Mapping PDB structures to gene-level in genome browsers, and visualizing SNPs
Adding aligned PDB data as tracks in such a visualization would help reveal relationships between protein structures and SNPs, including small variants of clinical interest.
One potentially popular use case is co-display of these tracks with protein domains from the Conserved Domain Database (CDD). This can be done on full-featured NCBI genome browsers (Genome Data Viewer, Variation Viewer, 1000 Genomes Viewer, etc.)
This can be implemented by:
- opening the tracks menu in any fully featured sequence viewer screenshot
- in the genes, select the current annotation and choose 'select all from the dropdown' screenshot
- below that, check the show all labels box screenshot
- click configure
Experimental support for remote tracks
In addition to aligning PDB files to the genome, this hackathon also resulted in a proof of concept for loading track data in NCBI genome browsers from any third-party web server. This enables users to load independently-hosted track data without the need for an intermediate host between the genome browser web application and the origin track host, i.e. without the need for a centralized track "hub".
Such functionality could be useful for those seeking to display modestly-sized (< 50 MB) sets of track data served from simple, external web hosts. Underlying data in this proof of concept is hosted on GitHub Pages (e.g. https://ncbi-hackathons.github.io/PDB2Genome/bedjs/chr2.bed.js), which provides static, CORS-enabled, secure web hosting for free.
To reproduce these results:
Open Chrome Developer Tools (Cmd-Alt-I)
Go to DevTools "Console" tab
Paste the following into DevTools console:
See how the PDB structure track(s) have been added to bottom of the Sequence Viewer track list
Files loaded to NCBI Genome Data Viewer can be recapitulated for other genome browsers by running the scripts in this repo against the parquet files found at dataframes.rcsb.org.