Model quality assessment using distance constraints from alignments

Abstract

Given a set of alternative models for a specific protein sequence, the model quality assessment (MQA) problem asks for an assignment of scores to each model in the set. A good MQA program assigns these scores such that they correlate well with real quality of the models, ideally scoring best that model which is closest to the true structure.In this paper, we present a new approach for addressing the MQA problem. It is based on distance constraints extracted from alignments to templates of known structure, and is implemented in the Undertaker [9] program for protein structure prediction. One novel feature is that we extract noncontact constraints as well as contact constraints.We describe how the distance constraint extraction is done and we show how they can be used to address the MQA problem. We have compared our method on CASP7 targets and the results show that our method is at least comparable with the best MQA methods that were assessed at CASP7 [7].We also propose a new evaluation measure, Kendall's τ, that is more interpretable than conventional measures used for evaluating MQA methods (Pearson's r and Spearman's ρ).We show clear examples where Kendall's τ agrees much more with our intuition of a correct MQA and we therefore propose that Kendall's τ be used for future CASP MQA assessments. IntroductionMost search algorithms for protein structure prediction are guided by cost functions that assess how "protein-like" particular conformations of the polypeptide chain are. In theory, a perfect cost function would guide a good search algorithm to the native state of the protein, but such a cost function has yet to be discovered.One of the obstacles is that many low-cost structures usually exist in the conformational search space and even good cost functions have trouble identifying the most native-like structure among them. For a given set of alternative models for some specific protein target, the model quality assessment (MQA) problem asks for an assignment of a score to each model in the set, such that the scores correlate well with the real quality of the model (that is, the similarity with the native structure). This assignment of scores is, of course, done without knowing the native structure of the protein.A good MQA is crucial when one has to choose the best model among several different models-for example, in a metaserver for protein structure prediction. Metaservers use structure models generated by other methods and either choose one of the models using an MQA or construct a consensus model to make a predicted structure. The most successful NIH Public Access MQA methods in the past have been either consensus methods (looking for features shared by many models in the set) or similarity to a single predicted model [7,18].The Lee group has been fairly successful at predicting the tertiary structure of CASP targets. Their method for MQA therefore first predicts the structure of the target and then measures the similarity between their prediction and the models to be assessed [7], a method which a...

Improved model quality assessment using ProQ2

“…The first of them is single-model methods that only use information from the actual model, such as evolutionary information [8-10], residue environment compatibility [11], statistical potentials from physics [12] or knowledge-based ones [13,14], or combinations of different structural features [15-19]. The second class is consensus methods that primarily use consensus of multiple models [1] or template alignments [20] for a given sequence to pick the most probable model. Finally, there are also hybrid methods that combine the single-model and consensus approaches to achieve improved performance [21-24].…”

SAM-T08, HMM-based protein structure prediction

“…The constraints have also been used for model quality assessment in evaluating models from other servers (22,13), but that application is not provided by the web server.…”

Template-Based Protein Modeling: Recent Methodological Advances

“…In another report addressing improvement of the undertaker scoring function, weighted distance constraints generated from alignment to different templates were used for model quality assessment. [118] The globularity index, a combined score including hydrogen bonding information, solvent accessible surface area, voids and the number of water molecules within 5 Å of the protein, has been used for evaluation of the quality of protein models. [119] The ModFOLD[108] server combines ModSSEA[120], MODCHECK[121] and ProQ[122] scores with secondary structure information.…”