G2S (Genome to Structure)

An auto-updating web-service to annotate genomic variants on protein structures

There is a great interest and need of mapping and analyzing germline and somatic genomic variants generated by recent large-scale sequencing efforts in the context of 3D protein structures. Several tools have been developed to address the need, including SIFTS (https://www.ebi.ac.uk/pdbe/docs/sifts/) , a service by PDB (http://www.rcsb.org/pdb/chromosome.do) , and G23D (http://www.sheba-cancer.org.il/G23D). However, none of them provides a web service that support automated analysis programmatically.

We have developed a comprehensive web service that support automated mapping and annotating genomic variants in 3D protein structures,

There are two components in G2S web services:

• G2S API (link): Query by Uniprot/Human Ensembl/Genomic Position.
• Sequence Web Interface (link) : Query by any protein sequences.

G2S web service supports:

• Retrieving aligned protein structure chains for an Ensembl/UniProt entry.
• Retrieving aligned protein structure chains for any protein sequence.
• Retrieving residue mapping between protein sequences and protein structures.
• Retrieving structural position of a genomic variant.

The web service will return high confidence pre-calculated mapping of the protein/residue aligned to the protein structures from the PDB in the JSON format. A fixed e-value score threshold are predefined with several alignment quality metrics to enable further filtering if desired. Our system makes frequent (weekly) updates to the set of pre-computed alignments as new pdb structures are added to RCSB, while keeping the set of protein sequences from public protein sequence databases fixed. Less frequently, the system is re-initialized using an updated set of protein sequences, and during re-initialization an all-against-all alignment process is required. This web service was initially designed and implemented for mapping and visualizing cancer mutations in the cBioPortal, a leading resource for studying cancer genomics (http://www.cbioportal.org/). The cBioPortal has been used by thousands of researchers (currently there are in average about 3,000 visit daily), mostly individuals outside of the authors’ group. Recently, we separated and generalized the web service from the cBioPortal and established the G2S web service in the project Genome Nexus for annotating genomic variants. The web service will continue serving the cBioPortal as well as any other applications. The codes of the web service and website are open-source and available at GitHub.

History:

2019-02-12 - Fix a bug of sequence id mixmatch in residue mapping.

2018-12-19 - Fix a bug of alignment mismatch in residue mapping.

2018-10-29 - Fix minor bug API 500 error when upstream API return error.

2018-10-05 - Add support of ENST/ENSG mapping in all versions.

2018-10-04 - Add support of mapping dbSNP ID to PDB.

2018-06-30 - Fix small bug of mapping hgvs-grch37 genes.

2017-11-30 - Initial Version released.

Citation:

G2S: A web-service for annotating genomic variants on 3D protein structures, Juexin Wang, Robert Sheridan, S Onur Sumer, Nikolaus Schultz, Dong Xu, Jianjiong Gao, Bioinformatics, 34(11), 1949-1950, https://doi.org/10.1093/bioinformatics/bty047