BACKGROUND: Transmembrane proteins (TMPs) constitute about 20~30% of all protein coding genes. The relative lack of experimental structure has so far made it hard to develop specific alignment methods and the current state of the art (PRALINE™) only manages to recapitulate 50% of the positions in the reference alignments available from the BAliBASE2-ref7. METHODS: We show how homology extension can be adapted and combined with a consistency based approach in order to significantly improve the multiple sequence alignment of alpha-helical TMPs. TM-Coffee is a special mode of PSI-Coffee able to efficiently align TMPs, while using a reduced reference database for homology extension. RESULTS: Our benchmarking on BAliBASE2-ref7 alpha-helical TMPs shows a significant improvement over the most accurate methods such as MSAProbs, Kalign, PROMALS, MAFFT, ProbCons and PRALINE™. We also estimated the influence of the database used for homology extension and show that highly non-redundant UniRef databases can be used to obtain similar results at a significantly reduced computational cost over full protein databases. TM-Coffee is part of the T-Coffee package, a web server is also available from http://tcoffee.crg.cat/tmcoffee and a freeware open source code can be downloaded from http://www.tcoffee.org/Packages/Stable/Latest.

MOTIVATION: Evaluating alternative multiple protein sequence alignments is an important unsolved problem in Biology. The most accurate way of doing this is to use structural information. Unfortunately, most methods require at least two structures to be embedded in the alignment, a condition rarely met when dealing with standard datasets. RESULT: We developed STRIKE, a method that determines the relative accuracy of two alternative alignments of the same sequences using a single structure. We validated our methodology on three commonly used reference datasets (BAliBASE, Homestrad and Prefab). Given two alignments, STRIKE manages to identify the most accurate one in 70% of the cases on average. This figure increases to 79% when considering very challenging datasets like the RV11 category of BAliBASE. This discrimination capacity is significantly higher than that reported for other metrics such as Contact Accepted mutation or Blosum. We show that this increased performance results both from a refined definition of the contacts and from the use of an improved contact substitution score. CONTACT: [email protected] AVAILABILITY: STRIKE is an open source freeware available from www.tcoffee.org SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

BACKGROUND: Recent approaches for predicting the three-dimensional (3D) structure of proteins such as de novo or fold recognition methods mostly rely on simplified energy potential functions and a reduced representation of the polypeptide chain. These simplifications facilitate the exploration of the protein conformational space but do not permit to capture entirely the subtle relationship that exists between the amino acid sequence and its native structure. It has been proposed that physics-based energy functions together with techniques for sampling the conformational space, e.g., Monte Carlo or molecular dynamics (MD) simulations, are better suited to the task of modelling proteins at higher resolutions than those of models obtained with the former type of methods. In this study we monitor different protein structural properties along MD trajectories to discriminate correct from erroneous models. These models are based on the sequence-structure alignments provided by our fold recognition method, FROST. We define correct models as being built from alignments of sequences with structures similar to their native structures and erroneous models from alignments of sequences with structures unrelated to their native structures. RESULTS: For three test sequences whose native structures belong to the all-alpha, all-beta and alphabeta classes we built a set of models intended to cover the whole spectrum: from a perfect model, i.e., the native structure, to a very poor model, i.e., a random alignment of the test sequence with a structure belonging to another structural class, including several intermediate models based on fold recognition alignments. We submitted these models to 11 ns of MD simulations at three different temperatures. We monitored along the corresponding trajectories the mean of the Root-Mean-Square deviations (RMSd) with respect to the initial conformation, the RMSd fluctuations, the number of conformation clusters, the evolution of secondary structures and the surface area of residues. None of these criteria alone is 100% efficient in discriminating correct from erroneous models. The mean RMSd, RMSd fluctuations, secondary structure and clustering of conformations show some false positives whereas the residue surface area criterion shows false negatives. However if we consider these criteria in combination it is straightforward to discriminate the two types of models. CONCLUSION: The ability of discriminating correct from erroneous models allows us to improve the specificity and sensitivity of our fold recognition method for a number of ambiguous cases.