Banca de QUALIFICAÇÃO: ELIONAI MOURA CORDEIRO

Uma banca de QUALIFICAÇÃO de DOUTORADO foi cadastrada pelo programa.
STUDENT : ELIONAI MOURA CORDEIRO
DATE: 11/12/2023
TIME: 10:00
LOCAL: Google Meet, https://meet.google.com/kzs-stsg-bai
TITLE:

Deep Learning applied to Waste Interaction Networks with a focus on redirecting drugs and bioligands

KEY WORDS:

Graphs,, Residue Interaction Networks, Mutations Translational Science.

PAGES: 42
BIG AREA: Ciências Biológicas
AREA: Biologia Geral
SUMMARY:

Biological data for drug development and the study of new ligand-protein interactions can be represented as labeled graphs with relationships and interdependencies between the objects studied as a consensual methodology. In recent decades, several tools for building waste interaction networks (RINs) have been described in the literature. However, none of them took advantage of the full potential of such an approach for identifying critical residues (and probably more restricted to mutations and polymorphisms), for studying the variation of interactions between different conformations, and mainly, for using such information to search for new drugs or repositioning existing ones on new targets. The present work proposes a mathematical and computational definition to calculate biochemical interactions of the molecules of interest and construct highly connected graphs that reliably represent the types of interactions, more specifically hydrogen bonds, salt bridges, van der Waals interactions, stacking interactions 𝝅 -𝝅. We thus built a database with graph networks (calculated from .PDBx/mmCIF files) based on structures from the RCSB PDB databases, as well as models from structural predictions from AlphaFold2, focusing on human proteins. . With this in mind, we developed a new tool for constructing and comparing RINs, as well as for describing the interactions between proteins and ligands and, mainly, for identifying residues important for function. Additionally, the tool features modularity, scalability and data output in simple files or web visualization. We also carried out some comparative tests of our solution in contrast to approaches found in the literature. During the validation of our approach, we conducted comparative tests with other approaches found in the literature, demonstrating the effectiveness and accuracy of our tool in identifying residues critical to protein function. These initial findings highlight the significant potential of our methodology in researching new drugs, repositioning existing drugs, and understanding molecular interactions at the biochemical level.

COMMITTEE MEMBERS:
Interno - 1893445 - EUZEBIO GUIMARAES BARBOSA
Presidente - 1513597 - JOAO PAULO MATOS SANTOS LIMA
Interno - 3063244 - TETSU SAKAMOTO