Scholarlink

Pharmaceuticals play a critical role in both the prevention and treatment of diseases. However, the drug development process is often lengthy, complex, and unpredictable, despite the growing demand for novel and effective treatments. To streamline this process, there is a pressing need for virtual in-silico screening frameworks that can reduce computation time, minimize the selection of non-viable compounds, and improve the accuracy of potential drug candidates. This framework is based on the hypothesis that low binding free energy is a key indicator of a compound's ability to function as an effective inhibitor for a specific target protein. By focusing on this metric, we aim to efficiently classify compounds with high therapeutic potential while eliminating those with less favorable profiles. Our study focuses on the Sirtuin6 protein, a member of the class-III histone deacetylase enzyme family. Sirtuin6 is critical for regulating a wide range of biological processes, including transcription, cell survival, and the de-acylation of proteins such as histones and transcription factors. Given its role in longevity and metabolic regulation, Sirtuin6 is an attractive target for drug development. To enhance the efficiency of the virtual screening process, we employ a machine-learning-driven approach utilizing supervised learning algorithms, including Decision Trees, Random Forests, Logistic Regression, and Support Vector Machines (SVM). These models help optimize resource usage, accelerate computational tasks, and improve the selection of novel inhibitors with low binding free energy.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Developing a Framework Driven by Analytics to Forecast the Binding Free Energy Level and, Consequently, the New Inhibitor Nature of Medicinal Chemical Compounds

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Ashutosh Kar,
  • Dipayan Chakraborty,
  • Sangita Mazumder,
  • Pratyay Ranjan Datta

摘要

Pharmaceuticals play a critical role in both the prevention and treatment of diseases. However, the drug development process is often lengthy, complex, and unpredictable, despite the growing demand for novel and effective treatments. To streamline this process, there is a pressing need for virtual in-silico screening frameworks that can reduce computation time, minimize the selection of non-viable compounds, and improve the accuracy of potential drug candidates. This framework is based on the hypothesis that low binding free energy is a key indicator of a compound's ability to function as an effective inhibitor for a specific target protein. By focusing on this metric, we aim to efficiently classify compounds with high therapeutic potential while eliminating those with less favorable profiles. Our study focuses on the Sirtuin6 protein, a member of the class-III histone deacetylase enzyme family. Sirtuin6 is critical for regulating a wide range of biological processes, including transcription, cell survival, and the de-acylation of proteins such as histones and transcription factors. Given its role in longevity and metabolic regulation, Sirtuin6 is an attractive target for drug development. To enhance the efficiency of the virtual screening process, we employ a machine-learning-driven approach utilizing supervised learning algorithms, including Decision Trees, Random Forests, Logistic Regression, and Support Vector Machines (SVM). These models help optimize resource usage, accelerate computational tasks, and improve the selection of novel inhibitors with low binding free energy.