In the original PSS article, we illustrated the interest of the program through several examples of application to pertinent biological problems. PSSweb is based on PSS ( 2), a previously published command-line program, and is intended for facilitating its use by a wider audience. Results are presented in an HTML report containing a convenient summary with interactive figures and tables, as well as 3D visualization of superposed structures. Given a set of protein structures, PSSweb automatically performs multiple sequence alignment, structure superposition, calculates Cartesian and dihedral coordinate statistics and performs cluster analysis.

We present here PSSweb ( ), a web server for protein structural statistics. Finally, a cluster analysis can be performed to classify the structures. Complementary information can be obtained from internal coordinates, such as the dihedral angles of the backbone (ϕ, ψ, ω) and the sidechains (χs). Once superposed, quantitative information on the structural variability can be obtained by calculating the average and standard deviation of atomic Cartesian coordinates. The comparison of the structures can then be addressed, first, through structure superposition, which provides additional insights than sequence alignment alone. The sequence alignment is useful in itself to highlight conserved regions and derive structural, evolutionary or functional hypotheses.

One way to obtain such correspondence is to perform a multiple alignment of the sequences. When comparing structures with similar but different sequences, it is useful to establish a relationship between the residues of the structures. Such analysis can be useful for understanding protein flexibility, to select templates for homology modeling, or to study the effects of ligand binding or mutations. As the number of available protein structures continues to increase ( 1), it is becoming more critical to have tools that efficiently allow the comparison and analysis of structural ensembles.

Jmol ramachandran plot