Parallel Decomposition of 3D Surfaces in Images of Local Descriptors for Molecular Screening

摘要

An emerging application field for structure matching is related to in silico studies of molecular biology. Considering that protein function is mainly related to its external morphology, the possibility to match macromolecular surfaces is very important to infer information about the possible interaction and function of biological components. Most of the present approaches aim to match complete objects and they exploit a complete 3D representations of them. These aspects do not suit well with the local formulation of the matching issue in the biological application domain. The problem is that the shapes of macromolecules that are experimentally known to dock cannot present high similarities except in small important functional regions. This is the reason why we chose to adapt a well known image based method, originally studied for matching similar structures, for screening possible interactions of couples of macromolecule. In this context the proposed work presents a parallel algorithm able to efficiently decompose high detailed 3D surfaces into a proper set of images to use for the macromolecular screening.