Abstract
Predicting protein structure from sequence remains a major open problem in protein biochemistry. One component of predicting complete structures is the prediction of inter-residue contact patterns (contact maps). Here, we discuss protein contact map prediction by machine learning. We describe a novel method for contact map prediction that uses the evolution of logic circuits. These logic circuits operate on feature data and output whether or not two amino acids in a protein are in contact or not. We show that such a method is feasible, and in addition that evolution allows the logic circuits to be trained on the dataset in an unbiased manner so that it can be used in both contact map prediction and the selection of relevant features in a dataset.
Highlights
Proteins are important biological molecules that perform many functions in an organism
The evolution continued for 100 k updates, and at intervals of 25 k updates, results on the test set for committee sizes up to 60 were recorded
Even though the treatments ended at 100 k updates, the best results for all treatments were from 75 k updates; this is probably due to overfitting of Markov networks to the training set in their evolution
Summary
Proteins are important biological molecules that perform many functions in an organism These molecules are composed of a string of amino acids comprised of a 20-letter ‘‘alphabet’’ of amino acids. The sequence of amino acids is referred to as the primary structure of the protein Beyond this primary structure, proteins are arranged in a higher-order, threedimensional secondary structure composed of motifs such as alpha-helices and beta sheets. Proteins are arranged in a higher-order, threedimensional secondary structure composed of motifs such as alpha-helices and beta sheets This secondary structure in turn is arranged into a tertiary structure that forms protein domains, which in turn can form a quaternary structure that is composed of multiple protein domains (McNaught & Wilkinson, 1997). Protein structure in turn greatly influences the function of a protein
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have