One final update: UCSC is looking into the feasibility of putting BLAT into the Cloud, to allow users to purchase compute time as needed. We do not have a schedule yet, but it is definitely something we are looking at seriously to help the users that do not have the resources to purchase and maintain compute hardware. This will not help you right now, but perhaps in the future.
Stay tuned for updates on this subject - if/when we go forward, details will be announced on the UCSC Browser's main home page. Best of luck with your work! Jennifer --------------------------------- Jennifer Jackson UCSC Genome Informatics Group http://genome.ucsc.edu/ On 4/26/10 5:00 PM, Galt Barber wrote: > > Hi, Inbal! > > Sounds like you are an ideal candidate for running your own > blat index, gfServer. Then your cgi could query it with > gfClient to get the alignment results. This would be great > for interactive queries such as your own. > > You would need to find a machine with 4 to 8 GB ram. > This however would give you good interactive performance > without limitations. Good luck with your application. > > The BlatBot.pl program is written to do batch queries. > It might have some rate-limiting code in it to help > keep from annoying our bottleneck processes. > > However, when it is run from a CGI, > and apache starts up a new process every time someone > submits a query, this could easily create a situation > where multiple instances of BlatBot.pl are running > at the same time. > > -Galt > > Ar 4/26/2010 2:42 PM, scríobh Jennifer Jackson: >> Hello Inbal, >> >> When using the BlatBot.pl program, it is important to be certain that it >> is run only once during any time period and that it fully completes >> before launching it again with another set of queries. >> >> http://genomewiki.cse.ucsc.edu/index.php/Blat_Scripts >> >> Your block will decay in about 12 hours, at which time you can start >> again. >> >> If you would like to share details about how you utility SFmap is using >> BlatBot.pl, because you feel that the usage criteria has been met and >> you are still being blocked, we would be interested in learning more. >> You can send the information to me directly and I will forward to the >> appropriate staff here for review and feedback. >> >> Please do not send anything large in an attachment. If you feel like you >> need to share data, please set up an ftp location and send access >> instructions. >> >> Thank you, >> Jennifer >> >> --------------------------------- >> Jennifer Jackson >> UCSC Genome Informatics Group >> http://genome.ucsc.edu/ >> >> On 4/26/10 6:35 AM, Inbal Paz wrote: >>> Hi, >>> >>> My name is Inbal Paz and I'm a programmer at Prof. Yael >>> Mandel-Gutfreund's >>> lab located in the Technion- Israel Institute of Technology. We have >>> recently developed a web-server called 'SFmap' for the mapping and >>> prediction of splicing factors binding sites >>> (http://sfmap.technion.ac.il/). >>> The related paper was accepted by Nucleic Acids Res. and is currently >>> under >>> the final stages of review. >>> >>> SFmap uses BLAT in order to determine the genomic coordinates of the >>> input >>> sequence we get from the user. We run BLAT using the BlatBot.pl >>> script on >>> the UCSC servers in order to get fast results, since our available >>> hardware >>> is not strong enough to run BLAT on whole genomes locally in a >>> reasonable >>> time. >>> >>> Yesterday we tried to run SFmap with a long list of input sequences >>> (for QA >>> purposes) and our web-server's IP was apparently blocked . This is >>> the error >>> message we got from the BLAT output: >>> >>> "There is an exceedingly high volume of traffic coming from your site >>> (IP >>> address 132.68.104.195) as of Mon Apr 26 03:06:05 2010 (California >>> time). >>> It looks like a web robot is launching queries quickly, and not even >>> waiting >>> for the results of one query to finish before launching another >>> query. We >>> cannot service requests from your IP address under these conditions. >>> (code >>> 844360)" >>> >>> We would be very happy to get the IP released again asap (if this is >>> indeed >>> the problem) and to get your advise on which is the best way to avoid >>> similar problems in the future. >>> >>> Thanks a lot in advance, >>> >>> Inbal >>> >>> Inbal Paz >>> Programmer and Webmaster >>> Yael Mandel-Gutfreund's lab >>> Technion - Israel Institute of Technology >>> Tel: +972-4-8293701 >>> >>> >>> >>> >>> >>> >>> >>> _______________________________________________ >>> Genome maillist - [email protected] >>> https://lists.soe.ucsc.edu/mailman/listinfo/genome >> _______________________________________________ >> Genome maillist - [email protected] >> https://lists.soe.ucsc.edu/mailman/listinfo/genome > _______________________________________________ Genome maillist - [email protected] https://lists.soe.ucsc.edu/mailman/listinfo/genome
