| |||||
 |
|
|
|||
|
|
Why does my search timeout on the BLAST ... To see other biotech frequently asked questions,
please visit http://biotech.fyicenter.com/faq/
(Continued from previous question...) Why does my search timeout on the BLAST servers? Certain combinations of BLAST searches with large sequences against large databases can cause the BLAST servers to timeout. This has to do with a limit on the server CPU's which prevents sequences which generate many HSPs from hoarding server resources. However there are some things you can do to prevent timeout and generate results from large sequences. - Some sequences contain large regions of ALU repeats. In this case you can select the "Human Repeat" filtering option on the main BLAST search page. This will mask repeat regions which generate a large number of biologically uninteresting hits to the databases. - Increase the Word Size to 20 - 25. With a default Word Size of 7, the BLAST algorithm finds initial HSPs of 7 bases in length and begins extension of these from either end. In a large sequence this can generate 100's of initial HSPs between the query sequence and even a single large genomic sequence in the databases. Increasing the Word Size to 25 makes the initial HSP smaller, limiting the number small initial fragments to be extended. - Decrease the Expect value to 1.0 or lower. Many hits from large sequences are to many small fragments in the database. The expect value for these searches is such that decreasing the expect value will eliminate these results, and concentrate on results which are more likely to contain large coding regions and genomic fragments. If you are still seeing a "timeout" error message after making the above changes, please contact blast-help@ncbi.nlm.nih.gov with the RID of your search. (Continued on next question...)
Other Frequently Asked Questions
|
||||
|
|
|
|||
