The part not found is only a perfect match of 14 bases: tagccctagtgtag
With tileSize of 11 and stepSize of 5, the minimum exact match is 2*5 + 11 - 1 = 20. (Even then, over-used tiles may get filtered out.) BLAT may not be a good tool for rna-seq. You can get around the splicing issue somewhat if you build a database of rna instead of using genomic dna. Then the splicing issue goes away. For example, for BLAT hgPcr, we provide the UCSC Genes as an RNA database target option. http://genome.ucsc.edu/cgi-bin/hgPcr Unfortunately this option is not available for hgBlat. However you should be able to construct your own rna database and use it with blat command line tools. -Galt Ar 8/30/2010 4:33 PM, scríobh [email protected]: > Hi again, > I'm having another problem with BLAT: > The following 50bp read maps without mismatches across a splice junction: > TAGCCCTAGTGTAGAGAAATACAGGTATCAGGATGAAGATACACCTCCTC > > However, only the part mapping to one exon maps with BLAT, and the part from > the other exon is not mapping, even though it is a perfect match. > > This is true in both the web-based BLAT and the command-line with the same > parameters. How should I alter BLAT's parameters so it is able to map the > entire length of such a read? > > Thanks. > > > > _______________________________________________ Genome maillist - [email protected] https://lists.soe.ucsc.edu/mailman/listinfo/genome
