An Overview of Information Flow in AMBER
Overview
Teaching: 30 min
Exercises: 5 minQuestions
What files are needed to run an MD simulation with AMBER?
Objectives
Learn what types of files are needed to run an MD simulation with AMBER
Introduction
In this lesson we will go through the steps of setting up a fully solvated protein system for simulation with AMBER/NAMD/OPENMM. There are many commercial programs and interactive graphical interfaces designed to assist with system preparation. While these tools are easy to use and don’t require as much learning efforts as command line tools, they are offer only a limited functionality, and most importantly results obtained with WEB/GUI tools are not reproducible and prone to human error. When a user needs to generate multiple systems with different proteins or membrane compositions or requires different starting configurations, relying on GUI/WEB interface becomes a challenge, as the process becomes time-consuming. Therefore, we will focus on system preparation using only scriptable command line driven tools. The emphasis of this lesson is to expose you to the various methods that can be used to create a reproducible molecular modeling workflow by automating preparation and simulation steps. One of the advantages of such approach is that once a workflow script have been developed it can be easily modified for other systems or conditions (for example if an updated version of pdb file is released, you can prepare a new simulation system with a single click).
Information flow in AMBER
Simulation workflow in MD usually involves three steps: system preparation, simulation, and analysis. Let’s take a closer look at these steps.
coordinates| C{TLEAP,
XLEAP} B([FF files]) --> |Load
parameters|C H([LEaP commands])-->|Build
simulation system|C end subgraph "Run MD" C-->|Save
topology|D(["prmtop"]) C==>|Save
coordinates|E(["inpcrd"]) E==>|Load
coordinates|F Q([disang])-.->|Load
NMR restraints|F T([groupfile])-.->|Setup
multiple
simulations|F G(["mdin"])-->|Load
run options|F D-->|Load
topology|F{SANDER,
PMEMD} F==>|Save
restart|P([restrt]) P==>|Load
restart|F end subgraph Analyze N(["CPPTRAJ
commands"])-->|Run
analysis|J F===>|Save
trajectory|I([mdcrd, mdvel]) F--->|Print
energies|L([mdout,
mdinfo]) L-->|Load
energies|J I==>|Load
frames|J{CPPTRAJ,
PYTRAJ} I==>|Load
frames|K{MMPBSA} R([MM-PBSA
commands])-.->|Compute
PB energy|K J-->S[Results] K--->S((Results)) end D---->|Load
topology|J linkStyle 1,3,9,21 fill:none,stroke:blue,stroke-width:3px; linkStyle 2,8,12 fill:none,stroke:red,stroke-width:3px;
System Preparation
AMBER package includes two utilities for simulation preparation: tLEaP and xLEaP. The command line version (tLeap) is very efficient for executing scripts automatically. The graphical version (xLEaP) is ideal for building simulation systems interactively and visualizing them. The first step in creating a simulation is to load coordinates in PDB format and a Force Field of your choice. Adding ions, setting up periodic boundary conditions, adding solvent are the following steps. As needed, you can modify the system, for example, by adding extra bonds, changing charges, or modifying some amino acids’ ionization states. Finish by creating a system topology and coordinate files. You can perform each of these actions by executing LEaP commands interactively or importing them from a file.
- Two utilities for simulation preparation: tLEaP and xLEaP.
- The command line version (tLeap) is very efficient for executing scripts automatically.
- The GUI version (xLEaP) is good for building simulation systems interactively.
Simulation
A package of AMBER includes two MD engines: SANDER and PMEMD. Serial and parallel versions are available for both programs. AMBER also provides a GPU-accelerated version of PMEMD. Along with the topology file and the coordinate or resume file, you will also need an input file describing the parameters of your simulation, such as integrator, thermostat, barostat, time step, cut-off distance, etc. It’s also possible to use bond and distance constraints derived from NMR experiments. PMEMD and SANDER can run multiple simulations, including replica exchange and constant pH. Parameters for such simulations are contained in a special groupfile. A simulation program will save energy components and MD trajectories.
PMEMD is faster than SANDER, but there are limitations to the simulation types available in PMEMD. Some algorithms (such as DFT QM-MM with GPU-accelerated DFT code QUICK) are only available in SANDER.
- Two MD engines: SANDER and PMEMD.
- Parallel (MPI) SANDER
- Parallel and GPU-accelerated PMEMD
- PMEMD is faster than SANDER, but it has fewer options
Analysis
Analysis of simulation output files can be done using CPPTRAJ, PYTRAJ. CPPTRAJ is a command line program for trajectory analysis, and PYTRAJ is its Python front end.t end to ptraj. A program for estimating binding energies is called MMPBSA.
- CPPTRAJ (command-line pytraj)
- PYTRAJ (python front end to ptraj)
- MMPBSA
Other useful utilities:
- antechamber (topology builder)
- parmed (utility for topology editing and conversion)
- quick (GPU-accelerated DFT QM program)
- packmol-memgen (utility for preparing membrane systems)
Key Points
To run an MD simulation with AMBER 3 files are needed: an input file, a parameter file, and a file describing coordinates/velocities .