GIGP-510, Computer Applications in Biomedical Research

Computer Modeling of Proteins and Nucleic Acids

Free software for viewing .pdb files

There are several free programs you can download that will let you visualize and manipulate structure files that are in .pdb format. This is a widely used format, introduced by the Protein Data Bank (remember that PDB is a database of X-ray and NMR structures of biological macromolecules). So you can download a structure file from PDB or some other source, and then view the structure on your own computer.

In the lab, we will work with PyMOL. This is an open-source program that will run on Linux/Unix, Windows, or Mac platforms. It is free for students and teachers; researchers are asked to subscribe on an honor system.

Other free viewing software includes:

• RasMol, originally developed by Roger Sayle. This runs on Unix, Macintosh, Windows, and other systems. Not actively being developed, so it's a frozen-in-time product.
• DeepView Swiss-PdbViewer, available from the Swiss Institute of Bioinformatics. This comes in Macintosh, Windows, and (unsupported) Linux versions. More features, including energy minimization, but harder to learn.
• The Richardson Lab at Duke University offers KiNG, Mage, and other software that runs on several different systems.
• Chimera is produced by a group at UCSF. It has Windows, Mac, and Linux/Unix versions.
• Cn3D is a helper application for web browsers. It's available from NCBI. It includes windows for viewing structure and sequences.
• FirstGlance is a web-based java applet you can use to view PDB files or any other .pdb file you can find.
• QuteMol is a new one that looks nice, but I haven't had a chance to test-drive it yet. It comes in Windows and Mac OS X flavors.

Homology Modeling

A reasonable model of a protein can sometimes be constructed computationally using (a) the sequence of the desired protein, and (b) the crystallographic structures of one or more homologous proteins. This works because tertiary and secondary structure tend to be more highly conserved than primary structure. Using crystallographic structure(s) as a starting point, side chains can be modified as needed; deletions can be removed and gaps bridged; and insertions can be squeezed in. Possible conformations for insertion sequences can be located among structures in the Protein Data Bank or several alternative conformations can be built and tested (e.g., by Monte Carlo methods). Unfavorable steric and electronic interactions can be alleviated by several cycles of minimization and molecular dynamics.
The Swiss-Model Repository is a database of annotated three-dimensional comparative protein structure models generated by the fully automated homology-modelling pipeline SWISS-MODEL. The repository is developed at the Biozentrum Basel within the Swiss Institute of Bioinformatics. As of October 2007, it contains 1,341,793 models.
MODBASE is another database of protein models generated by automated homology modeling. As of October 2007, it contains models for 1,340,687 proteins.

Computer Modeling in Drug Design

For a discussion of computer modeling in drug design, take a look at materials I prepared for Biochemistry Course BC 5220, Techniques in Biomedical Research. Two topics covered there are receptor-based drug design (HIV protease inhibitors) and ligand-based drug design (the discovery of Viagra).