Current projects deal with:

  • Discovering the biochemical function of Structural Genomics proteins that are currently of unknown function
  • Exploring the multilayer nature of enzyme active sites
  • Using THEMATICS and POOL for the analysis of specific enzymes of biological and medical importance
  • Application of our computational methods to protein design
  • Computational guidance of drug discovery

Our novel machine learning (ML) methodologies utilize THEMATICS and other bioinformatics tools to identify protein interaction sites.

We are working with collaborators in Mathematics and in Computer Science at Northeastern University on novel methodologies for predicting protein function, adapting our SALSA method for high-throughput screening. We are also working in collaboration with experimentalists to test and verify our predictions pertaining to multilayer active sites and also to test our predicted functional annotations of Structural Genomics proteins.

Figure 1

Functionally important residues predicted in and around the active site of PNGase F.

Figure 2

Active site residues and corresponding theoretical titration curves in Arginine Kinase.