Loading...
Publication
Chemslotprotein : expanding the druggable universe to unlock chemists´creativity
| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 1.79 MB | Adobe PDF |
Authors
Abstract(s)
A descoberta de novos medicamentos tem-se centrado predominantemente num
subconjunto limitado de alvos do proteoma humano, representando apenas 15% do total
de proteínas. Isto acontece porque os cientistas apenas se focam num pequeno conjunto de
proteínas, de forma a evitar alvos e vias de sinalização para as quais ainda pouco se
conhece. Consequentemente, existem profundas lacunas de identificação e validação de
novos alvos farmacêuticos, o que leva a que cerca de 85% dos alvos associados a doenças
permaneçam inexplorados. De forma a contribuir para melhorar o conhecimento nesta
área, esta tese realiza uma série de estudos computacionais com o principal objetivo de
identificar e analisar sítios de ligação de proteínas de forma a facilitar a descoberta de
novos alvos cuja função possa ser modulada. Para atingir estes objetivos, este projeto
categoriza as proteínas em três grupos distintos, com base no trabalho de Oprea e
colaboradores, bem como em dados dos bancos de dados UniProt, ChEMBL e Protein Data
Bank (PDB). O primeiro grupo contém proteínas para as quais existem medicamentos já
aprovados ou que se encontram em ensaios clínicos, classificando-os como “Alvos Bem
Conhecidos” (WKT). O segundo grupo contém proteínas para as quais ainda não foram
identificados compostos ativos, designados como “Alvos Yet NOT druggable” (YNOTs).
Por fim, o último grupo contém alvos que apresentam desafios para modelação, que
classificamos como de “Difícil Obtenção de Efeito Farmacológico” (DOPE). A abordagem
analítica utilizada envolve a identificação de cavidades para cada um dos PDB’s
associados aos três grupos. Para isso, utilizámos o software CAVIAR, que possibilita a
identificação de cavidades nas proteínas, incluindo os seus sítios de ligação.
Posteriormente, realizámos uma procura por semelhança de cavidades com o grupo WKT.
Os alvos mais promissores encontrados no grupo dos YNOTs foram, a estrutura cristalina
da isobutiril-CoA desidrogenase complexada com o seu substrato e a estrutura cristalina
da decarboxilase do ácido UDP-glucurónico humano. Para o grupo DOPE, o alvo mais
promissor encontrado foi a subunidade delta da fosfodiesterase 6 (PDE6δ). Além disso,
obtivemos um conjunto de 178 proteínas que consideramos como promissores alvos
farmacológicos.
Recurrent research in the discovery of new drugs has predominantly focused on a limited subset of the human proteome, accounting for just 15%. This emphasis arises from researchers’ focus on a small set of genes, to evade unexplored research paths. Consequently, a significant gap exists in the validation process for new pharmaceutical compounds, with an estimated 85% of disease-associated targets remaining underexplored. To address this critical issue, this thesis undertakes a series of computational studies with a primary objective of detecting and analyzing protein binding sites to facilitate the discovery of novel drug targets. In pursuit of these goals, this project categorizes proteins into three distinct groups, drawing from the work by Oprea and collaborators, as well as data from the UniProt, ChEMBL, and Protein Data Bank (PDB) databases. The first group contains targets that have either approved drugs or those currently undergoing clinical trials, classifying them as Well-Known Targets (WKT). The second group contains targets for which no active compounds have yet been identified, designated as Yet NOT druggable targets (YNOTs). Finally, the last group contains targets posing modeling challenges, which we classify as Difficult to Obtain Pharmacological Effect (DOPE). The analytical approach involves performing pocket detection for each PDB structure associated with protein targets. To achieve this, we employed the CAVIAR software, enabling the identification of cavities within each protein, including their binding sites. Subsequently, we conducted a search for similarities among these pockets. The most promising targets identified in the YNOTs group were the crystal structure of isobutyryl-CoA dehydrogenase complexed with substrate and the crystal structure of Human UDP-glucuronic acid decarboxylase. For the DOPE group, the most promising target discovered was the phosphodiesterase 6 delta subunit (PDE6δ). Furthermore, we compiled a database featuring 178 targets that also exhibit promise as potential pharmacological targets.
Recurrent research in the discovery of new drugs has predominantly focused on a limited subset of the human proteome, accounting for just 15%. This emphasis arises from researchers’ focus on a small set of genes, to evade unexplored research paths. Consequently, a significant gap exists in the validation process for new pharmaceutical compounds, with an estimated 85% of disease-associated targets remaining underexplored. To address this critical issue, this thesis undertakes a series of computational studies with a primary objective of detecting and analyzing protein binding sites to facilitate the discovery of novel drug targets. In pursuit of these goals, this project categorizes proteins into three distinct groups, drawing from the work by Oprea and collaborators, as well as data from the UniProt, ChEMBL, and Protein Data Bank (PDB) databases. The first group contains targets that have either approved drugs or those currently undergoing clinical trials, classifying them as Well-Known Targets (WKT). The second group contains targets for which no active compounds have yet been identified, designated as Yet NOT druggable targets (YNOTs). Finally, the last group contains targets posing modeling challenges, which we classify as Difficult to Obtain Pharmacological Effect (DOPE). The analytical approach involves performing pocket detection for each PDB structure associated with protein targets. To achieve this, we employed the CAVIAR software, enabling the identification of cavities within each protein, including their binding sites. Subsequently, we conducted a search for similarities among these pockets. The most promising targets identified in the YNOTs group were the crystal structure of isobutyryl-CoA dehydrogenase complexed with substrate and the crystal structure of Human UDP-glucuronic acid decarboxylase. For the DOPE group, the most promising target discovered was the phosphodiesterase 6 delta subunit (PDE6δ). Furthermore, we compiled a database featuring 178 targets that also exhibit promise as potential pharmacological targets.
Description
Tese de mestrado, Química Medicinal e Biofarmacêutica, 2023, Universidade de Lisboa, Faculdade de Farmácia.
Keywords
Binding site Pocket detection Computational methods Teses de mestrado - 2023