Analysis of Interface Patterns between Biomolecules Based on Alpha Shape Models
DescriptionMany biological processes are carried out through interactions between biomolecules. The interactions take place when one biomolecule binds onto another one and the characteristics of the interface surface between the biomolecules greatly influence the binding process. For example, drug molecules can bind to a disease causing protein and stop its functions. However, a mutation of the protein can change the properties of the interface surface, cause drug resistance and make the drug ineffective. Thus, it is important to investigate the static and dynamic properties of interface patterns between biomolecules.In this project, we will develop advanced computational methods for the analysis of interface patterns in biomolecular interactions. We will investigate the relationships among different types of atoms in biomolecules using alpha shape models. Various spatial and physicochemical features will be used to distinguish between different interface patterns. Dynamic alpha shapes will be employed to study the flexibility of interface patterns between molecules. We will introduce the novel concepts of EigenInterface and TensorInterface to characterize the interface spaces and extract fingerprints of different types of interface patterns. Our mathematical models and computational algorithms will be used to investigate the relationships between interface patterns and biological functions and to predict potential interactions between two molecules. For example, we can study the interface pattern between a protein mutant and a drug molecule to analyze the drug therapeutic effectiveness and identify possible drug resistance. Our methods will be developed and verified with many real-world datasets. We will address several statistical criteria in performance evaluation of our algorithms. The proposed approach has several advantages. Shape and charge complementarities are two important aspects of biomolecular interactions. We can obtain such complementary features directly and easily from alpha shape models. The dynamic properties of alpha shapes will provide valuable information on different molecular conformations and flexible structures. The EigenInterface and TensorInterface patterns can reveal the interface properties of many molecules in multidimensional feature spaces. In addition, our methodology will be established based on robust mathematical frameworks, including computational geometry and matrix and tensor decomposition. This work will lead to a deep understanding of the principles of biomolecular interactions. The project will also produce valuable computer software for biomedical researchers. The research results will find many practical applications, such as the study of protein functions, identification of drug targets, prediction of drug resistance and selection of personalized medicine. Our research work will bring great benefits to the society.
|Effective start/end date||1/01/16 → 2/06/20|
- Scientific computing,Interface patterns,Protein related interactions,Alpha shape,