how to add Plasmodium to organism choice

[David Hoover]

The new hgGateway has eliminated my custom genome for Plasmodium. What steps do I need to take to restore it to availability? I am running v335.

David Hoover
HPC @ NIH

[Angie Hinrichs]

Hi David,

Sorry to hear that.  If you type the genome name or database name in the box labeled "Enter species or common name", it should auto-suggest the database.  Is it not doing that?  I assume that your custom genome has an entry in your local hgcentral database, correct?  

Angie

--

[David Hoover]

Yes, sort of.  If I type 'pf5', it displays the genome.  If I type 'falciparum' or 'Plasmodium', it shows nothing, even though these strings appear in the entry that appears when I search for pf5.  How does hgGateway.js construct its look ahead database?  Where in hgcentral can I change things?

David

--

---
You received this message because you are subscribed to the Google Groups "UCSC Genome Browser mirror site discussion list" group.
To unsubscribe from this group and stop receiving emails from it, send an email to genome-mirro...@soe.ucsc.edu.

[Angie Hinrichs]

Hi David,

Can you share the hgcentral.dbDb row for pf5?  hgGateway.js sends auto-suggest requests to the server -- hgGateway.c.  The function searchDbDb tries to match your search term to dbDb.name, and -- if dbDb.taxId is a valid taxId number -- dbDb.genome and dbDb.scientificName.  If dbDb.taxId is a placeholder like 0, then only the database name is matched.  

For Plasmodium falciparum, it looks like 5833 is a valid taxId (there are also more specific strains with their own taxIds): http://www.ncbi.nlm.nih.gov/Taxonomy/Browser/wwwtax.cgi?id=5833

Hope that helps,

Angie

[David Hoover]

> select * from dbDb where name = 'pf5'\G

*************************** 1. row ***************************

          name: pf5

   description: Dec. 2007

       nibPath: /data/genome/gbdbLocal/pf5

      organism: Plasmodium falciparum

    defaultPos: chr13:1,635,875-1,639,211

        active: 1

      orderKey: 80

        genome: Plasmodium falciparum

scientificName: Plasmodium falciparum

      htmlPath: /data/genome/gbdbLocal/pf5/html/description.html

      hgNearOk: 0

        hgPbOk: 0

    sourceName: Ares lab

         taxId: 0

1 row in set (0.00 sec)

Setting taxId to 5833 did the trick!  Searching with either 'pf5' or 'Plasmodium' worked.  hgGateway.js is full of taxonomy IDs, so I'll look there in the future if we have other custom genomes.

The next step is to have it display on the taxonomic chart.  I see there is 'Ebola' on the UCSC site, but not on ours.  I guess that would take some magic to add another color and image, unless I classified Plasmodium as either 'Worm' or 'Yeast'.

Thanks,

David

[Angie Hinrichs]

Great, glad that fixed it.  The tree is generated from dbDb and the NCBI Taxonomy Database as described in kent/src/hg/makeDb/doc/dbDbTaxonomy.txt:

http://genome-source.cse.ucsc.edu/gitweb/?p=kent.git;a=blob_plain;f=src/hg/makeDb/doc/dbDbTaxonomy.txt;hb=HEAD

It uses the script kent/src/hg/utils/dbDbTaxonomy.pl.  You'll need to do a couple things differently, i.e. use hgcentral instead of hgcentraltest and instead of overwriting ~/kent/src/hg/js/dbDbTaxonomy.js, create a new file in some directory that your CGIs can access.  

Then add or change this setting in cgi-bin/hg.conf so that the tree is generated using your custom file instead of our js/dbDbTaxonomy.js:

# Display phylogenetic tree in hgGateway (URL or path relative to cgi-bin)

hgGateway.dbDbTaxonomy=/path/to/custom/dbDbTaxonomy.js

That should make Plasmodium appear in the tree (although, in our Human-centric view, pretty far down the list).  

Our colors and icons are pretty much hardcoded in hgGateway.js as you can see.  We contracted a graphic designer to produce the icons and select the colors, and I just coded those up.  It would require non-trivial code changes to add your own coloring rules and icon, and those might be a pain to merge with future changes that we make to hgGateway.js and/or the sprite image.  But if you really want to go for it, I guess you could replace the Ebola icon in the sprite and change Ebola references to Plasmodium references in iconToSpeciesName, spriteRowCol and iconFudgeY.  

Angie

[David Hoover]

One last esoteric question.  Is there any hope of having UCSC support Plasmodium genomes?  I know your bag is mainly vertebrate sequences, but there are a number of researchers here at NIH interested in parasites.

David

[Angie Hinrichs]

Our funding mandate is to focus on human data.  Our genome-preview server happens to have a plasmodium assembly, but that was made as a research project a couple years ago, has only a few tracks, and has not been through our QA process.