RESUMO
We present a novel and interpretable approach for assessing small-molecule binding using context explanation networks. Given the specific structure of a protein/ligand complex, our CENsible scoring function uses a deep convolutional neural network to predict the contributions of precalculated terms to the overall binding affinity. We show that CENsible can effectively distinguish active vs inactive compounds for many systems. Its primary benefit over related machine-learning scoring functions, however, is that it retains interpretability, allowing researchers to identify the contribution of each precalculated term to the final affinity prediction, with implications for subsequent lead optimization.
Assuntos
Redes Neurais de Computação , Ligação Proteica , Proteínas , Bibliotecas de Moléculas Pequenas , Ligantes , Bibliotecas de Moléculas Pequenas/química , Bibliotecas de Moléculas Pequenas/farmacologia , Bibliotecas de Moléculas Pequenas/metabolismo , Proteínas/química , Proteínas/metabolismo , Aprendizado de MáquinaRESUMO
Convergent evolution is a widespread phenomenon. While there are many examples of convergent evolution at the phenotypic scale, convergence at the molecular level has been more difficult to identify. A classic example of convergent evolution across scales is that of the digestive lysozyme found in ruminants and Colobine monkeys. These herbivorous species rely on foregut fermentation, which has evolved to function more optimally under acidic conditions. Here, we explored if rodents with similar dietary strategies and digestive morphologies have convergently evolved a lysozyme with digestive functions. At the phenotypic level, we find that rodents with bilocular stomach morphologies exhibited a lysozyme that maintained higher relative activities at low pH values, similar to the lysozymes of ruminants and Colobine monkeys. Additionally, the lysozyme of Peromyscus leucopus shared a similar predicted protonation state as that observed in previously identified digestive lysozymes. However, we found limited evidence of positive selection acting on the lysozyme gene in foregut-fermenting species and did not identify patterns of convergent molecular evolution in this gene. This study emphasizes that phenotypic convergence need not be the result of convergent genetic modifications, and we encourage further exploration into the mechanisms regulating convergence across biological scales.
Assuntos
Muramidase , Roedores , Animais , Muramidase/genética , Muramidase/química , Estômago , Primatas , Ruminantes/genética , Evolução Molecular , Filogenia , Evolução BiológicaRESUMO
We present a novel and interpretable approach for predicting small-molecule binding affinities using context explanation networks (CENs). Given the specific structure of a protein/ligand complex, our CENsible scoring function uses a deep convolutional neural network to predict the contributions of pre-calculated terms to the overall binding affinity. We show that CENsible can effectively distinguish active vs. inactive compounds for many systems. Its primary benefit over related machine-learning scoring functions, however, is that it retains interpretability, allowing researchers to identify the contribution of each pre-calculated term to the final affinity prediction, with implications for subsequent lead optimization.
RESUMO
Background: Minimally invasive techniques should be employed to maintain pulp vitality without affecting physiological tooth resorption for pediatric patients. This study aims to evaluate clinical and radiographic success of light cure calcium silicate and resin-modified glass ionomer (RMGI) as indirect pulp-capping agent in primary molars in children between 5 and 9 years. Materials and Methods: In this randomized clinical trial study, 40 primary molars with the International Caries Detection and Assessment System criteria 4-6 score were divided randomly into two groups after computerized randomization method. After caries excavation by minimally invasive dentistry principle, affected dentin was lined by either light cure calcium silicate (Group I) or RMGI (Group II) liner followed by composite restoration. Clinical success was determined with the absence of symptoms and bitewing radiographs were taken at baseline, 3, and 6 months' interval to measure increase in dentin increment using ImageJ software. Statistical analysis for intergroup comparison was done using Paired t-test, and Independent t-test was used for intragroup comparison. The level of statistical significance was set at P < 0.05. Results: Change in dentin increment seen from baseline to 6 months in Group I was 0.19 mm and in Group II was 0.20 mm (P = 0.924). Intragroup increment was statistically significant during all follow-up intervals for both the groups (Group I - P ≤ 0.001, Group II - P = 0.009). For baseline remaining dentin thickness >1.5 mm, statistically significant increase was observed in the dentin increment at 3 months' interval between both the groups. Conclusion: Both TheraCal LC and Vitrebond show acceptable clinical and radiographic results when used in primary molars as indirect pulp treatment agents. Both the liners are equivalent to each other in terms of feasibility and cost-effectiveness but TheraCal LC can be preferred due to better handling and avoiding the manipulation step thereby reducing treatment time.