Molmodel Workshop

March 21, 2008Christopher Bruns

Schedule

• 8:30-10:00 Molmodel examples• 10:00-10:15 Break• 10:15-11:30 More Molmodel examples• 11:30 Lunch• 12:00 Your projects

What is Molmodel?

• Domain-specific modeling layer on Simbody• Part of SimTK core libraries• You should have Programmer’s Guide

SimTK Molmodel molecular modeling

SimTK Simbody order(n) multibodydynamics

SimTK Lapack linear algebra

SimTK Common Vector, Matrix classesand other data structures

SimTK Math numerical methodsand linear algebra

Principal SimTK Core Libraries

Molecular Physical Units

• Length: nanometers (10-9 meters)• Time: picoseconds (10-12 seconds)• Mass: atomic mass units

(1/(Avogado’sConstant) grams)

Argon

• Inert “noble” gas• Has no chemistry• van der Waals forces only

Molecular Force Field

• What is a molecular force field?– A specification of all forces affecting a

dynamic simulation. For conventional molecular simulation this includes non-bonded electrostatic forces and van der Waals forces, and bonded stretch, bend, torsion, and improper torsion parameters

Biotype

• Bridge between structural model and force field implementation

• Concept borrowed from TINKER molecular dynamics package

Argon Exercises

• Compile and run TwoArgons example program• Add a third argon atom.• What happens when atoms start too close

together? Too far apart?• Increase the attractive force between atoms

ten-fold

Ethane

• Two carbons, six hydrogens• One dihedral degree of freedom• 3 kcal/mol energy barrier to rotation

Internal coordinates• Instead of explicit Cartesian X, Y, Z

coordinates for an atom• Given three other atom positions, one

distance, one bond angle, and one dihedral angle

?

2 3

1

1 distance

bond angle 23 dihedral angle

Molecule internal coordinates

• Bond lengths, bond angles, dihedral angles• Mandatory for Compound construction• Optional for Compound simulation

Ethane Exercises

• Compile and run TwoEthanes example program

• Add a third ethane molecule

Macromolecules, residues, and atoms

• RNA, DNA, and protein are large molecules called macromolecules

• Macromolecules consist of chains of residues

Atoms

Residues}Macromolecule

Top level API for constructing proteins

#include “SimTKmolmodel.h”Protein(“ACDEFGHIKLMNPQRSTVWY”);

Simple protein

• Construct protein from sequence• All atom parameters are built into AMBER99

force field• Note end caps

Protein Exercises

• Compile and run SimpleProtein example program

• Try a different sequence• Add a second protein

Simple RNA

• RNA from sequence• Writing PDB files

RNA Exercises

• Compile and run SimpleRNA example program• Try a different sequence• Change the frequency of PDB writing

PDB: The Protein Data Bank

• http://www.rcsb.org/pdb/• Repository of DNA, RNA, and protein atomic

structures• Contains experimental results, not perfect

models• Entries identified by 4 character ID, e.g.

“1MRP”, “1GRZ”• Use “advanced search” at PDB web site to find

structures.

Search box

4-character ID

PDB File format

• Text format – humans can read it too• Most of the information is atomic coordinates• PDB files usually contain multiple molecules,

including water molecules

…ATOM 1751 N GLY C 250 32.286 1.882 43.206 1.00 22.00ATOM 1752 CA GLY C 250 32.365 1.086 41.969 1.00 21.39ATOM 1753 C GLY C 250 31.538 1.735 40.864 1.00 20.79ATOM 1754 O GLY C 250 30.621 2.527 41.152 1.00 21.58…

LoadPdb Exercises

• Download 1AKG from Protein Data Bank http://www.rcsb.org/pdb

• Simulate 1AKG using Example

Coarse-grained representations in biology

Natural lump sizes:– atoms– residues– base pairs– duplexes, alpha helices,

beta strands– macromolecules– molecular assemblies– organelles

– cells– tissues– organs– limbs– organisms– societies– ecosystems– biospheres– intergalactic federations

(picometer to micron scale) (micron to mega-light-year scale)

Coarse-grained simulation of atom-based molecular models

• Full atom force field• coarse-grained motion permits longer

integration time steps• Increasing coarseness

– full Cartesian – torsion only (internal coordinate)– full rigid

Molecule Mobility/Partitioning

• What is molecule partitioning?– defining the movable degrees of freedom

that will be used in a molecule simulation

• Examples of partitioning strategies– Internal coordinates (fixed bond lengths and

fixed bond angles)– Full Cartesian (conventional molecular

dynamics)– Rigid molecules

Protein Internal Coordinates

Nucleic Acid Internal Coordinates

RigidProtein Exercises

• Simulate rigid protein• Create second protein• Convert to full Cartesian model• Which way is faster?

Custom Molecule Construction

• BondCenters, Atoms, Bonds, Compounds• Inboard bond center• The first few atoms• Ring closing bonds

Molecule Construction nomenclature1: Atoms(3)

3: Bonds(2)

2: BondCenters(9)

4: Compound(1)

Submitting Bug Reports