed by the UCSC browser as > single white letters in a black bar representing the read. A heterozygous > genotype in the input reads is easily apparent as a "mismatch" in about half > the reads. Combined with a vaiety of other tracks in the browser, this > provides a very nice view of the data at a given locus. > > Due to the limit on 25 reads, this approach is often not very useful. I > have been running BLAT locally on hundreds on short reads, and am now trying > to upload this BLAT output to the UCSC genome browser, and have it display > the same as output from the UCSC BLAT server is displayed. The best I can > do is use a modified MAF file, in which mismatches are shown as a letter, > and bases the same as reference are displayed as a dot. While this works, > the output would be much more visually appealing if I could adjust some > settings to display the reads as black bars, with white letters illustrating > the mismatches. I have looked at the UCSC browser source code (thanks for > making it available!!), but I can't figure out a way to use your settings > for a custom track. Maybe it isn't possible, but any insight would be > appreciated! I've looked at the BLAT and custom track FAQs, and they > weren't able to help. > > Thanks for your help! > Jamie > > > On 2/20/09 6:45 PM, "Jennifer Jackson" <[email protected]> wrote: > > >> Hello Jamie, >> The output on our server is a combination of file format input and the >> cgi./html that creates the browser page. >> >> There may be something in the Blat FAQ to help? >> http://genome.ucsc.edu/FAQ/FAQblat.html >> >> Or better yet, the web Blat instructions on this page: >> http://genome.ucsc.edu/goldenPath/help/blatSpec.html >> >> If you still need technical help after ready about the web Blat, please >> write back and let us know, >> Jennifer Jackson >> UCSC Genome Bioinformatics Group >> >> Jamie Teer wrote: >> >>> I am trying to mimic the visual output of BLAT output from the UCSC BLAT >>> server. Specifically, I want to see black bars, with letter highlighting >>> variations in the sequence (what one would see after clicking on "browser" >>> in the BLAT output window.) I am trying to run BLAT locally, to get past >>> the 25 sequence limitation. I have searched the mailing list, and have >>> tried the pslPretty -axt | axtToMaf scripts, but do not get the same black >>> bar output with the differing bases being the only visible bases. I feel >>> that I might be misunderstanding the final modification needed in the maf >>> file itself. Can you go through the process in a bit more detail? Thanks! >>> Jamie >>> >>> _______________________________________________ >>> Genome maillist - [email protected] >>> http://www.soe.ucsc.edu/mailman/listinfo/genome >>> >>> > > _______________________________________________ Genome maillist - [email protected] http://www.soe.ucsc.edu/mailman/listinfo/genome
