PQG 2019

In Structural Bioinformatics, several problems do not present a computational method that can guarantee the minimum quality of their solutions. Structural bioinformatics deals with problems where the rules governing biochemical processes and their relationships are partially known, making it challenging to develop efficient computational strategies. Among these problems, two are the most challenging: predicting the three-dimensional structure of macromolecules and the molecular docking problem. Predicting the structure of a polypeptide/protein just from its linear sequence of amino acid residues represents a challenging problem in the field of mathematical optimization, being classified in computational complexity as an NP-complete problem. The challenge occurs due to the explosion of possible conformations that a long chain of amino acid residues can assume. Another significant problem is related to the construction of algorithms for assisted drug/drug development. In this context, molecular docking (docking) algorithms are needed to determine the preferred orientation of a small molecule, drug against a receptor molecule (e.g., protein). With the increasing amount of biological molecule structures available, effective and efficient techniques are needed to treat these structures. Because of this, there is a great need to build more intelligent strategies that can deliberate better solutions to the problem. In this project, we are working on the development of metaheuristics for Structural Bioinformatics problems. Metaheuristics are among the most frequent and powerful techniques used in situations where knowledge about the problem is restricted and exact solutions are not currently computable. Metaheuristics do not guarantee the optimal solution, but they provide a good approximation solution with little computational effort. To tackle the above problems, we propose the development of multimodal and multi-objective metaheuristics models. The new metaheuristics models will enable better exploration of the search space compared to state-of-the-art strategies.