Introduction

The TOPOLOG algorithm is a structure alignment algorithm for measuring the similarity between protein structures. It uses a similarity metric that is based on encoding the local twist, or "writhe", of the protein backbone into a structure-sequence.

Usage

This website provides protein structure comparison facilities that are based on the TOPOLOG algorithm. The facilities on this website can be accessed by clicking on cells of the facilities table at the top of the page. To perform a database search, for example, click on the cell at the top of the page titled:

The available facilities fall into two major categories: comparison of selected protein structures and searching based on a particular query structure. In the comparison category, the facilities include pairwise and multiple sequence/structure alignment. In the search category, the facilities include searches of a recent version of the PDB database and the ability to combine saved search results using standard set operations.

The TOPOLOG Algorithm

The TOPOLOG algorithm uses a structure metric based on the local twist, or "writhe", of the protein backbone. The algorithm associates a "writhe value" with each alpha-carbon of the protein backbone. Each writhe value is calculated from the "writhe" of the protein backbone in the vicinity of its associated alpha-carbon. The size of the window used to calculate the writhe values is an adjustable parameter of the TOPOLOG algorithm. The writhe values are encoded in a 20 letter writhe alphabet so that the structure of each protein is encoded in a "writhe sequence". These writhe sequences (and specialized BLOSUM style scoring matrices) are then used in standard sequence analysis software to perform the structure comparison. The SeqAln program is used for pairwise alignment, the ClustalW program is used for multiple alignment and the SSEARCH program is used for database searching.

References

The TOPOLOG algorithm is described in detail in the following paper

Structure alignment based on coding of local geometric measures
Chang PL, Rinne AW, Dewey TG
BMC BIOINFORMATICS 7: Art. No. 346 JUL 14 2006

A local copy of the paper (PDF format) is available here

Examples

To get started, examples of "writhe sequences" can be looked up by PDB identifier. Note that the chain specifier is the mmCIF 'asym_id' rather than the PDB author chain ID.

Access Counter: 13993 page loads from 1520 IP addresses since 7/26/2006