Prediction of protein secondary structure using two-layered neural networks
Abstract
Prediction of protein 2° structure can be treated as a problem in pattern recognition for which neural networks can be applied, since certain amino acid sequences favor the formation of certain 2° structures. To determine the optimum parameters for 2° structure prediction by two-layered neural networks, 3-state (helix, sheet, and coil) predictive accuracy was tested depending on the state threshold, the number of inputs used, the number of hidden nodes used, and the number of output nodes used.
Downloads
Issue
Ins-FIRE-ing excellence in physics education and research
23-25 October 2002, Ateneo de Naga University