next up previous contents index
Next: 8.3 Escape of Unsafe Up: 8.2 Error Checking Previous: 8.2.2 Errors at the   Contents   Index


8.2.3 Strategy for Coping with CPU and HSP Related Errors

As the database size and the genomic proportion continue to grow, it is more likely that a misconfigured search would encounter CPU and HSP related errors. Here are some general suggestions in coping with such type of errors:

a. Reduce the size of the query by breaking a large query into smaller chunks and search with each smaller chunk indivivually.

b. Search a smaller and well curated database or a more relevant subset of the original database using ENTREZ_QEURY parameter.

c. Increase the search stringency by using smaller EXPECT, larger WORD_SIZE, and more appropriate FILTER settings.

d. Use a more stringent algorithm if applicable.

For identification of candidate genes contained within a nucleotide query sequence, a recommended way to search is to gradually escalate the search intensity. Use megablast first, followed by discontiguous megablast, blastn, and blastx. Proceed to next algorithm if the result from a strigent algorithm is not informative. The tblastx search should be reserved as the last resort and it should always be combined with ENTREZ_QUERY to focus the search to a more relevant subset of the target database.


next up previous contents index
Next: 8.3 Escape of Unsafe Up: 8.2 Error Checking Previous: 8.2.2 Errors at the   Contents   Index
Tao Tao 2007-08-03