MacMatch: a tool for pattern-based protein secondary structure prediction
1Department of Pharmaceutical Chemistry, University of CaliforniaSan Francisco San Francisco, CA 94143-0446, USA
2Department of Medicine, University of CaliforniaSan Francisco San Francisco, CA 94143-0446, USA
3Departments of Biochemistry and Biophysics,University of CaliforniaSan Francisco San Francisco, CA 94143-0446, USA
A program is described for predicting the secondary structure of globular proteins on an Apple Macintosh computer. MacMatch produces turn predictions using augmented regular expression pattern matching. Helix and strand predictions are based on a neural network. Protein structural class (
/
), ß/ß,
/ß can be exploited to improve secondary structure prediction. The program is simple to use, and the package includes sets of tested patterns and trained neural networks.
Received on July 1, 1992; accepted on October 15, 1992