FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W1582362079> ?p ?o ?g. }

• W1582362079 endingPage "1654" @default.
• W1582362079 startingPage "1651" @default.
• W1582362079 abstract "A chilly, but still very attractive, Asilomar hosted the seventh edition of the Critical Assessment of Protein Structure Prediction Experiment (CASP) [[1]]. A few months ago, 96 protein sequences of unknown, but soon to be determined structure, were distributed to more than 250 groups, whose task was to predict their structure and a number of other features, such as their domain boundaries, the presence of disordered regions, contacts between specific amino acids and their molecular function. As soon as the structure of a target protein was available, the prediction centre at UC Davis (http://www.predictioncenter.org) compared it with all submitted predictions, computed several parameters and passed them to a set of experts, who were given the task to draw conclusions about the state of the art in the field. This year the assessors for structure prediction were Neil Clarke (Genome Institute of Singapore, Singapore), Torsten Schwede (Swiss Institute of Bioinformatics, Basel, Switzerland) and Randy Read (Cambridge Institute for Medical Research, Cambridge, UK). Alfonso Valencia (Spanish National Cancer Research Centre, Madrid, Spain), Lorenza Bortoli (Swiss Institute of Bioinformatics), Neil Clarke, and Mike Tress (Spanish National Cancer Research Centre, Madrid, Spain) assessed the so-called ‘other categories’, namely function prediction, disordered regions, contacts and domain boundary prediction. A new category was introduced this year: predictors were given a set of models produced by automatic servers and were asked to predict the ‘quality’ of each model (i.e. its distance from the native structure). After the structures were made available, Andriy Kryshtafovych (University of California, Davis, CA, USA) and I analysed the correctness of the predictions in this area. This complex process culminated in a meeting, held in Asilomar (CA, USA), where the scientists involved in the experiment convened, discussed, sometimes disagreed, sometimes ‘stamped’. The latter is a CASP tradition that takes advantage of the wooden floor of the meeting room and of a quite demanding audience to notify speakers when they are not following the rule of the experiment, and the rule is that they should discuss what went right and what went wrong in their modelling efforts as revealed by the blind tests of the experiment, and refrain from claiming successes other than those certified by the results of the blind test. A few ‘external’ seminars were held this year in the context of the meeting, by people involved in the folding problem from a less heuristic point of view, i.e. physics or principle-based structure prediction, and by experimentalists who tried to tell the audience what is needed in the field from their point of view. One may think that, after 12 years, we might not need a CASP experiment anymore, that the concept of blind test has permeated the field and that it is clear that the scientific community would accept no unsubstantiated claims on structure prediction methods. However, my personal view is that we are still far away from reaching this goal and that CASP is as necessary as ever to maintain a serious approach to the evaluation of prediction results. As CASP's originator, John Moult (CARB, Rockville, MD, USA), continuously states, CASP is not a competition. However, there are people who performed substantially better than others. You may choose not to call them ‘winners’, but the fact remains that their methods have an edge on the others. In this year's experiment, two groups clearly stood out from the crowd. The group of Yang Zhang (University of Kansas, Lawrence, KS, USA) improved the Tasser method [[2]], originally developed in Jeff Skolnick's group (Georgia Institute of Technology, Atlanta, GA, USA) where Zhang was a postdoc, and produced models of impressive quality. The other group, unsurprisingly, was David Baker's (Washington University, Seattle, WA, USA) who improved the Rosetta method [[3]] mainly by recruiting hundreds of thousands of CPUs around the world in his Rosetta@home project (http://boinc.bakerlab.org/rosetta/) and therefore was able to sample the conformational space of proteins much more thoroughly than was possible before. It should be mentioned that assessment in CASP has become very professional, is always based on sound statistical analysis, and has essentially become uncontroversial; therefore the conclusions reported here are not ‘my view’ of what happened, but a real sound and provable result. Nevertheless there are complex issues, not directly related to the numerical assessment but still very relevant to the field. For example, some of the predictions are based on, or sometimes just taken from, the results of publicly available servers. In these cases, it is really unclear how to assign credit and how to compare the results of a genuinely original method from those obtained by selecting the best model among several alternative ones. A substantial fraction of the time at this year's meeting was devoted to the problem of establishing whether there has been progress since the last experiment. This is a difficult question: the targets are different in each experiment and therefore the complexity of predicting them can be different. Furthermore, sequence and structure databases grow continuously, so that the task of identifying suitable structural templates and obtaining a good sequence alignment for a given protein target becomes progressively easier. For these reasons, comparing results from different experiments is a tricky business, and methods for defining a ‘difficulty scale’ for a protein structure prediction are continuously being proposed. The discussions at the meeting were animated, and no firm consensus was reached about the method that should be used. Nevertheless, it is refreshing to see a community interrogate itself to find objective and reliable ways to define whether they are making progress or not. A few comments are required on the results in the other categories. In domain boundary predictions, not many conclusions could be derived. Most of the target proteins came from structural genomics projects where, apparently, multidomain proteins are rarely selected and therefore the results in this area were inconclusive. Nothing much happened in disorder prediction either: no new methods appeared and the old ones performed exactly as expected, i.e. at the same level as in the previous CASP experiment. Assessing function prediction was as difficult as in the past experiment, to the point that it was decided that, in the next CASP experiment, predictors will be asked only to concentrate on the prediction of residues involved in binding sites rather than in predicting the fully fledged molecular function of the target proteins. The problem is that it is difficult to assess the predictions, as the assessors do not have any more information than the predictors at the end of the experiment. The issue has been discussed before [[4]], and we published a retrospective analysis of the CASP6 function predictions in this journal [[5]], concluding that the predictions can be useful, provided that a large number of groups participate. Yet this did not convince many members of the community to take part in the challenge, and too few predictions were again submitted this time around. This is not only a CASP problem. The only way to improve methods for function prediction is to test methods blindly, and the scientific community has not yet found a good way to do this, something we should all worry about. Finally, it is extremely relevant to have an idea of the quality of the thousands of models produced every day all over the world if we want to make use of them in a sensible way. The category of ‘quality prediction’ was introduced exactly for this reason, and we are happy to report that there are methods that are able to evaluate ‘a priori’ the quality of a model on the basis of the atomic coordinates (Fig. 1), the best ones being those available at the Stockholm Bioinformatics Centre (http://www.sbc.su.se/) [[6]]. We urge our experimental and computational colleagues to use them in place of other more popular but much less effective methods for model evaluation. Unfortunately, this, and the other successful methods, are able to select with reasonable accuracy among alternative models, but no really good method is available for assessing the quality of a single model. Example of the prediction of a model's quality. The graph shows the predicted (green line) and observed (blue line) deviation of the model from the experimental structure, which is shown in ribbon representation. The red regions are those predicted to be incorrectly modelled. The prediction was submitted by group 634 (Pcons6) and refers to the model of the protein target T0334 (RebH from Saccharothrix aerocolonigenes) by group 609 (GeneSilico Metaserver). A simple back of the envelope calculation shows that 250 groups working on a project for 3 months every 2 years (the effort of the CASP community in the prediction season, without taking into account all the time and effort needed to develop and test the methods) represents a substantial investment in time and money, probably comparable to some of the large worldwide genomic and postgenomic projects. It is legitimate to question whether the results are worth the effort. Undoubtedly, everybody has a different answer to this question. Personally I strongly believe that CASP is necessary to the community for at least three reasons: first, for removing false claims of success from such an important field as structure prediction; second, to highlight areas where we are not able to make progress in a measurable way; and, last but not least, to urge the prediction community to continuously develop and improve the methods. I believe that, if the readers look at some of the predictions, such as the one shown in Fig. 2, they will agree with me. Comparison of the structure (green) and a model (red) of the target protein T0350.The target protein is SR482 from Bacillus subtilis, available from the PDB with the code 2HC5. The model was submitted by group 113 (Bates), but it is extremely similar to one of the models automatically produced by a server (Baker server). Note that this protein is not homologous to any protein of known structure and has a novel fold. CASP7 was sponsored by the US National Library of Medicine (NIH/NLM), the National Institute of General Medical Sciences (NIH/NIGMS) and by BioSapiens, a Network of Excellence funded by the European Commission within its FP6 Programme, under the thematic area ‘Life sciences, genomics and biotechnology for health’, contract number LSHG-CT-2003-503265. AT is grateful to Dr. Domenico Raimondo for helping in preparing the figures." @default.
• W1582362079 created "2016-06-24" @default.
• W1582362079 creator A5073708939 @default.
• W1582362079 date "2007-02-20" @default.
• W1582362079 modified "2023-10-18" @default.
• W1582362079 title "An account of the Seventh Meeting of the Worldwide Critical Assessment of Techniques for Protein Structure Prediction" @default.
• W1582362079 cites W2017285544 @default.
• W1582362079 cites W2052112948 @default.
• W1582362079 cites W2057272128 @default.
• W1582362079 cites W2101495661 @default.
• W1582362079 cites W2145356476 @default.
• W1582362079 cites W2154106236 @default.
• W1582362079 doi "https://doi.org/10.1111/j.1742-4658.2007.05718.x" @default.
• W1582362079 hasPubMedId "https://pubmed.ncbi.nlm.nih.gov/17313483" @default.
• W1582362079 hasPublicationYear "2007" @default.
• W1582362079 type Work @default.
• W1582362079 sameAs 1582362079 @default.
• W1582362079 citedByCount "10" @default.
• W1582362079 crossrefType "journal-article" @default.
• W1582362079 hasAuthorship W1582362079A5073708939 @default.
• W1582362079 hasConcept C18051474 @default.
• W1582362079 hasConcept C41008148 @default.
• W1582362079 hasConcept C47701112 @default.
• W1582362079 hasConcept C55493867 @default.
• W1582362079 hasConcept C70721500 @default.
• W1582362079 hasConcept C86803240 @default.
• W1582362079 hasConceptScore W1582362079C18051474 @default.
• W1582362079 hasConceptScore W1582362079C41008148 @default.
• W1582362079 hasConceptScore W1582362079C47701112 @default.
• W1582362079 hasConceptScore W1582362079C55493867 @default.
• W1582362079 hasConceptScore W1582362079C70721500 @default.
• W1582362079 hasConceptScore W1582362079C86803240 @default.
• W1582362079 hasIssue "7" @default.
• W1582362079 hasLocation W15823620791 @default.
• W1582362079 hasLocation W15823620792 @default.
• W1582362079 hasOpenAccess W1582362079 @default.
• W1582362079 hasPrimaryLocation W15823620791 @default.
• W1582362079 hasRelatedWork W1564749278 @default.
• W1582362079 hasRelatedWork W1975122118 @default.
• W1582362079 hasRelatedWork W2043066834 @default.
• W1582362079 hasRelatedWork W2058542300 @default.
• W1582362079 hasRelatedWork W2136856901 @default.
• W1582362079 hasRelatedWork W2294851134 @default.
• W1582362079 hasRelatedWork W2368468053 @default.
• W1582362079 hasRelatedWork W2946599741 @default.
• W1582362079 hasRelatedWork W3171039768 @default.
• W1582362079 hasRelatedWork W2593264178 @default.
• W1582362079 hasVolume "274" @default.
• W1582362079 isParatext "false" @default.
• W1582362079 isRetracted "false" @default.
• W1582362079 magId "1582362079" @default.
• W1582362079 workType "article" @default.