Validate gene symbols - Githubissues

iVis-at-Bilkent / pathway-mapper

PathwayMapper: An interactive and collaborative graphical curation tool for cancer pathways

https://www.pathwaymapper.org/

GNU Affero General Public License v3.0

55 stars 27 forks source link

Validate gene symbols #12

Closed ugurdogrusoz closed 7 years ago

ugurdogrusoz commented 8 years ago

We should implement an operation (Edit > Validate Gene Symbols) that somehow checks (BioGene?) whether or not each of the gene symbols is valid and highlights any that is not!

cannin commented 7 years ago

This came up on a call today as being important because an analysis missed results due to mapping issues. One of the CBioPortal people showed me the API they have, but also mentioned it is not completely stable. Niki will likely have more information on this. The alternative is to get symbols directly from HGNC:

http://www.genenames.org/cgi-bin/statistics

They provide a JSON download.

ugurdogrusoz commented 7 years ago

This could go as View > Check Gene Symbols. As a result, genes with invalid symbols would be highlighted (using cytoscape.js-view-utilities extension) with a message on the lower right corner that says either one of the following:

All gene symbols are valid
Highlighted genes have invalid symbols

cannin commented 7 years ago

This should be a good feature. There is some R code for the way I thought about the CBioPortal gene name validation.

https://bitbucket.org/cbio_mskcc/pancanatlas/src/master/checkGeneNames.R

istemi-bahceci commented 7 years ago

The beta API at cBioPortal seems promising:

'http://www.cbioportal.org/beta/api/genes/TP53'

It seems we can validate genes with the above API call in pathway mapper easily.

ugurdogrusoz commented 7 years ago

Done with https://github.com/iVis-at-Bilkent/pathway-mapper/commit/b1636bc9a0ec0bda8d333cbf3585531e9c86cceb