Docking

1
Docking

• Where we are and where were going

2
Energy Landscape
Ligand in complex w/ protein (in solvent)
Ligand and protein separate (in solvent)
Internal Energy
Conformers
Docked Structure
Configuration
3
Configuration

• Chemical Shifts
• Tautomers
• PKa shifts

• Flexibility
• Protein
• Rigid
• Flexible side chains
• Full xyz flexibility
• Ligand
• Rigid
• Enumerated conformers
• Full xyz flexibility

• Secondary Molecules
• H2O placement

Configuration is the set of independent variables
an energy function takes
4
Sampling more better?
True Positive
False Positive
Sampled Configurations
Unsampled Configurations
Full Sampling
Limited Sampling
Limited Sampling
1 True Positive 1 False Positive
0 True Positives 1 False Positive
1 True Positive 4 False Positives
5
Energy Functions
Accuracy
Speed

• First Principal Approximations
• AM1
• PBSA

slow
good

• Semi Empirical
• MMFF

• Scoring Functions
• Chemscore
• ChemGauss
• PLP
• GSF

poor
fast
6
Staged Docking
Second Stage Energy Function
Energy
First Stage Energy Function
Configuration
7
Where we are now?
8
How are we doing?
Configuration Sampling
Scoring
Scoring functions do a pretty good job Juan
Alvarez
9
Self Rating
10
Where are we going?

• Configuration
• Protein flexibility
• PKa shifts upon binding
• Tautomers
• Energy Function
• Solvent
• Ligand free energy
• Free energy not Minimum energy

Dealt with to some extent currently
11
PKa(whered dat hydrogen go?)
12
How often are PKa shifts important?
13
Potential PKa shifts
14
Solvent(Isnt there some water around here?)

• MMFF/PBSA
• Merck VdW and intramolecular terms
• ZAP PB electrostatics
• Relatively smooth energy landscape
• Explicit Solvent
• Jagged energy landscape
• Difficult minimizations

15
Calculating Free Energy
Given an energy function and configuration space
how do we calculate free energy?
Statistical
Internal Energy
Mechanics
G A PV
Configuration
16
Minimum Energy
Internal Energy
Internal Energy
Configuration
Configuration
Typically only the lowest energy is retained
17
Numerical Integration
Internal Energy

• MD / MC
• Slow
• Longer time ? better sampling
• Exhaustive Search (FRED)
• Fast
• Rigid Sampling

Configuration
18
Normal Mode Analysis
Internal Energy
Internal Energy
Harmonic potential
Configuration
Configuration
Free energy of a set of uncoupled harmonic
oscillators
19
What goes into ?G of binding?
What are we trying to do?
?G Gcomplex Gprotein Gligand
20
Gligand

• Most scoring functions have crude approximations
  of Gligand if any!
• This should be easier than Gcomplex

Why do people think there is a large vs. large
problem?
G Actual Free Energy GC Calculated Free
Energy O constant error
Constant errors still cancel
21
Roadmap