Wormhole: Ortholog Prediction
through Machine Learning

    Home     Downloads     About     Help


WORMHOLE - the WORM Human OrthoLogy Explorer-is a meta-tool that uses machine learning to predict novel least diverged orthologs (LDOs) by integrating ortholog predictions from 17 algorithms. Support vector machine (SVM) classifiers are trained to distinguish whether a gene is or is not an LDO by comparing the predictions of the consituent algorithms across a set of high- confidence examples of known LDOs (the PANTHER LDOs). Originally conceived to predict orthologs between humans and worms, the scope was later expanded to include five commonly used eukaryotic model organisms: humans (Homo sapiens), mice (Mus musculus), zebrafish (Danio rerio), fruit flies (Drosophila melanogaster), and nematodes (Caenorhabditis elegans). The WORMHOLE SVMs are used to calculate LDO confidence scores (aka WORMHOLE Scores) for genome-wide gene pairs between combination of species.


George L. Sutphin, The Jackson Laboratory
J. Matthew Mahoney, University of Vermont, Department of Neurological Sciences
Keith Sheppard, The Jackson Laboratory
David O. Walton, The Jackson Laboratory
Ron Korstanje, The Jackson Laboratory


For a detailed description WORMHOLE, please refer to our paper in PLOS Computational Biology (download the PDF). If you use WORMHOLE in your research, please reference this paper.

Citation: Sutphin GL, Mahoney JM, Sheppard K, Walton DO, Korstanje R (2016) WORMHOLE: Novel Least Diverged Ortholog Prediction through Machine Learning. PLoS Comput Biol 12(11): e1005182. doi:10.1371/journal.pcbi.1005182

In 2021 Steve Grubb adapted this web application to an updated technology stack.