Machine Learning QSAR Study for Investigating the Desired Characteristics of a Potential Inhibitory Molecule for PIM-1 Kinase

Overview

This repository contains a comprehensive study focused on the design of inhibitory molecules targeting the PIM-1 kinase protein, which plays a significant role in various types of cancer. Leveraging machine learning techniques, particularly Quantitative Structure-Activity Relationship (QSAR) modeling, this project aims to identify potential inhibitors that can be further validated through molecular docking simulations.

Table of Contents

• Background
• Objectives
• Methodology
• Usage
• To-Do

Background

PIM-1 kinase is a serine/threonine kinase that is implicated in the regulation of cell survival, proliferation, and metabolism. Overexpression of PIM-1 has been linked to several cancers, making it a promising target for therapeutic intervention. This study utilizes machine learning approaches to predict the activity of potential inhibitors based on their chemical structure.

Objectives

• Develop a robust QSAR model to predict the inhibitory activity against PIM-1 kinase.
• Identify and prioritize candidate molecules for further validation.
• Set the foundation for future molecular docking studies to evaluate binding affinities.

Methodology

1. Data Collection:

   • Gathered a dataset of known PIM-1 inhibitors and their corresponding biological activity from CHEMBL database.

2. Feature Engineering:

   • Generated Pubchem molecular fingerprints to represent chemical structures.

3. Model Development:

   • Employed a regression based machine learning algorithms (Random Forest) to build a predictive model.
   • Conducted model evaluation using performance metrics (R²).

4. Model Interpretation:

   • Analyzed feature importance to understand key molecular features influencing activity.

Usage

To run the code in this repository, open the Jupyter notebook and run the cells in the exact order, one by one.

To Do

• Perform molecular docking studies to evaluate the binding affinities of the identified inhibitors.
• Optimize the QSAR model based on feedback and additional data.
• Document the molecular docking methodology and publish the results in a peer-reviewed journal