Raschid Helps Lead UMIACS Research in Bioinformatics

Louiqa Raschid (left), a professor in the Smith School of Business with an appointment in UMIACS, recently collaborated with Maria Esther Vidal and Guillermo—both researchers at Palma Universidad Simon in Bolivar, Venezuela—on an innovative approach to finding drug-target communities.

Their semEP method to community detection is based on edge partitioning; that is, a drug or target that can participate in multiple communities. The researchers are exploiting semantic knowledge from ontologies and similarities and believe their work can increase the prediction quality of other machine learning based methods.

The team’s research was presented at the 13th International Semantic Web Conference in Trentino, Italy, in October. The project is supported by a National Science Foundation (NSF) grant.

In related news, research by computer science graduate student Shobeir Fakhraei, who is advised by adjunct computer science professor and former UMIACS researcher Lise Getoor, was chosen as the lead article for the September 2014 issue of the IEEE/ACM Transactions on Computational Biology and Bioinformatics. It was also featured on the cover.

Fakhraei's approach is based on collective inference and used structured knowledge, including the semantic similarities between pairs of drugs and pairs of targets. The implementation in PSL (Probabilistic Soft Logic) outperformed state-of-the-art drug-target interaction prediction methods.

Raschid and Bert Huang of Virginia Tech are co-authors on this research, which is supported by an NSF grant.

Read more about Fakhraei's research here.