As a ClinVar user, what if you could see the position of a variant of interest highlighted in a 3-dimensional view of the variant's protein structure?
One of the biggest challenges facing researchers and clinicians in medical genomics and personalized medicine is the interpretation of variants. Efforts to prioritize variants for interpretation or further study are often restricted to: allele frequency; the biochemical nature of an amino acid substitution (for example); or one of several available bioinformatic estimates of evolutionary conservation.
However, every protein-altering variant exerts its effect in the biological context of a 3-dimensional protein. iCn3D is a Web-GL-based 3D protein structure viewer which allows users to interact with the protein sequence and structure, highlighting desired amino acids so that their influence on the protein within the 3D structure can be better understood. iCn3D's usefulness has so far been limited, however, by the lack of annotations for existing variants. In this hackthon we leveraged a cgi to retrieve variant annotations from databases such as dbSNP and ClinVar, among others, which has enabled us to project these annotations directly on the protein structure using iCn3Ds interactive 3D view.
This improved version of iCn3D can be implemented as a widget and embedded directly in dbSNP and ClinVar webpages, offering users a new, intuitive level of insight into the 3-dimensional biological context of any variant of interest for which a 3D protein model exists.
iCn3D is an interactive WebGL-based 3D protein structure viewer that displays a given protein in 1D (amino acid sequence), 2D (interaction diagram), and 3D (rotating 3D model) simultaneously. Users can interact with a protein by, for example, highlighting several amino acid residues in the sequence view; the corresponding residues in the remaining views are automatically highlighted. Our goal in this hackathon is to enhance the functionality of iCn3D by importing annotations from other rich sources of information about proteins, such as dbSNP rs numbers, ClinVar assertions of pathogenicity, and known structural features such as ligand binding pockets, protein interaction surfaces, and other functional domains.
In order for annotations to be viewed in iCn3D they must be retrieved from databases (e.g., dbSNP and ClinVar) using Ajax calls.
Without the aid of a 3D protein structure viewer like iCn3D, researchers and clinicians may be limited in their attempts to evaluate a variant's role in causing disease by the information at their disposal, such as whether a particular amino acid change is a conservative one, or on various estimates of evolutionary conservation. If they could instead view variants in the context of a protein's biological 3D structure, they might gain a deeper understanding of how a variant is likely to affect the protein, and thus better understand the mechanism by which it may cause disease. This could in turn lead to an ability to make more informed choices in asking research questions, and even in making more appropriate treatment choices for patients.
dbSNP developers provided us access to a cgi that retrieves ClinVar and dbSNP data for annotations.
A test page can be viewed here. Novel annotations resulting from the achievements of this hackathon are shown at the bottom of the page. Users can highlight desired SNP residues or domains on the 3D protein structure simply by selecting the corresponding amino acids in the sequence pane.