Evaluation of methods for prediction of conformation of loops, and their application to the modelling of 3-d structure of proteins

Lines connect the coordinator with its partners. CORDIS does not always give exact coordinates for projects before 2014. These points are placed at city or country level.

Project objective

Research objectives and content The aim of the proposed project is to up-date the loop-data base and to evaluate the loop selection procedures. Loop regions are non-repetitive conformations connecting regular secondary structures. They are the hardest to model and most prone to errors. However the accurate modelling of loop regions may be crucial as they are often involved in molecular recognition-processes. In order to model loops it is necessary to select protein fragments from a wide database of all known protein structures, and not just members of the homologous family. Generally they are selected using a geometric filter. When the fragments are sought out side the family, a large number of segments usually results posing the problem of selection of the most appropriate fragments. Two powerful methods have been developed: the fragment ranking method (Topham et al, 1993) and the SLOOP method (Rufino et al, 1997). Both methods when individually used are useful in the selection of appropriate fragments for loop modelling. However, it is necessary to evaluate the performance of both. For this we are proposing to carry out an exhaustive study on a large number of loop sequences from proteins of known-three dimensional structure. We also propose to derive at a new scoring scheme in which the contributions from each method are appropriately weighted. Finally, we will apply this procedures to the model of unknown three-dimensional structure proteins. Training content (objective, benefit and expected impact) The objective of the proposal is to achieve a deeper knowledge in loop-prediction. As loop regions are the most difficult to model, any improvement in the prediction of these regions will be worth. Thus, if we develop a new scoring scheme, it will be a useful tool for protein modellers. Regarding my own training, I will get a deeply and broader knowledge in modelling of three dimensional structures of proteins. Links with industry I industrial relevance (22) The Cambridge group is funded by Oxford Molecular Ltd. Who are developing their software commercially in parallel to its release freely to academia. The software for modelling protein is of interest to the pharmaceutical and biotechnology industries. The Cambridge group has collaborative research programmes and is funded by Glaxo Wellcome, Smithkline Beecham, Parke Davies and Pfizer (UK).

Original text from CORDIS.

Participants

• University of Cambridge · CambridgeCoordinatorUnited Kingdom
• Not availableCity levelSpain

Links

• View on CORDIS