University of Dundee

Artificial Intelligence

Geoff Barton 

Developing methods to analyse large biological datasets

Geoff Barton's computuational biology group develop techniques to analyse large biological datasets and predict the structure and function of biological molecules and systems. His group has applied machine learning methods, principally Artificial Neural Networks (ANNs) to a wide range of predictive challenges.  The best known is the JPred protein secondary structure predictor that is currently the most accurate method in its class and provides up to 500,000 predictions per month for scientists worldwide.  His group has particular skill in the critical step of developing and optimising datasets suitable for training ANNs and other classes of Machine Learning algorithm.