[Genome] search for il10 yields unexpected/wrong result

[Ross Lazarus]

Whew. Thanks for the excellent explanation and quick response.
I didn't know any STS markers had names like that....

Think I might turn of the statMapALiasMrnaAcc short circuit and see what breaks since IL10 
is one of those genes I use for example searches :)


Ann Zweig wrote:
> Hello Ross,
> 
>     I asked one of our developers to determine what was going on here 
> and she provided this excellent tutorial:
> 
> For performance, searches are lumped into two categories:
> 
> short-circuit: fast, exact matches of accessions -- quit if and when
> something matches.
> 
> other: slow, exhaustive search -- if something matches, add it to the
> list and keep searching.
> 
> Successful short-circuit searches usually take a small fraction of a
> second, while exhaustive search (or unsuccessful search) can take
> several seconds.
> 
> Here are the searches performed in this case -- there are lots of other 
> possible short-circuit searches, but many are skipped because IL10 
> doesn't match their
> termRegex:
> 
> SHORT-CIRCUIT clonePosNoVers
> SHORT-CIRCUIT mrnaAcc
> SHORT-CIRCUIT stsMapAliasMrnaAcc
> SHORT-CIRCUIT switchDbTss
> SHORT-CIRCUIT wgRnaExact
> 
> If there is no short-circuit match, then it proceeds to the usuals:
> 
> ADDITIVE knownGene
> ADDITIVE ccdsGene
> ADDITIVE gencodeKnown2
> ADDITIVE refLink
> ADDITIVE xenoRefGene
> ADDITIVE transMapRefGene
> ....
> 
> The knownGene search is additive because often when there's a
> knownGene match, it may have relevant matches in other tracks too.
> 
> In making stsMapAliasMrnaAcc short-circuit, there's an assumption that
> if something matches an STS marker name, it couldn't possibly be
> anything else.  Most of the time that's fine... but here's a case
> where it's not.
> 
> ----------
> 
>     So, in the case of the IL10 known gene, it's finding a match in the 
> stsMapAliasMrnaAcc table and stopping the search.  If you don't want 
> this to happen, you can change it by removing the "shortCircuit 1" line 
> from the
> trackDb.ra spec for stsMapAliasMrnaAcc.  However, there will be an 
> impact on other searches, as explained above.
> 
> 
> Regards,
> 
> ----------
> Ann Zweig
> UCSC Genome Bioinformatics Group
> http://genome.ucsc.edu
> 
> 
> 
> 
> 
> Ross Lazarus wrote:
>> I have a local mirror, and I noticed that if I use il10 (IL10 at 
>> chr1:205007571-205012462) as the search term in the 'position or 
>> search term', the browser goes straight to chr7:34,489,430-34,689,640
>>
>> Worried that my MySQL tables were broken, I tried it on the main site 
>> - same result. This is odd since IL10 is an official hugo symbol - if 
>> I do a search for (eg) il1 then il10 does appear - as a ucsc gene 
>> (IL10 (uc001hen.1) at chr1:205007571-205012462) and in the list of 
>> refseq genes.
>>
>> I know that the search function is complicated (!), but is this expected?
>>

-- 
Ross Lazarus MBBS MPH, Director of Bioinformatics
Channing Laboratory, 181 Longwood Ave., Boston MA 02115, USA.
Voice: +617 525 2730  Fax: +617 525 0958