Attila Gursoy & Ozlem Keskin Discuss Understanding Drug-Protein Interactions
Fast Breaking Paper Commentary, December 2010
|
Article: Identification of computational hot spots in protein interfaces
Authors: Tuncbag, N;Gursoy,
A;Keskin, O |
Attila Gursoy & Ozlem Keskin talk with ScienceWatch.com and answer a few questions about this month's Fast Breaking Paper paper in the field of Computer Science.
Why do you think your paper is highly
cited?
Drugs can target protein-protein interfaces, and certain amino acids in interfaces are the key contact points (hot spots). We describe an efficient algorithm to predict hot spots of protein interfaces. Its efficiency allows us to apply the method on a large scale (many proteins).
Does it describe a new discovery, methodology, or
synthesis of knowledge?
It is partly methodology, partly synthesis of knowledge. Most machine learning-based hot-spot prediction methods learn complex relations between training data and hot spots; however, it is very difficult to translate these relations into simple, intuitive rules. We present a new efficient algorithm based on some simple rules to determine and mine computational hot spots in protein interfaces.
Would you summarize the significance of your paper
in layman's terms?
Coauthor Ozlem Keskin.
Drugs bind to protein surfaces. Some amino acids, called hot spots, at the surfaces are more important in binding. Prior knowledge of which amino acids are hot spots makes it possible to design new and more efficient drugs and understand how the drug acts to inhibit the protein function. For predicting hot spots, we determined a few simple rules involving physico-chemical properties of individual residues. Computational effectiveness of the model makes it favorable for hot spot prediction at a large scale.
How did you become involved in this research, and
how would you describe the particular challenges, setbacks, and
successes that you've encountered along the way?
Biology is increasingly becoming an information-driven science as enormous of amount of data is produced with the recent advances in experimental biology. This naturally leads to multidisciplinary research involving computer science as well. I believe that challenging complex problems require interdisciplinary approach for innovative solutions.
Our multidisciplinary computational biology group and our work on protein-protein interactions resulted in this particular hot-spot prediction work. A major challenge is to create an harmonious group with complementary partners.
Where do you see your research leading in the
future?
Our research has important practical applications, particularly in drug discovery and design. The method may be used for the design of specific therapeutic agents for protein interactions. In the future, part of our research will be on understanding drug-protein interactions in a larger context (protein interaction networks).
Do you foresee any social or political
implications for your research?
The drug discovery and design is a challenging and expensive process.
Society will benefit significantly from any development in this area. Our
research may help narrow down potential binding sites to be used for drug
design.
Attila Gursoy
Professor of Computer Science
College of Engineering
Koc University, Istanbul
Ozlem Keskin
Assoc. Professor of Chemical and Biological Engineering
College of Engineering
Koc University, Istanbul
KEYWORDS: GLOBULAR-PROTEINS; SUBUNIT INTERFACES; RECEPTOR COMPLEX; BINDING-ENERGY; SEQUENCE; PREDICTION; CONSERVATION; DATABASE; SITES; HORMONE.