Computational Biology

(BIOSC 1540)

Apr 1, 2025

Lecture 12A

Docking

Foundations

Announcements

Drug targets

After today, you should have a better understanding of

Proteins

Proteins are the primary targets of most small-molecule drugs

Over 90% of FDA-approved drugs act on proteins—enzymes, receptors, ion channels, and transporters.

These proteins play key roles in signaling, metabolism, immune response, and other vital functions.

Modulating protein activity with small molecules allows us to influence biological pathways precisely.

A protein is a good drug target when it is causally linked to disease

Not every protein is "druggable"—target selection must be biologically and chemically justified.

Genomics, proteomics, and phenotypic screens help identify candidate targets.

Criteria for Selecting a Protein Target

• Disease Relevance: The protein plays a critical role in the disease mechanism.
• Druggability: The target has a structure that allows it to bind with drug-like molecules.
• Specificity: Targeting the protein minimizes effects on healthy cells, reducing side effects.

Target-centric drug discovery begins with the biology, not the molecule

In target-centric approaches, researchers start with a well-understood protein and search for molecules that modulate its activity.

This contrasts with phenotypic screening, which starts with observed effects and works backward to find the target.

Target-centric methods allow rational design, structure-based modeling, and docking campaigns.

Drug targets

After today, you should have a better understanding of

Bruton’s tyrosine kinase (BTK)

Identifying the right protein target is crucial for developing effective and safe drugs

(Vetrie et al., 1993 & Tsukada et al., 1993)

Family-based genotyping of affected males revealed tight linkage between the disease and genetic markers.

Positional and sequence-based analyses identified BTK as the XLA gene

Bioinformatic comparison of the novel gene’s sequence to known kinases identified conserved domains with non-receptor tyrosine kinases.

In silico analysis of XLA patient sequences showed missense mutations affecting critical residues (e.g., Lys430 in the ATP-binding site).

(Aokl et el., 1994; Weers et al., 1994; Saouaf et al., 1994)

(Singh et al., 2018)

Tyrosine phosphorylation datasets and known B-cell signaling proteins placed BTK at a convergence point of multiple BCR-related pathways

Network-based analyses revealed BTK as a central hub in BCR signaling

Comparative genomics shows high conservation of BTK’s domains across vertebrates that are overrepresented among central signaling hubs.

Dihydrofolate Reductase (DHFR)

Drug targets

After today, you should have a better understanding of

THF production is crucial for cellular growth

Schnell et al., 2004

DHFR is responsible for synthesizing THF

Schnell et al., 2004

DHF

NADPH

DHFR has been extensively studied as an antibiotic (e.g., trimethoprim) and cancer (e.g., methotrexate) target

DHFR conservation complicates drug design

DHFR conservation complicates drug design

Bhabha et al., 2013

Bacteria and humans have similar structures, but their dynamics are different

Structure-based drug design

After today, you should have a better understanding of

We can modulate protein's function by binding small molecules to specific sites

A protein’s activity can be altered by binding a small molecule—often called a ligand—to a functional site on its surface.

This interaction can inhibit, activate, or subtly reshape the protein’s behavior.

These small molecules act like “molecular switches” that control protein action without altering the underlying gene.

Binding affinity and specificity determine a molecule’s effectiveness as a modulator

Not every small molecule that binds a protein is useful—effective modulation depends on how tightly and selectively it binds.

High-affinity binding ensures that a drug is effective at low doses, while specificity minimizes off-target effects.

These properties can often be optimized through structure-based design and screening campaigns.

Critical ligand interactions in DHFR

Chemical space contains an astronomical number of possible compounds to explore

Estimated to be between 10⁶⁰ to 10²⁰⁰ possible small organic molecules

We need methods to navigate chemical space and identify promising leads accurately and efficiently

High-throughput screening (HTS) allows testing of thousands of compounds against the target protein

• Library Preparation: Collection of diverse compounds
• Assay Development: Design of biological assays to measure compound activity against the target
• Screening: Compounds are tested in miniaturized assays
• Data Analysis: Identification of "hits" that show desired activity

Virtual screening evaluates vast libraries to identify potential leads efficiently

Thermodynamics of binding

After today, you should have a better understanding of

Selective binding to a protein is governed by thermodynamics (and kinetics)

Binding occurs when a compound/ligand interacts specifically with a protein

We can model this as a reversible protein-ligand binding

Gibbs free energy combines enthalpy and entropy

Entropy

Enthalpy

Accounts for energetic interactions

How much conformational flexibility changes

By predicting the free energy of binding, we can identify small molecules with high affinity to our drug target

Accurate and efficient binding predictions are essential

Objective: Directly predict binding affinity from protein and ligand structures with high accuracy and minimal computational resources.

Docking simplifies the binding free energy prediction problem to enhance speed

Identifying relevant protein conformations

After today, you should have a better understanding of

Accurate, reproducible docking requires a relevant protein conformation

Sources of Protein Conformational Data

Experimental Structure Selection Criteria

• Resolution and Quality
  • Prefer structures with higher resolution (e.g., <2.5 Å).
  • Assess reliability using R-factors and validation reports.
• Ligand-Bound vs. Apo Structures
  • Ligand-Bound (Holo) Structures
    • Provide direct insight into binding site conformation.
  • Apo Structures
    • May reveal binding site flexibility in the absence of ligands.
• Relevance to Target Ligand
  • Choose structures co-crystallized with ligands similar to those of interest.

Importance of Water Molecules

Role in Binding: Structured water molecules can mediate interactions between the protein and ligand.

Handling Water in Docking

• Some docking programs allow explicit water molecules in the binding site.
• Alternatively, consider their effect implicitly in scoring functions.

​Inclusion Criteria: Retain water molecules that are conserved across multiple crystal structures.

Binding pockets

After today, you should have a better understanding of

Types

Accurate Binding Pocket Detection is Crucial for Docking

The binding pocket is the specific region where a ligand interacts with a protein

Accurate identification of binding pockets is essential for successful docking and virtual screening.

Understanding Protein Surface Topography

Binding Pocket: A cavity that can accommodate a ligand.

Binding pockets are classified by location, accessibility, and regulatory function

Orthosteric sites are the natural binding sites of endogenous ligands or substrates

Cryptic Sites: Binding pockets not apparent in the unbound protein structure but form upon ligand binding or conformational change.

Allosteric sites are spatially distinct from the orthosteric site and modulate protein activity indirectly.

Binding pockets

After today, you should have a better understanding of

Detection

Alpha shapes detect pockets by reconstructing molecular surface topology

Alpha shapes extend the idea of a convex hull to capture the "shape" of a protein surface with cavities and tunnels.

Think of shrinking a sphere around atom centers—small alpha values allow more detailed surface features to be resolved.

Cavities enclosed by the alpha shape can be interpreted as binding pockets.

Alpha shape–based methods reveal pockets as topological voids

Proteins are modeled as a union of spheres (atoms), and the alpha shape filters through those to reveal cavities.

The algorithm identifies pocket volume, enclosure, and surface accessibility—critical for ligand fit.

These pockets are purely geometry-based, independent of electrostatics or residue type.

Grid-based methods map the protein onto a 3D lattice to identify cavities

The protein is embedded in a 3D grid, and each voxel is labeled as protein, solvent, or cavity.

Grid points near the surface but not occupied by protein atoms are clustered into potential pockets.

Detecting Cryptic Binding Sites

Scoring functions

After today, you should have a better understanding of

Scoring functions estimate how well a ligand binds to a protein

Scoring functions assign a numerical value to a protein-ligand pose, reflecting its favorability.

Scoring functions attempt to approximate this balance using simplified models.

The best pose maximizes favorable interactions (e.g., hydrogen bonds, hydrophobic packing) and minimizes clashes or strain.

A good scoring function may rank correctly even if absolute energies are wrong.

Force-field-based scoring calculates interactions from first principles

They consider van der Waals attraction/repulsion, electrostatics, and sometimes torsional strain.

Advances in scoring functions are driven by data and AI

Machine learning models (e.g., RF-Score, DeepDock) are trained directly on binding data.

These models learn nonlinear relationships and use structural and chemical features.

Require large, high-quality datasets and careful validation.

Ligand pose optimization

After today, you should have a better understanding of

Accurate Docking Depends on Optimized Ligand Poses

Docking needs to generate diverse conformations

Search strategies

• Systematic
• Stochastic
• Empirical
• Machine learning

Systematic searches numerically iterate over all possible conformations

Identify important degrees of freedom

• Angles
• Dihedrals

Scan along each angle with a step size of a N degrees

Remove structures with high strain

Systematic searches are only possible for very small molecules

How many different conformations would we have in this molecule if we scanned only dihedrals every 45 degrees?

Systematic searches are only possible for very small molecules

8 dihedrals

1

2

3

4

5

6

7

8

8 angles

8 × 8 × 8 × 8 × 8 × 8 × 8 × 8 = 16,777,216

That's a lot of structures, and many of them will clash!

We almost never do a systematic search in practice without some precautions to combinatorics

Stochastic algorithms provide better balance of sampling and cost

We also have conformer libraries

Use pre-generated libraries