Tuning local blat to perform more like UCSC's version

Question

[dchen71]

I am trying to find where the 20bp probes I have, lie on the mouse genome. I am trying to tune the parameters to try to match the results that I get when I search it on UCSC. I am hoping someone has some hints into tuning blat.

For reference I am using the following commands for my current results:

gfServer start -tileSize=11 -minMatch=1 -maxGap=2 localhost 3000 Index/mm10.2bit
gfClient localhost 3000 ../mouse Output/test.fa Output/test.psl -t=dna -q=dna -minScore=10

Answer 1

[mastal511]

Have you seen

Replicating web-based Blat parameters in command-line version

on the UCSC FAQ pages?

[dchen71]

I missed that topic but unfortunately, it still gives me a slightly different answer. Hadn't though of trying to use -minIdentity so useful all the same.