Sequence harmony: detecting functional specificity from alignments

Abstract

Multiple sequence alignments are often used for the identification of key specificity-determining residues within protein families. We present a web server implementation of the Sequence Harmony (SH) method previously introduced. SH accurately detects subfamily specific positions from a multiple alignment by scoring compositional differences between subfamilies, without imposing conservation. The SH web server allows a quick selection of subtype specific sites from a multiple alignment given a subfamily grouping. In addition, it allows the predicted sites to be directly mapped onto a protein structure and displayed. We demonstrate the use of the SH server using the family of plant mitochondrial alternative oxidases (AOX). In addition, we illustrate the usefulness of combining sequence and structural information by showing that the predicted sites are clustered into a few distinct regions in an AOX homology model. The SH web server can be accessed at www.ibi.vu.nl/programs/seqharmwww.

Figure 1.

Input screen of the Sequence Harmony server, showing part of the AOX alignment and model input. (A) Overview of the Sequence Harmony web page; and (B) the Sequence Harmony input options and advanced features.

Figure 2.

Output screen of the Sequence Harmony server with results of the AOX analysis. Included are a summary of input parameters, a PyMol rendered image, the Jmol applet and controls, and the output table. Links to the raw table with all alignment positions and to additional output files are also provided.

Figure 3.

Model of the AOX iron-binding domain with low-harmony sites shown in spheres and the iron-binding Glu and His residues in sticks. Figure created using PyMol (www.pymol.org).