Sorry, I did not copy anyone else. Gary Begin forwarded message:
From: Gary Olsen <[email protected]> Subject: Re: [Microbe] [Kbase-devel] ID mapping in KBase Date: October 25, 2012 23:29:05 CDT To: Robert Olson <[email protected]>
kb|g.1870 Escherichia coli str. K-12 substr. MG1655
Gary
On Oct 25, 2012, at 23:03 , Robert Olson wrote:
Which KB genome(s) should these map to?
----- Original Message -----
Hi all,
From Ecogene is possible to download a file with every possible E.coli aliases/IDs http://www.ecogene.org/ecodownload/crossref
How difficult is to get this in?
José
On Tue, Oct 23, 2012 at 5:10 AM, Robert Olson < [email protected] > wrote:
If there are IDs missing, and we have a reliable location for the definition of the IDs, we can look into the right way to get them loaded. As I pointed out in the original note, the CDM has two primary mechanisms for these mappings, and Tom's design for the translation service is probably the place where solutions other than the CDM ones would be based.
--bob
----- Original Message -----
Really nice work Bob, and great explanation.
So I'd like the science users to kick the tires on this. Sam, does it have the plant gene IDs we need? Janaka, does it have gene IDs you need for all your template genomes used for your core model work? Jose, does it have blatner numbers? bsu loci for B. subtilis? Ben Bowen, does it have your shewenella genes? Gavin, how about all the gene IDs in the datasets you're loading? Liz, I know you in particular are an alias finatic. Do you see missing content?
Someone should probably make a tutorial on ID querying?? The text Bob has below is pretty darn close.
Bob, if we discover that there are important IDs we're missing, how do we go about getting them in? Do we do something crazy like setting up a git repo where poeple check in new ID sets for loading?? Or is there a more sane solution?
On Oct 22, 2012, at 3:48 PM, Robert Olson wrote:
Since these questions come up a lot, and since we now have some additional data loaded, I wanted to give some examples of doing ID lookups using the KBase Central Store.
There are two forms of external IDs loaded into the CDM.
The first is a set of aliases that may be associated with features. These are data like gene names and locus tags that for instance come along with the Genbank files that originally sourced the genomes.
These may be searched with the new script (and API function) aliases_to_fids:
$ echo b0010 | aliases_to_fids b0010 kb|g.1870.peg.4458 $ echo dnaK | aliases_to_fids | head dnaK kb|g.5.peg.1109 dnaK kb|g.11.peg.1619 dnaK kb|g.22.peg.3037 dnaK kb|g.27.peg.3820 dnaK kb|g.34.peg.1696 dnaK kb|g.35.peg.966 dnaK kb|g.37.peg.1437 dnaK kb|g.39.peg.2253 dnaK kb|g.39.peg.2487 dnaK kb|g.42.peg.766 $ echo dnaK | aliases_to_fids | wc -l 1658
(this is also available in Iris).
The second source of external IDs is the external protein database data that are loaded in the HasAssertedFunctionFrom relationship. The relevant part of the schema is here:
<PastedGraphic-1.png>
What this tells us is if you have a feature, you can find the protein sequence for that feature, and map it to the set of assertions from some external database:
$ echo 'kb|g.3785.peg.1578' | get_relationship_Produces -rel to_link | get_relationship_HasAssertedFunctionFrom -rel function,external_id,to_link kb|g.3785.peg.1578 feb01774b3c5bb6e84ca9ae888514c5f multidrug resistance protein A ref|ZP_05465429.1 NCBI kb|g.3785.peg.1578 feb01774b3c5bb6e84ca9ae888514c5f multidrug resistance protein A gb|EEZ16943.1 NCBI kb|g.3785.peg.1578 feb01774b3c5bb6e84ca9ae888514c5f secretion protein HlyD family protein gb|ADZ67367.1 NCBI kb|g.3785.peg.1578 feb01774b3c5bb6e84ca9ae888514c5f secretion protein HlyD family protein gb|ADZ88235.1 NCBI kb|g.3785.peg.1578 feb01774b3c5bb6e84ca9ae888514c5f Multidrug resistance protein gb|AEQ09823.1 NCBI kb|g.3785.peg.1578 feb01774b3c5bb6e84ca9ae888514c5f Multidrug resistance protein A tr|D0GC62|D0GC62_BRUML TrEMBL kb|g.3785.peg.1578 feb01774b3c5bb6e84ca9ae888514c5f Multidrug resistance protein tr|G4PJZ8|G4PJZ8_BRUML TrEMBL kb|g.3785.peg.1578 feb01774b3c5bb6e84ca9ae888514c5f Secretion protein HlyD family protein tr|F2GWN2|F2GWN2_BRUM5 TrEMBL kb|g.3785.peg.1578 feb01774b3c5bb6e84ca9ae888514c5f Secretion protein HlyD family protein tr|F2HWT4|F2HWT4_BRUMM TrEMBL
Since going in the other direction requires querying on a relationship field, and we may want to do prefix matches there, you will need the new external_ids_to_fids script (or API routine):
$ echo 'gb|ADZ88235.1' | external_ids_to_fids gb|ADZ88235.1 kb|g.1940.peg.347 gb|ADZ88235.1 kb|g.2192.peg.1337 gb|ADZ88235.1 kb|g.29.peg.2088 gb|ADZ88235.1 kb|g.30.peg.1725 gb|ADZ88235.1 kb|g.31.peg.2171 gb|ADZ88235.1 kb|g.3785.peg.1578
Or for a prefix match:
$ echo 'gb|ADZ88235' | external_ids_to_fids -prefix gb|ADZ88235 kb|g.1940.peg.347 gb|ADZ88235 kb|g.2192.peg.1337 gb|ADZ88235 kb|g.29.peg.2088 gb|ADZ88235 kb|g.30.peg.1725 gb|ADZ88235 kb|g.31.peg.2171 gb|ADZ88235 kb|g.3785.peg.1578
--bob
_______________________________________________ Kbase-devel mailing list [email protected] https://lists.kbase.us/mailman/listinfo/kbase-devel
Microbe mailing list [email protected] https://lists.kbase.us/mailman/listinfo/microbe