Arby:: automatic protein structure prediction using profile-profile alignment and confidence measures

Motivation: Arby is a new server for protein structure prediction that combines several homology-based methods for predicting the three-dimensional structure of a protein, given its sequence. The methods used include a threading approach, which makes use of structural information, and a profile-profile alignment approach that incorporates secondary structure predictions. The combination of the different methods with the help of empirically derived confidence measures affords reliable template selection. Results: According to the recent CAFASP3 experiment, the server is one of the most sensitive methods for predicting the structure of single domain proteins. The quality of template selection is assessed using a fold-recognition experiment.