Researchers develop novel technique to map protein interactions

By Sathya Achia Abraham

Virginia Commonwealth University is part of an international research team that developed a technique to map the protein interactions in a species of budding yeast, a discovery that may be applied to further understand cell biology processes and genes involved in human disease.

The study, published online in the Early Edition of the Proceedings of the National Academy of Sciences the week of Jan. 19-23, is a collaboration between researchers based at Unidade de Sistemas Biológicos, Universidade de Coimbra, and BIOCANT, a research and development center, all located in Portugal, and the VCU Center for the Study of Biological Complexity and the Department of Computer Science in the VCU School of Engineering.

The collection of physical interactions among the proteins in a cell is called the interactome. In this study, the team reported on the development of a novel computational algorithm that is shown to be able to produce highly reliable protein-protein interaction and the resulting protein complexes from publicly available raw high throughput data. The authors also propose to structure an interactome in terms of predicted permanent protein complexes and predicted transient, nongeneric interactions between these complexes.

Studying how proteins interact on the molecular level is a systems biology approach that enables researchers to perform an integrated analysis of the cell rather than just an isolated study of individual components. By understanding complex interactions among the proteins, researchers can gain further insight into genes that may be involved in human diseases.

Investigators have been facing a number of challenges in analyzing high throughput datasets that aim to study these interactions systematically. For example, in a high throughput protein-protein interaction study, a protein is used as a bait to pull down the interacting partners, and then the interacting partners are identified by some method, such as mass spectrometry. One of the problems with such an approach is high “noise,” meaning that proteins which do not interact will appear to be interacting by the pull down assay. Such noise can come from many sources, for example, proteins can stick to another protein non-specifically.

“The algorithm we have created allows us to obtain reliable interaction data and structure them in a meaningful form, amenable and valuable for further biological research,” said Yuan Gao, Ph.D., an assistant professor in the Department of Computer Science in VCU’s School of Engineering and the VCU Center for the Study of Biological Complexity, who co-led the study.

“We demonstrated that when we apply our algorithm to the construction of an interactome for S. cerevisiae, it yields reliability typical of low throughput experiments, using high throughput data.” 

This work was supported by grants from the National Institutes of Health, BIOCANT, and the Universidade de Coimbra.

Andre X. C. N. Valente, Ph.D., with the Unidade de Sistemas Biológicos at BIOCANT and  Universidade de Coimbra in Portugal co-led this investigation with Gao. They collaborated with Seth Roberts, Ph.D., and Gregory Buck, Ph.D., director of VCU Life Sciences’ Center for the Study of Biological Complexity.