O'Reilly Bioinformatics Technology Conference.
Books Safari Bookshelf Conferences O'Reilly Network

Arrow Home
Arrow Registration
Arrow Speakers
Arrow Keynotes
Arrow Tutorials
Arrow Sessions
Arrow At-a-Glance
Arrow BOFs
Arrow Posters
Arrow Community
Arrow Events
Arrow Exhibitors
Arrow Sponsors
Arrow Hotel/Travel
Arrow Venue Map
Arrow See & Do
Arrow Press
Arrow Join Mailing List 
Arrow Related Reading

Practical Innovation at BioCon 2003

Poster Session

Neural Network-Based Prediction of Mutation Induced Protein Stability Changes

By Christopher Frenz
Graduate Student, Albert Einstein College of Medicine

An increasing amount of protein products are being used in the development of human therapeutics and pharmaceutical treatments. The viability of these treatments, however, can be greatly limited if stability issues prevent the storage and effective dispensement of the protein product. Protein engineering for increased stability becomes essential in these cases. While many factors that effect protein stability have been elucidated, the engineering of stability mutants is often a complicated task, since the complex intermolecular interactions make mutation induced stability changes difficult to predict. This study utilizes a feed-forward neural network to predict mutation induced stability changes in staphylococcal nuclease. The input to the neural network consisted of sequences of evolutionary-based amino acid similarity scores that were obtained through the comparison of the amino acids in a mutation containing sequence to their positional counterparts in a baseline amino acid sequence. A training set was created which consisted of similarity score sequences for which the stabilities of the corresponding amino acid sequences were known paired with the relative stabilities of the sequences to that of the baseline. Back-propagation of error was used to train the network to output accurate relative stability scores for the sequences in the training set. When the network was used to predict relative stability values for sequences containing mutation combinations not found in the training set, an accuracy of 93% was achieved.

oreilly.com Home | O'Reilly Bookstores | How to Order | O'Reilly Contacts
International | About O'Reilly | Affiliated Companies | Privacy Policy

© 2002, O'Reilly Media, Inc.