Drug Discovery using AI & Machine Learning Session
Kindly submit your Doubts here (Regarding any task ): Click here
Hi Interns ,
Welcome to the Internship Task Portal
Here you will get all the intimations regarding your tasks , sessions , assignments and any updated briefing .
(The information and content provided in this portal is purely confidential and is under protected surveillance by the technical support team.)
🧬 Drug Discovery using AI & Machine Learning – Internship Update
The Drug Discovery using AI & Machine Learning Internship is designed to help students build strong foundations in modern computational pharmacology while applying advanced machine learning concepts to real-world therapeutic and clinical pipelines. This program focuses on combining theoretical chemical-biological knowledge with a practical understanding of advanced computational technologies such as molecular modeling, virtual screening, ADMET prediction, deep learning for generative chemistry, and algorithmic target identification.
Throughout the internship, participants will:
-
Learn the fundamentals of traditional and AI-accelerated drug discovery funnels.
-
Understand molecular representations, chemical structure formatting, and structural informatics.
-
Explore virtual compound screening, molecular docking simulations, and quantitative structure-activity relationship (QSAR) modeling.
-
Study applications of deep learning networks and generative models in designing novel, drug-like therapeutic molecules.
-
Gain exposure to production-ready informatics workflows, chemical databases, and modern pharmaceutical research benchmarks.
-
Build a capstone mini-project or case study based on real-world predictive analytics or molecular generation challenges.
Each task includes:
-
📘 Session-based learning
-
📝 Notes preparation
-
💻 Hands-on practical assignments
-
🔬 Case study analysis
-
📤 Submission via Google Forms
By the end of the internship, students will gain practical exposure in computational chemistry and artificial intelligence while developing skills useful for careers in computer-aided drug design (CADD), machine learning engineering, pharmaceutical industries, biomedical research, and healthcare innovation.
TASK I: Drug Discovery – AI-Driven Target Identification & Development Processes
🎯 Objective:
Understand the fundamental, multi-stage pipeline of traditional drug discovery and analyze how Artificial Intelligence and Machine Learning accelerate target identification and validation to reduce early-stage attrition.
🎥 Session:
-
Session 01 – Drug Discovery and Development Process
-
Session Link: click here to access
📝 Task:
Write Notes On:
-
The classic drug discovery funnel: Target identification, lead optimization, preclinical testing, and clinical trial phases ($Phase\ I$ to $Phase\ IV$).
-
The economic bottlenecks (high cost, long timelines, high failure rates) of traditional pharma pipelines.
-
How ML models (e.g., Deep Learning, Graph Neural Networks) are used to scan biomedical literature, genomics data, and biological networks to predict disease-associated targets.
🔬 Practice / Research:
-
Research a real-world case study of an AI-designed drug candidate entering clinical trials (e.g., candidates designed by Insilico Medicine or Exscientia).
-
Explain the difference between phenotypic screening and target-based drug discovery in an algorithmic context.
📂 Assignment:
-
Task: Write a short overview report (200–300 words) comparing the traditional target identification timeframe with an AI-accelerated approach. Highlight at least two specific machine learning architectures used to map protein-protein interactions or gene expression signatures for target discovery.
📂 Internship Task Completion Status Form:
After completing this task, interns must fill out the Internship Task Completion Status Form and upload their notes/practice work.
👉 Form Link:
https://forms.gle/dcfdpoDgksgP9KUY8
TASK II: Computational Chemistry – How Drugs Are Discovered and Developed using ML
🎯 Objective:
Explore the computational and machine learning approaches used for virtual screening, molecular property prediction (ADMET), and generative molecular design.
🎥 Session:
-
Session 02 – How Drugs Are Discovered and Developed
-
Session Link: click here to access
📝 Task:
Write Notes On:
-
Molecular representations for machine learning: SMILES strings, molecular fingerprints (e.g., ECFP4), and 3D geometric coordinates.
-
ADMET Prediction: Using classification and regression models to predict Absorption, Distribution, Metabolism, Excretion, and Toxicity profiles before chemical synthesis.
-
An introduction to generative chemistry: How Variational Autoencoders (VAEs) and Generative Adversarial Networks (GANs) design novel, drug-like molecules from scratch.
🔬 Practice / Research:
-
Look up an open-source chemical dataset repository (such as ChEMBL or PubChem) and describe how molecular structures are indexed and filtered.
-
Explain the concept of “Chemical Space” and why searching it efficiently requires advanced optimization algorithms.
📂 Assignment:
-
Task: Create a workflow blueprint or block diagram detailing an ML-based virtual screening pipeline. The pipeline must demonstrate how a library of millions of compounds (SMILES strings) is filtered down using quantitative structure-activity relationship (QSAR) models and ADMET filters to yield top lead candidates.
📂 Internship Task Completion Status Form:
After completing this task, interns must fill out the Internship Task Completion Status Form and upload their notes/practice work.
👉 Form Link:
https://forms.gle/dcfdpoDgksgP9KUY8
TASK III: Drug Targets and Target Discovery
🎯 Objective:
Deepen your understanding of biological targets (proteins, nucleic acids, and receptors) and learn how machine learning algorithms identify and validate disease-relevant nodes in complex biological pathways.
🎥 Session:
-
Session 3 – Drug Targets and Target Discovery. The search for new drugs.
-
Session Link: click here to access
📝 Task:
Write Notes On:
-
What constitutes a “druggable” target versus an “undruggable” target.
-
The types of biological targets: Enzymes, G-protein coupled receptors (GPCRs), ion channels, and nuclear receptors.
-
Network biology and graph machine learning techniques used to discover novel disease-associated targets.
🔬 Practice / Research:
-
Research the therapeutic target for a widely prescribed medication (e.g., Imatinib or Metformin) and find its entry on Open Targets Platform or DrugBank.
-
Explain how transcriptomic data profiling helps computational biologists validate a target’s role in a specific disease state.
📂 Assignment:
-
-
Task: Write a brief research summary (200–300 words) on how deep learning models (such as alpha-fold or similar protein structure prediction tools) assist researchers in locating viable binding pockets on newly discovered target proteins.
-
📂 Internship Task Completion Status Form:
After completing this task, interns must fill out the Internship Task Completion Status Form and upload their notes/practice work.
👉 Form Link:
https://forms.gle/dcfdpoDgksgP9KUY8
TASK IV: Lead Generation in Drug Discovery
🎯 Objective:
Explore the computational strategy of finding “lead” compounds—molecules that show specific activity against a target—and how high-throughput screening data feeds machine learning hit-to-lead prediction engines.
🎥 Session:
-
Session 4 – Lead generation in drug discovery and development.
-
Session Link: click here to access
📝 Task:
Write Notes On:
-
The transition from a “Hit” compound to a “Lead” candidate.
-
Core principles of Lipinski’s Rule of 5 and why it serves as a baseline filter for drug-likeness.
-
How classification algorithms (like Random Forests or Support Vector Machines) are trained on screening libraries to predict whether a molecule will be active or inactive.
🔬 Practice / Research:
-
Look into the concept of “privileged scaffolds” in medicinal chemistry and how machine learning generative frameworks use them to build intelligent libraries.
-
Describe how false positives (e.g., PAINS molecules) can disrupt screening datasets and how they are algorithmically filtered out.
📂 Assignment:
-
Task: Draft a workflow diagram representing an AI-driven Hit-to-Lead optimization pipeline. Show how hit compounds from a high-throughput screening database are passed through structural filters, predictive QSAR models, and toxicity screens to generate optimized lead derivatives.
📂 Internship Task Completion Status Form:
After completing this task, interns must fill out the Internship Task Completion Status Form and upload their notes/practice work.
👉 Form Link:
https://forms.gle/dcfdpoDgksgP9KUY8
TASK V: Molecular Methods in Drug Discovery & Development
🎯 Objective:
Analyze the essential biophysical and molecular methods utilized to evaluate drug-target interaction networks, structural configurations, and computational modeling simulations.
🎥 Session:
-
Session 5 – Molecular methods in drug discovery & development.
-
Session Link: click here to access
📝 Task:
Write Notes On:
-
Key structural biology methods supporting computer-aided drug design: X-ray Crystallography, NMR Spectroscopy, and Cryo-EM.
-
Principles of Molecular Dynamics (MD) simulations and how they model the physical movements of atoms within drug-target complexes over time.
-
The role of geometric deep learning on 3D molecular meshes and point clouds.
🔬 Practice / Research:
-
Explore the Protein Data Bank (RCSB PDB) website and note the typical file structure (.pdb format) used to store the 3D coordinates of co-crystallized drug-target complexes.
-
Investigate the computational cost differences between static molecular docking and dynamic molecular simulations.
📂 Assignment:
-
-
Task: Prepare a technical comparison matrix comparing X-ray Crystallography, Cryo-EM, and Computational Structural Prediction models. Focus on their limitations, output resolution, and how machine learning uses their experimental structural data to make automated predictions.
-
📂 Internship Task Completion Status Form:
After completing this task, interns must fill out the Internship Task Completion Status Form and upload their notes/practice work.
👉 Form Link:
https://forms.gle/dcfdpoDgksgP9KUY8
TASK VI: Biochemical and Cell-Based Assay Techniques
🎯 Objective:
Understand how biological assay techniques screen molecules for target binding and cellular efficacy, and study how these laboratory assays generate data for quantitative modeling pipelines.
🎥 Session:
-
Session 6 – Assay Techniques. Methods used in drug discovery.
-
Session Link: click here to access
📝 Task:
Write Notes On:
-
The distinction between biochemical assays (cell-free, looking at direct target binding) and phenotypic/cell-based assays (measuring physiological effects in living cells).
-
Key laboratory indicators: Half-maximal inhibitory concentration ($IC_{50}$), $EC_{50}$, and $K_d$ values.
-
How computer vision models automate the analysis of high-content cell imaging assays to identify compound-induced changes.
🔬 Practice / Research:
-
Research how High-Throughput Screening (HTS) data plates are set up and how $Z’$-factor analysis is calculated to evaluate assay quality.
-
Explain why a compound with excellent enzyme inhibition in a biochemical assay might completely fail in a cell-based assay.
📂 Assignment:
-
Task: Write a short analytical report (200–250 words) detailing how regression models use experimental $IC_{50}$ dose-response curves to train algorithms that predict the potency of completely unsynthesized, virtual chemical structures.
📂 Internship Task Completion Status Form:
After completing this task, interns must fill out the Internship Task Completion Status Form and upload their notes/practice work.
👉 Form Link:
https://forms.gle/dcfdpoDgksgP9KUY8
TASK VII: Ligand Binding Dynamics – Specific and Non-specific Binding
🎯 Objective:
Examine ligand binding assay parameters to mathematically distinguish between specific target engagement and non-specific background binding, a critical step for refining predictive docking scores.
🎥 Session:
-
Session 7 – Specific and Non-specific Binding in a ligand binding assay. Drug discovery…
-
Session Link: click here to access
📝 Task:
Write Notes On:
-
Specific binding (saturable, high affinity, occurring at the active site) versus Non-specific binding (unsaturable, low affinity, background noise).
-
Saturation binding curves and competitive inhibition assay mechanics.
-
How machine learning scoring functions in molecular docking try to approximate true binding affinity while penalizing non-specific interactions.
🔬 Practice / Research:
-
Review the mathematical formula for Scatchard plots or non-linear regression equations used to isolate specific binding values from total binding experiments.
-
Research how hydrophobic effects and electrostatic interactions contribute to non-specific compound accumulation on lipid membranes or unintended proteins.
📂 Assignment:
-
Task: Create a graph placeholder or sketch showing a standard radioligand binding assay experiment. Plot out three separate lines representing: Total Binding, Non-specific Binding, and calculated Specific Binding. Write a paragraph explaining how data points from these curves help refine automated computer-aided drug discovery algorithms.
📂 Internship Task Completion Status Form:
After completing this task, interns must fill out the Internship Task Completion Status Form and upload their notes/practice work.
👉 Form Link:
https://forms.gle/dcfdpoDgksgP9KUY8
⏱ Estimated completion time: 5 working days per portal.