Cancer3D: understanding cancer mutations through protein structures.
Bottom Line: This approach allows users to find novel candidate driver regions or drug biomarkers that cannot be found when similar analyses are done on the whole-gene level.In addition, it displays mutations from over 14,700 proteins mapped to more than 24,300 structures from PDB.This helps users visualize the distribution of mutations and identify novel three-dimensional patterns in their distribution.
Affiliation: Bioinformatics and Systems Biology Program, Sanford-Burnham Medical Research Institute, 10901 North Torrey Pines Road, La Jolla, CA 92037, USA.Show MeSH
Related in: MedlinePlus
Mentions: Cancer3D integrates data from TCGA and CCLE and allows users to explore the biomarker and driver problems at the same time through two novel algorithms: e-Driver (13) and e-Drug (15). These algorithms are unique in using information about the modular structure of a protein to predict novel cancer drivers or drug biomarkers, respectively. Statistics are calculated separately for each region in each protein, including known PFAM domains (16), predicted intrinsically disordered regions and over 1300 potential novel domains in the human proteome detected by AIDA (17). Another important feature of Cancer3D is that it maps somatic missense mutations from over 18 000 human proteins to, wherever available, experimental or predicted protein three-dimensional structures. The Cancer3D database not only displays the mutated positions of a protein in their corresponding structures, but also conveys important information such as drug activity or mutation frequency by color-coding. This helps users to quickly identify three-dimensional patterns in data such as clustering of mutation hotspots around particular regions (Figure 1).
Affiliation: Bioinformatics and Systems Biology Program, Sanford-Burnham Medical Research Institute, 10901 North Torrey Pines Road, La Jolla, CA 92037, USA.