This project will map PDB accessions onto human genome viewers. The eventual goal is to make the accession and features selectable and viewable in a pop-up iCN3D window.
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
scripts
src
LICENSE
README
README.md
Screen Shot 2016-07-09 at 11.03.37 PM.png
Screen Shot 2016-07-10 at 12.11.33 AM.png
Screen Shot 2016-07-10 at 12.22.14 AM.png

README.md

PDB2Genome

Mapping PDB structures to gene-level in genome browsers, and visualizing SNPs

Background

Genome browsers like NCBI Genome Data Viewer enable users to view tracks of genomic data, e.g. for gene KIF1A:

https://www.ncbi.nlm.nih.gov/genome/gdv/?context=genome&acc=GCF_000001405.33&q=KIF1A

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.

Use cases

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

Results

##Results 1

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:

  1. Go to https://www.ncbi.nlm.nih.gov/genome/gdv/?context=genome&acc=GCF_000001405.33&q=KIF1A

  2. Open Chrome Developer Tools (Cmd-Alt-I)

  3. Go to DevTools "Console" tab

  4. Paste the following into DevTools console:

    jQuery.getScript("https://ncbi-hackathons.github.io/PDB2Genome/remote-tracks-experiment.js");

  5. Press Enter

  6. See how the PDB structure track(s) have been added to bottom of the Sequence Viewer track list

Results 3

Future work

The PDB accessions displayed on the genome will be linked to structures displayed by iCn3D after streaming from the PDB in compressed format using MMTF.

Data

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.