IN SILICO MOLECULAR DESIGN

DEPARTMENT OF CHEMISTRY
BASIC SCIENCES

1
 SYLLABUS

Introduction to Drug Discovery and Design, Molecular Biology, Fundamentals of Medicinal

Chemistry and Pharmacology, Molecular Modelling and Docking, Molecular Representation
and Visualization, Structure- and Ligand-Based Drug Design, AI and ML in Drug Design, Data
Engineering in CADD, AI-Driven Case Studies and Future Trends.

2
Hour Topic
1 Introduction to Drug Discovery and Design
2 Molecular Biology
3 Fundamentals of Medicinal Chemistry and Pharmacology
4 Molecular Modelling and Docking
5 Molecular Representation and Visualization
6 Structure-Based and Ligand-Based Drug Design
7 AI and ML in Drug Design & Data Engineering in CADD
8 AI-Driven Case Studies and Future Trends
 How Drugs Work

Drug : A substance used in the diagnosis, treatment, or prevention of a disease or as a component

of a medication.

Introduction to Drug Discovery and Design

 The goal of Drug Discovery
Identify a drug candidate with unique profile to treat unmet needs of a disease
The drug candidate:

✔ Designed to meet unmet medical needs.

✔ Has the right properties to work effectively at a practical dosage.
✔ Provides benefit without causing excessive side effects.
✔ Works consistently across different patient groups.
✔ Can be patented.
✔ Either a first-of-its-kind drug or competitive with others in development.

Introduction to Drug Discovery and Design

Drug Development Process

Find compounds effective

against disease protein File IND
Disease FDA approval
(2-5 years) Animal toxicity
(2-3 years)

Isolate protein Preclinical testing Human clinical trials

involved in disease Animal efficacy Phase 1
(2-5 years) (1-3 years) Phase 2
Phase 3
(2-10 years)

Introduction to Drug Discovery and Design

Drug Development Process

You have to test A LOT of

compounds to find a drug

10 – 15 years,
500 - 1000 million US $

Introduction to Drug Discovery and Design

Molecular Biology
DNA, RNA, and Genes
•DNA (Deoxyribonucleic Acid): Blueprint of life, contains instructions for proteins.
•RNA (Ribonucleic Acid): Messenger that carries DNA instructions to ribosomes.
•Genes: Segments of DNA that code for specific proteins.
•Diseases like cancer, sickle-cell anemia, and cystic fibrosis are caused by gene
mutations.
•Understanding genes helps design targeted drugs (e.g., cancer gene therapy).
Central Dogma Definition
“Central dogma is the process in which the genetic information
flows from DNA to RNA, to make a functional product protein.”

DNA → RNA → Protein

Molecular Biology
DNA is the genetic material within the nucleus.

Replication creates new copies of DNA.

Transcription creates an RNA using DNA information.

Translation creates a protein using RNA information.

Molecular Biology
Amino Acids – Building Blocks of Proteins

Definition: Small molecules that link together to form proteins via peptide bonds.
Structure:
• Central α-carbon
• Amino group (–NH₂)
• Carboxyl group (–COOH)
• Hydrogen atom (H)
• R-group (side chain) – defines the amino acid type and properties.
Types of Amino Acids (Based on R-group):
• Nonpolar / Hydrophobic: Glycine, Alanine, Leucine
• Polar / Hydrophilic: Serine, Threonine
• Acidic: Aspartic acid, Glutamic acid
• Basic: Lysine, Arginine
• Special: Cysteine (forms disulfide bonds), Proline (rigid ring)
• 11 non- essential, 9-essential
Proteins & Enzymes

Proteins: Perform key functions like transport, defense

(antibodies), signaling (hormones/receptors).
Enzymes: Proteins that act as biological catalysts; speed
up reactions without being consumed.
Drug Targets:
• Enzymes (e.g., Aspirin inhibits COX enzyme →
reduces pain/inflammation).
• Receptors (e.g., β-blockers bind heart receptors →
lower blood pressure).
• Ion channels, transporters, structural proteins.
Ligands & Binding Sites

•Ligand: Molecule that binds to a target protein.

• Natural ligands: Hormones, neurotransmitters,
substrates.
• Synthetic ligands: Drugs, inhibitors, activators.
•Binding Site / Active Site:
• Specific pocket or cavity on the protein where
the ligand fits.
• Determines selectivity and strength of binding.
•Types of Binding Interactions:
• Strong Interactions:
• Hydrogen bonds, ionic (electrostatic) bonds
→ high specificity.
• Weak Interactions:
• Hydrophobic forces, van der Waals forces →
help stabilize the complex.
Fundamentals of Medicinal Chemistry and Pharmacology
PHARMACOLOGY
Drug: “ A drug is any chemical molecule, which can be administered exogenously into
the living system produce biological response or effect.”

Pharmacology is the science of drugs (Greek: Pharmacon—drug; logos—discourse in).

Pharmacology

Pharmacodynamics Pharmacokinetics
(Greek: dynamics—power) (Greek: Kinesis—movement)

—What the drug does to the body. —What the body does to the drug.
Fundamentals of Medicinal Chemistry and Pharmacology
PHARMACOKINETICS

Pharmacokinetics is the quantitative study

of drug movement in, through and out of the
body.

A – Administration

D – Distribution

M - Metabolism(Biotransformation)

E - Excretion

Fundamentals of Medicinal Chemistry and Pharmacology

Lipinski's rule of five
The rule was formulated by Christopher A. Lipinski in 1997, based on the observation that
most medication drugs are relatively small and lipophilic molecules [Lipinski et al. 1997,
2001 & 2004].

✔ Not more than 5 hydrogen bond donors (nitrogen or oxygen atoms with one or more
hydrogen atoms)
✔ Not more than 10 hydrogen bond acceptors (nitrogen or oxygen
atoms)
✔ A molecular mass less than 500 daltons
✔ An octanol-water partition coefficient log P not greater than 5

Fundamentals of Medicinal Chemistry and Pharmacology

C13H18O2

Cyclosporine

C62H111N11O12
How this rule benefits DRUG Design?

The rule describes molecular properties important for the pharmacokinetics of a drug in
the human body, including their absorption, distribution, metabolism, and excretion
(“ADME”). However, the rule does not predict if a compound is pharmacologically
active.

This rule helps Pharmaceutics/Industrial Pharmacy students in proper selection of

the drug and knowing whether the drug is suitable for oral formulations.

For Medicinal chemistry students involved in drug designing, CADD, understanding

this rule will help you a lot in designing suitable homologues of rugs and fine tuning
your drug with suitable modifications

Fundamentals of Medicinal Chemistry and Pharmacology

 This seems like a lot to remember!

There are various guidelines to help, the most well-known of which is the
Lipinski Rule of Five

➢ molecular weight < 500

➢ logP < 5
➢ < 5 H-bond donors (sum of NH and OH)
➢ < 10 H-bond acceptors (sum of N and O)

An additional rule was proposed by Veber

➢ < 10 rotatable bonds

Otherwise absorption and bioavailability are likely to be poor. NB This is for oral
drugs only. Fundamentals of Medicinal Chemistry and Pharmacology
 PHARMACODYNAMICS

Pharmacodynamics is the study of drug effects.

• It describes, what the drugs do and how they do.

“what the drug does to the body when they enter”

• Drugs (except those gene based) do not impart new functions to any system, organ
or cell; they only alter the pace of ongoing activity

Principles of Drug Action:

• Stimulation
• Depression
• Irritation
• Replacement
• Cytotoxic action Fundamentals of Medicinal Chemistry and Pharmacology
 PHARMACODYNAMICS

Stimulation:
It refers to selective enhancement of the level of activity of specialized cells,
• E.g. adrenaline stimulates heart, pilocarpine stimulates salivary glands.
• However, excessive stimulation is often followed by depression

Depression:
It means selective diminution of activity of specialized cells,
E.g. barbiturates depress CNS, quinidine depresses heart, omeprazole depresses gastric
acid secretion

Fundamentals of Medicinal Chemistry and Pharmacology

 PHARMACODYNAMICS

Irritation:
Strong irritation results in inflammation, corrosion, necrosis and morphological damage.

This may result in diminution or loss of function.

Replacement:
This refers to the use of natural metabolites, hormones or their congeners in deficiency
states,

E.g. levodopa in parkinsonism, insulin in diabetes mellitus, iron in anaemia.

Fundamentals of Medicinal Chemistry and Pharmacology

 PHARMACODYNAMICS

Cytotoxic action:

Selective cytotoxic action on invading parasites or cancer cells, attenuating them

without significantly affecting the host cells is utilized for the cure/ palliation of
infections and neoplasms.

* E.g. penicillin, chloroquine, cyclophosphamide, zidovudine, etc.

Fundamentals of Medicinal Chemistry and Pharmacology

Molecular Modelling and docking
Molecular Modelling

Molecular modelling is a theoretical and computational method used to

mimic molecular behavior.

Aims to visualize molecules, discover new drug compounds, and simulate

properties at the atomic level.

Key goal is to develop sufficiently accurate models to replace physical

experiments.

Molecular Modelling and Docking

 Molecular Modelling

• Molecular docking is a process used to predict how a small molecule (ligand) binds to a specific site
on a protein (receptor).
• The interaction between the ligand and protein can involve different forces such as Van der Waals,
electrostatic, and hydrogen bonding.
• Docking methods work by searching different possible positions and shapes of the ligand in the
binding site and then using a scoring function to find the best fit.
• This technique is widely used in drug design because it helps scientists predict how well a drug
molecule might bind to its target.
• Understanding this binding behavior is important for designing new medicines and for studying
how biological processes work.

Molecular Modelling and Docking

Drug Design Strategies

A. Structure-Based Drug Design (SBDD)

•In this approach, the 3D structure of the target protein (receptor or enzyme) is known.
•The drug (ligand) is designed to fit precisely into the target’s active or binding site.
•Tools like molecular docking and computer simulations are used.
•Example: Designing HIV protease inhibitors using the 3D structure of the HIV enzyme.
B. Ligand-Based Drug Design (LBDD)
•Used when the structure of the target protein is not known.
•Relies on information from known active molecules (ligands) to design new ones.
•Techniques include QSAR (Quantitative Structure–Activity Relationship) and pharmacophore
modeling.
C. De Novo Drug Design
•Drugs are designed from scratch using computational tools to build new
molecules that could bind effectively to the target.
D. Fragment-Based Drug Design (FBDD)
•Small chemical fragments that bind weakly to the target are identified and then
linked or grown into more potent compounds.
E. Computer-Aided Drug Design (CADD)
•Uses computational methods like molecular modelling, docking, and
simulations to speed up and improve the design process.
Molecular Dynamics
• Molecular Dynamics (MD) is a computer simulation method used to study the
movement of atoms and molecules over time.
• It is based on Newton’s laws of motion, where the position and velocity of each
atom are calculated step by step.
• MD uses mathematical equations called force fields to describe the forces
between atoms (bonded and non-bonded interactions).
• It helps in understanding molecular behavior such as protein folding, drug
binding, and stability of molecules.
• Widely used in drug design, biomolecular research, and material science to
observe dynamic molecular changes.
There are two approaches in Molecular dynamics
1. Deterministic Approach
•In this method, the motion of atoms is completely determined by Newton’s laws of motion.
•Once the initial positions and velocities of atoms are known, their future motion can be predicted
exactly (no randomness).
•Example: Classical Molecular Dynamics where trajectories are calculated precisely using force
fields.
2. Stochastic Approach
•This method adds an element of randomness to account for effects like temperature and collisions
with surrounding particles.
•It uses random forces (noise) and friction terms to simulate more realistic environments.
•Example: Langevin Dynamics and Monte Carlo simulations.

Molecular Modelling and Docking

 Energy Minimization
It is systematic modification of the atomic coordinates of the model resulting in a
3D arrangements of atoms in the model representing an energy minimum (a stable
molecular geometry to be found without crossing a conformational energy barrier) is
called energy minimization and geometry optimization.

EM is used to
Locating a stable conformation
Locating global &local energy minima
Locating saddle point

Molecular Modelling and Docking

Why energy minimization ?
1. Comparison of Structures/Properties
checking how two or more molecules are similar or different in their shape, size, or chemical features.
2. Template Forcing
A method where a known structure (template) is used as a guide to build or modify another molecule to
have a similar shape or function.
3. Systematic Mapping of Space
This involves exploring all possible positions and shapes a molecule can take in a specific region (like a
binding site) to find the best fit.
4. Binding Energies
It is the energy released or required when a ligand binds to a protein. Lower (more negative) energy
means a stronger and more stable interaction.
5. Docking
A computer method used to predict how a small molecule (drug) fits into the binding site of a protein
and how strong the interaction is.
6. Harmonic Analysis
A technique used to study small vibrations or movements of atoms around their stable positions in a
molecule.
7. Comparing/Fitting Force Fields
This means testing different mathematical models (force fields) to see which one best describes the
Molecular Modelling and Docking
forces and interactions between atoms in a system.
 What is docking?
Docking is finding the binding geometry of two interacting molecules
with known structures

The two molecules (“Receptor" and “Ligand") can be:

-two proteins
-a protein and a drug
-a nucleic acid and a drug

Two types of docking:

- local docking: the binding site in the receptor is known, and

docking refers to finding the position of the ligand in
that binding site
- global docking: the binding site is unknown. The search
for the binding site and the position of the ligand in the binding site
can then be performed sequentially or simulaneously
Molecular Modelling and Docking
Docking Workflow

Molecular Modelling and Docking

 Types of docking

Lock and Key / Rigid Docking :

In rigid docking, both the internal
geometry of the receptor and ligand is
kept fixed and docking is performed.

Induced fit/ Flexible Docking :

An enumeration on the rotations of
one of the molecules usually smaller
one) is performed. Every rotation the
surface cell occupancy and energy is
calculated; later the most optimum
pose is selected.

Molecular Modelling and Docking

 Scoring Function
A scoring function is a mathematical formula used in molecular docking to measure how well a ligand
binds to a protein.
[Link] gives a numerical score to each possible ligand–protein complex.
[Link] score helps rank the different binding poses — lower (more negative) scores usually mean
stronger binding.
[Link] considers factors like hydrogen bonds, van der Waals forces, electrostatic interactions, and binding
energy.
[Link] goal is to find the best fit between the drug and its target protein.
[Link] functions are widely used in drug design to predict which compounds might work best before
lab testing.
 Scoring Function

A scoring function is like a grading system that tells us how well a drug (ligand) fits
into its target protein.
• A mathematical tool used in docking.
• Predicts how well a drug (ligand) binds to a target protein.
• Acts like a “score card” to rank drug candidates.
• Higher (or lower) score = better fit and more stable binding.

Molecular Modelling and Docking

EXAMPLE

Interpretation:
•Docking Score: Compound a has the most negative value (-6.581), so it has the best docking score
(strongest predicted binding).
•Binding Energy: Compound b has the most negative binding energy (-57.159), suggesting it forms the
most stable complex energetically.
Conclusion:
•If you prioritize docking score, compound a is best.
•If you prioritize binding energy, compound b is best.
•Overall, compound b could be considered the most promising since it has very favorable binding energy
while maintaining a good docking score.
Key Elements of Scoring Function
1. Shape Complementarity
The drug must fit into the protein’s binding pocket like a key in a lock. A good fit means
better chances of binding.
2. Binding Interactions
Strong attractions such as hydrogen bonds, hydrophobic forces, and electrostatic interactions
help the drug stay bound to the protein.
3. Energy Stability
Stable drug–protein complexes are low in energy. The lower the binding energy, the stronger
and more reliable the interaction.
4. Penalties
If the ligand clashes with the protein or does not fit properly, penalty points are applied,
reducing the overall score.
Example: Docking of Indinavir (HIV Protease Inhibitor)

Shape Complementarity:
Indinavir fits into the catalytic cleft of HIV-1 protease, occupying the subsites with high
steric complementarity.
Binding Interactions:
It forms hydrogen bonds with the Asp25/Asp25’, while hydrophobic moieties interact with
Val82 and Ile84 side chains. Electrostatic complementarity further stabilizes the complex.
Energy Stability:
The calculated binding free energy (ΔGbind) is highly negative, reflecting a strong and
stable drug–target interaction.
Penalties:
If bulky substitutions at the ligand’s P2 position cause steric clashes with the protease flap
residues, the docking score decreases, predicting weaker inhibition.
Molecular Representation and Visualization
Protein structure visualization tools are software applications that allow users to view, manipulate, and analyze the
three-dimensional (3D) structures of proteins and other biomolecules.

These tools are essential for understanding the molecular architecture, function, and interactions of proteins, as well
as for designing new drugs, vaccines, and biotechnologies.

There are many protein structure visualization tools available, each with different features, capabilities, and
limitations.

Some of them are standalone applications that can be installed on a local computer, while others are web-based
applications that can be accessed through a browser.

Some of them require additional plugins or applets to run, while others use HTML5 or JavaScript technologies that do
not need any installation.

Some of them are specialized for specific tasks, such as structure prediction, docking, or simulation, while others are
more
Molecular Representation
•Molecules can be represented in different formats to
capture structure and properties.
•2D Representation: Atoms and bonds drawn on a flat
plane (useful for chemical formulas, reaction schemes).
•3D Representation: Shows spatial arrangement of atoms
(crucial for docking and drug design).
•Coordinate Files: Molecules stored as Cartesian 2D
3D
coordinates (e.g., PDB, MOL2, SDF files).
•Simplified Notations: SMILES and InChI strings
represent molecules in text format for databases and
computational tools.
3D 2D
Molecular Representation and Visualization
Types of Molecular Visualization
•Wireframe Model: Shows bonds as lines;
simple but lacks detail.
•Stick/Ball-and-Stick Model: Atoms shown as
spheres, bonds as sticks; clear visualization of Wireframe Stick/Ball-and-Stic
k Model
connectivity.
•Space-Filling Model: Atoms shown with van
der Waals radii; highlights molecular size and
packing.
•Ribbon/Cartoon Model (for proteins): Shows
Space-Filling
secondary structures (α-helices, β-sheets) for
clarity.
Ribbon/Cartoon
•Surface Representation: Visualizes molecular Model
surface, binding pockets, and electrostatic
potential.
Surface Representation
Molecular Graphics Software – RCSB PDB

NCBI Structure Viewer

PyMOL
Molecular Graphics Software – RCSB PDB

• The RCSB Protein Data Bank (RCSB PDB) is a repository of experimentally determined 3D structures of proteins,
nucleic acids, and complex assemblies.

• It also provides access to computed structure models (CSMs) from AlphaFold DB and ModelArchive.

• The RCSB PDB website offers a variety of molecular graphics software for browsing, viewing, and downloading
protein structures and CSMs

Software :BioBlender, BioViz Studio, CCP4mg, Chimera, ChimeraX, Cn3D, CrystalMaker, Epmv, EzMol, Foldit,
ICM-Browser, IcmJS, iMol
NCBI Structure Viewer

• The NCBI Structure Viewer is a web-based application that allows users to view and manipulate 3D structures of
proteins and other biomolecules from the NCBI databases, such as Entrez, MMDB, PubChem, and CDD.

• The NCBI Structure Viewer uses the JSmol library, which is a JavaScript adaptation of the Jmol molecular graphics
software.

• The NCBI Structure Viewer does not require any installation or plugin, and works with most modern browsers.

• To use the NCBI Structure Viewer, users can go to the NCBI Structure Home Page and enter the PDB ID or other
query in the search box and press the Go button.

• Then, they can click on a structure image to access its record page and scroll to the molecular graphic section.

• There, they can click on the spin icon to load an interactive view of the structure within the web page.

• Users can also download the structure file in various formats and open it with the NCBI Structure Viewer or other
software on their local computer
PyMOL
▪ PyMOL is a popular and powerful molecular graphics software for viewing, editing, and analyzing 3D structures of
proteins and other biomolecules.

▪ PyMOL is written in Python and C, and can be extended and customized with Python scripts and plugins.

▪ PyMOL is available for Windows, Linux, and Mac OS X, and can be downloaded for free for educational and
non-commercial use, or purchased for commercial use.

▪ To use PyMOL, users can download and install the software on their local computer, and then open it and load a
structure file from their local disk or from the internet.

▪ Users can also use the PyMOL command line or graphical user interface to manipulate and modify the structure,
such as changing the display mode, color, style, size, orientation, lighting, and transparency.

▪ Users can also perform various calculations and analyses on the structure, such as measuring distances, angles,
dihedrals, contacts, surface area, volume, electrostatics, and hydrogen bonds.

▪ Users can also create animations, movies, and ray-traced images of the structure, and export them in various
formats.
Structure-Based and Ligand-Based Drug Design
Computer-Aided Drug Design (CADD): Use of computational methods to design and
evaluate drug molecules.
Ligand: A small molecule (drug candidate) that binds to a target protein.
Protein (Target): Biological macromolecule (enzyme, receptor) responsible for disease
pathways.
Docking: Process of fitting a ligand into the binding site of a protein to predict binding
mode.
Scoring Function: Mathematical model that evaluates how well the ligand binds (like a
score card).
Types of CADD Approaches

Structure-Based Drug Design (SBDD): Uses 3D structure of the target

protein.
Ligand-Based Drug Design (LBDD): Uses knowledge of active/inactive
molecules when protein structure is unknown.
ADMET Prediction: Computational evaluation of Absorption, Distribution,
Metabolism, Excretion, Toxicity properties of drug molecules.
Methodologies and strategies of CADD

Structure-Based and Ligand-Based Drug Design

Structure-Based Drug Design (SBDD)

Definition: Designing ligands using the 3D structure of the protein target.

Importance: Helps identify molecules that can specifically fit into protein binding sites.
Process:
• Obtain protein structure (X-ray, NMR, Cryo-EM, homology modeling).
• Identify active/binding site (lock).
• Dock ligands (keys) into site.
• Use scoring functions to rank best ligands.
•Advantages:
• Directly targets protein site (high specificity).
• Helps in novel/first-in-class drug design.
• Visualizes atomic-level interactions.
•Limitations:
• Needs accurate protein structure.
• Protein flexibility may reduce accuracy.
• Computationally intensive.
Ligand-Based Drug Design (LBDD)
Definition: Designing new ligands based on known active molecules when protein
structure is not available.
Importance: Useful when the target structure is unknown or difficult to determine.
Process:
• Collect known active ligands.
• Identify common features (pharmacophore).
• Develop QSAR models (structure–activity relationship).
• Design/test new molecules with similar features.
Advantages:
• Works without protein structure.
• Faster and simpler than SBDD.
• Can predict activity trends from large ligand datasets.
Limitations:
• Requires good dataset of active ligands.
• Predictions may fail if dataset is limited or biased.
• Less precise than SBDD for binding mode.
Structure-Based vs. Ligand-Based Drug Design

Feature Structure-Based Drug Design (SBDD) Ligand-Based Drug Design (LBDD)

Requires knowledge of active molecules
Starting Point Requires 3D structure of the target protein
(ligands)
“You know the lock, design the right “You don’t see the lock, but you study
Analogy
key.” working keys.”
Protein binding site structure (X-ray, Chemical structures & activities of known
Main Data Used
NMR, cryo-EM, homology model) ligands
Molecular Docking, Scoring Functions, QSAR, Pharmacophore Modeling, Similarity
Techniques
Molecular Dynamics Search
Directly designs molecules that fit
Useful when protein structure is unknown;
Advantages binding site; high accuracy if structure is
quick predictions
good
Needs accurate protein structure; may fail Depends heavily on quality and quantity of
Limitations
if binding site is flexible known ligand data
Examples HIV protease inhibitors, kinase inhibitors NSAIDs (aspirin → ibuprofen), β-blockers
ADMET Prediction
Definition: Computational prediction of a drug candidate’s Absorption, Distribution,
Metabolism, Excretion, and Toxicity.
Importance:
• Ensures drug reaches target site in body at effective concentration.
• Filters out molecules likely to fail in clinical trials due to poor safety/PK.
• Saves time and cost before lab experiments.
Tools:
SwissADME → Predicts absorption, lipophilicity, drug-likeness.
pkCSM → Predicts pharmacokinetic and toxicity profiles.
ADMETlab → Integrated ADMET property prediction.
Applications:
• Check if molecule obeys Lipinski’s Rule of Five (drug-likeness).
• Predict blood–brain barrier penetration, solubility, metabolism (CYP450 enzymes).
• Screen for toxicity (hepatotoxicity, cardiotoxicity, mutagenicity).
ADMET evaluation

Data mining….

Schrodinger
MOE
MedChem studio
Structure-Based
….. and Ligand-Based Drug Design
Available CADD software
[Link]
 Big data and AI methods
in the drug discovery process
Case Studies: Exploring AI in Drug Discovery
Objective:
Understand how computers and AI are used to speed up drug discovery.
Instructions:
1. Read a short summary about AI in drug discovery (use the 5 case studies we simplified).
2. Pick one case study you find most interesting.
3. Answer these 3 questions:
o What was the problem the scientists were trying to solve?
o How did AI or big data help solve it?
o Why is this important for making new drugs?
4. Optional activity: Draw a simple diagram showing “AI → Drug Discovery → Faster
Medicine.”
List of case studies:

Big Data and AI in Drug Discovery – Simple Case Studies

1. BenevolentAI – Fighting COVID-19

∙ AI looked at huge amounts of medical data to find existing drugs that could work against
COVID-19.

∙ Found baricitinib, a drug for arthritis, as a possible treatment in just a few days.

∙ Showed how AI can make drug discovery much faster.

2. DeepMind AlphaFold – Protein Shapes

∙ AI predicts the 3D shape of proteins from their sequence of amino acids.

∙ Knowing protein shapes helps scientists design new drugs more quickly.
3. Berg Pharma – Cancer Drugs

∙ AI analyzed data from over 1,000 patients to find promising cancer drugs.

∙ Helped develop BPM 31510, a drug now in clinical trials.

4. Iambic Therapeutics – Predicting Drug Success

∙ AI predicts which drug candidates are likely to work well and be safe.

∙ Reduces the risk of expensive failures in drug development.

5. Oxford Alzheimer’s Research

∙ AI helped identify 54 genes that could be drug targets for Alzheimer’s.

∙ Made research faster and more efficient by handling large amounts of data.
Thank you