News and Publications

Analysis and Predictions of Protein Structures

A. Godzik, L. Rychlewski, K. Pawlowski, B. Zhang, L. Jaroszewski*

* University of Warsaw, Warsaw, Poland

Improved accuracy in methods of recognizing protein folds and predicting protein structure developed in our group enabled us to assemble an automated set of programs for genomic-scale prediction of protein folds and verification of protein function. New advances include a reworked threading algorithm, an improved algorithm for predicting local structure, a new sequence alignment program based on a library of generalized sequence profiles for several thousand protein families, automated modeling tools that enable us to build low-resolution structures for all predicted folds, and tools for automated verification of the conservation of functionally important features in predicted structures.

Several algorithms developed in our group and some public-domain software were put together to form a cascade of fold-prediction methods. Estimates of the significance of the predictions are made for all methods. This aspect enabled us to combine predictions from different methods to develop a "jury" method, which has an accuracy higher than any of the single methods. Building full 3-dimensional models for all top predictions helps eliminate possible false-positives when alignments, which are optimal in the 1-dimensional sequences, lead to unsolvable sterical conflicts for the full 3-dimensional models. Finally, alignments are annotated by using databases of functionally important features, a step that shortens the time-consuming process of verifying the alignments and models.

This suite of programs was applied to proteins from genomes of several organisms, including Escherichia coli, Mycoplasma genitalium, Haemophilus influenzae, Methanococcus jannashii, Bacillus subtilis, and Helicobacter pylori. Three-dimensional folds could be assigned to almost 40% of proteins from each genome, which is a more than 2-fold increase over standard homology recognition methods. Many of the fold predictions could be indirectly verified by functional similarity between the new proteins and the protein group to which they are assigned. For others, fold predictions open a way to rationalizing, and in some cases predicting, the functions of new proteins.

A database of predictions of the folds and functions of proteins from the genomes of M genitalium and E coli are made available to the TSRI community via our World Wide Web server at cape6.scripps.edu. In addition, all software tools developed in our group are available on the server. Programs available on the server include a new method for predicting secondary structure, a new sequence alignment program called BASIC, a protein structure alignment program, and several others.

PUBLICATIONS

Fetrow, J.S., Godzik, A. Function driven protein evolution: A possible proto-protein for the RNA-binding proteins. In: Biocomputing 98. Altman, R.B., et al. (Eds.). World Scientific, River Edge, NJ, 1998, p. 485.

Fetrow, J., Godzik, A., Skolnick, J. Functional analysis of the Escherichia coli genome using the sequence-to-structure-to-function paradigm: Identification of proteins exhibiting the glutaredoxin/thioredoxin disulfide oxidoreductase activity. J. Mol. Biol., in press.

Jaroszewski, L., Pawlowski, K., Godzik, A. Multiple model approach: An extension of comparative modeling. J. Mol. Modeling, in press.

Jaroszewski, L., Rychlewski, L., Zhang, B., Godzik, A. Fold prediction by a hierarchy of sequence and threading methods. Protein Sci., in press.

Rychlewski, L., Godzik, A. Local sequence similarity secondary structure prediction. Protein Eng. 10:1143, 1997.

Rychlewski, L., Zhang, B., Godzik, A. Function and fold predictions for Mycoplasma genitalium proteins. Folding Design, in press.

Rychlewski, L., Zhang, B., Godzik, A. Searching for the optimal sequence similarity function. Protein Eng., in press.

Zhang, B., Jaroszewski, L., Rychlewski, L., Godzik, A. Similarities and differences between non-homologous proteins with similar folds: Evaluation of threading strategies. Folding Design 2:307, 1997.