||Prediction of MHC class II epitopes using genetic algorithms and other metaheuristics
||Mygind, Henrik Egeberg
||Fischer, Paul (Algorithms and Logic, Department of Informatics and Mathematical Modeling, Technical University of Denmark, DTU, DK-2800 Kgs. Lyngby, Denmark)
||Technical University of Denmark, DTU, DK-2800 Kgs. Lyngby, Denmark
||As a part of immunological bioinformatics research metaheuristics are used in
the prediction of amino acid chains binding to the MHC-II molecule. A method
of prediction the Gibbs sampler, developed by Nielsen et. al., uses simulated
annealing to optimise an objective function. We have replaced the simulated annealing
with a genetic algorithm, in an attempt to perform a better optimisation
and thereby achieve a better prediction.
The bioinformatical problem of MHC-II binding is interpreted from a computer
science perspective. The genetic algorithm has been implemented in Java. The
results have been thoroughly analysed and compared with the Gibbs sampler
using the statistical tool R.
The genetic algorithm has proven more effective at optimising the objective
function than the Gibbs sampler. This improvement in optimisation has not
entailed a better prediction. This leads to the conclusion that to get a better
prediction a better objective function has to be found.
Creation date: 2009-07-02
Update date: 2010-10-28