A Custom Bio-Inspired Algorithm for the Molecular Distance Geometry Problem

• Sarah Ribeiro Lisboa Carneiro CEFET-RJ https://orcid.org/0009-0004-6877-5205
• Michael Ferreira de Souza UFC https://orcid.org/0000-0001-6751-2877
• Douglas O. Cardoso Polytechnic Institute of Tomar https://orcid.org/0000-0002-1932-334X
• Luís Tarrataca CEFET-RJ https://orcid.org/0000-0001-9359-5143
• Laura S. Assis CEFET-RJ https://orcid.org/0000-0003-3081-9722

Resumo

Protein structure allows for an understanding of its function and enables the evaluation of possible interactions with other proteins. The molecular distance geometry problem (MDGP) regards determining a molecule’s three-dimensional (3D) structure based on the known distances between some pairs of atoms. An important application consists in finding 3D protein arrangements through data obtained by nuclear magnetic resonance (NMR). This work presents a study concerning the discretized version of the MDGP and the viability of employing genetic algorithms (GAs) to look for optimal solutions. We present computational results for input instances whose sizes varied from 10 to 10³ atoms. The results obtained show that approaches to solving the discrete version of the MDGP based on GAs are promising.