CRF
Home
Projects
BioTagger
MSTParser
CRAIG
StructLearn
CALO
EVIGAN
Publications
People
Computing
Links
Penn
Penn CIS
Machine Learning at Penn
Genomics and Computational Biology

EVIGAN Gene Predictor

Evigan is a eukaryotic gene predictor which produces consensus gene models by integrating multiple sources of evidence. Evigan can incorporate various types of evidence such as gene models predicted by gene finders, EST matches, protein-to-genome alignment, etc. Employing a Dynamic Bayesian Network(DBN), Evigan makes predictions by infering the most likely consensus gene models given observed source evidence. Its model parameters are estimated by the Expectation-Maximization(EM) algorithm, eliminating the need of curated training data.

You can download the source code here.