• Welcome
• Research
• Software
• People
• Publications
• Contact
• master/HPC

Master official page at Universitat Pompeu Fabra

High performance computing 2012

Synopsis
The focus of this course is to provide the tools, knowledge and practice to perform biochemical experiments on proteins and other molecules (drugs) in-silico using molecular dynamics simulations on a high performance computing infrastructure.

• Target students
  • Students with an interest in computing and simulation, structural biology, computational biophysics and biochemistry.
• Requirements
  • Courses: Attendance to the course MSI is recommended for this course.
  • Programming: Use of medium to high level scripting is used in order to perform analysis on large simulation data sets.
  • Knowledge: Interdisciplinary. Pharmaceutics, biochemistry, chemistry and biology will be an advantage to understand the molecular systems, computer science, physics and mathematics for the understanding of the fundamentals of the methodology.
• Practical information
  • Course length is 30 hours of classes (corresponding to 100 hours of personal work including classes) of which 4 hours are for the final project revision and defence.
  • Room: all lectures will be in Aula 60.122, except the seminar

  • Google calendar for the course

• Evaluation
  • 50% of final evaluation is based on exercises during the course
  • 50% of final evaluation is based on research projects carried out by groups and defended publicly.
• Picture and video gallery

  • You Tube channel: http://www.youtube.com/user/ps3grid

  • Flickr channel: http://www.flickr.com/photos/multiscalelab/with/3542248429/

  • Vimeo channel: http://vimeo.com/user862246

• Useful books:
  • Computer Simulation of Liquids, Allen and Tildesley.
  • Understanding Molecular Simulation, Second Edition: From Algorithms to Applications, Frenkel and Smit.
• Contacts: toni.giorgino at upf.edu and gianni.defabritiis at upf.edu

Resources

Hardware

You will do the exercises, mostly in group, on your laptop. Access to a GPU-equipped remote machine will be provided for running HPC simulations.

Operating systems

It is recommended that you use either Linux or OSX on the laptops you bring to class. This will greatly simplifies programming and scripting. While Windows may work for most of the exercises, it will be inconvenient. If you have a Windows-only machine, you may consider the following options (use Google to find them)

• Install Linux properly - preferred (suggested distributions: Ubuntu and Fedora)
• Make a bootable USB key, such as Puppy Linux

• Install a virtual machine, such as Oracle VM VirtualBox, and install Linux inside it

• Make a non-permanent install of Ubuntu via Wubi

Software

The following free software will be used during the classes: preferably, install and test it before the course.

• Virtual Molecular Dynamics (VMD): Software downloads and User guide

• Gnu Octave

• Java JRE (browser plugin): http://www.java.com

• ACEMD user guide http://multiscalelab.org/acemd/manual

The TCL Programming language

The TCL language will be used for the course practices. Useful resources -

• TCL Tutorial

• TCL Language reference

Protocols

• ACEMD protocols http://multiscalelab.org/acemd/protocols

Course material

• Class 1: Molecular dynamics hands-on
  • Concept of classical dynamics
  • The water molecule
  • Amber and Charmm forcefields (bond, angle, dihedral, improper, Lennard-Jones, Coulomb)

  • Charmm topology filestop_all27_prot_lipid.top

  • Charmm parameter filespar_all27_prot_lipid.prm

  • Molecular dynamics codes
• Class 2: Introduction to VMD

  • http://www.ks.uiuc.edu/Training/Tutorials/vmd/tutorial-html/ (tutorial files http://www.ks.uiuc.edu/Training/Tutorials/vmd/)

  • Use of the user guide http://www.ks.uiuc.edu/Research/vmd/current/ug.pdf

  • Topics: Load molecule and navigate it, representations, selections (resname, name, type, resid), within of, same residue as, loading trajectories, keys (r,t,s,1,2)
  • Example of Tcl scripting (atomselect, measure)

  • Practice: VMD_exercise.txt, 1NEY.pdb, TRJ: http://www.gpugrid.net/download/trj.dcd, PDB: http://www.gpugrid.net/download/TIM_sub_cry.pdb

• Class 3: Preparing a molecular structure for MD simulations

  • Built from PDB in the Charmm format http://www.multiscalelab.org/acemd/protocols/ACETK.BLDCHARMM

  • Relaxation and production run http://www.multiscalelab.org/acemd/protocols/ACETK.EQ

  • Practice
• Class 4: Scripting in VMD
  • Advanced problem solutions using VMD scripting
• Class 5: Practical
• Class 6: Molecule parametrization

  • Molecule parametrization using the Amber force field http://www.multiscalelab.org/acemd/protocols/ACETK.PRMAMBER

  • Built from PDB in the Amber format http://www.multiscalelab.org/acemd/protocols/ACETK.BLDAMBER

  • Practice
• Class 7: Advanced system building

  • Molecule parametrization using the Charmm force field http://www.multiscalelab.org/acemd/protocols/ACETK.PRMCHARMM

  • Built a membrane using charmm-gui.org and VMD
  • Equilibration
  • Embed a protein into a lipid bilayer
  • Practice
• Class 8: Markov state models
  • Theory behind Markov state models analysis
  • Practice using simple Brownian dynamics
• Class 9: Reconstructing the binding process by Markov state models
  • Seminar
  • Practice
• Class 10: Practical
• Class 11: Biased free energy calculations: Metadynamics
  • Theory behind metadynamics
  • Practice
• Class 12: Protein-ligand binding affinity calculations by umbrella sampling
  • Seminar
  • Practice
• Class 13: Seminar from PRBB Structural biology series

  • http://structure.prbb.org

• Class 14: Project review
• Class 15: Project defense