Catapult is a computer algorithm for associating new genes with traits, phenotypes, and diseases. Catapult leverages hundreds of thousands of known individual gene-phenotype associations from humans, mice, yeast, Arabidopsis, and C. elegans, and interprets them with the aid of a genome-scale functional network of human genes (HumanNet) in a graph-based formalism in order to rank new candidate genes for diseases/traits of interest. This is the homepage for Catapult. Here we provide:
- The predictions for all the OMIM diseases used.
- A query interface, where you can submit your own gene lists and have the results emailed to you.
- The project source code, if you want to test it all yourself.
- All gene-disease predictions in a 300MB tab-delimited text file (or .zipped 120MB version here). Rows represent human genes, indicating gene names in column 1 as NCBI Entrez genome/Locus link ID numbers. Columns represent human diseases or mouse phenotypes. Entries represent computed scores for gene-trait associations, where larger values indicate more likely associations. All values are scaled, such that sorting by values in a given column allows one to find the most likely genes for that disease, and sorting by values in a given row allows one to find the most likely linked traits for a given gene (obvious caveats being that the diseases/genes must be represented in this table, and the associations are subject to the reach and quality of the predictions captured by CATAPULT.)
Citation: Singh-Blom, M.*, Natarajan, N.*, Tewari, A., Woods, J. O., Dhillon, I. S., Marcotte, E. M. Prediction and validation of gene-disease associations using methods inspired by social network analyses. PLoS One (submitted)