Gene functional classification from heterogeneous data
RECOMB '01 Proceedings of the fifth annual international conference on Computational biology
Gene Classification Using Expression Profiles: A Feasibility Study
BIBE '01 Proceedings of the 2nd IEEE International Symposium on Bioinformatics and Bioengineering
Hi-index | 0.01 |
Existing data mining tools can only achieve about 40% precisionin function prediction of unannotated genes. We developed a genefunction prediction tool based on profile Hidden Markov Models(HMMs). Each function class was modelled using a distinct HMM whoseparameters were trained using yeast time-series gene expressionprofiles. Two structural variants of HMMs were designed and tested,each of them on 40 function classes. The highest overall predictionprecision achieved was 67% using double-split HMM withleave-one-out cross-validation. We also attempted to generaliseHMMs to dynamic Bayesian networks for gene function predictionusing heterogeneous data sets.