The function of RNA molecules depends on their interaction with one or many partners. Upon interaction, RNA molecules often undergo large conformation changes. Understanding how these molecules interact with proteins would allow better targeting for therapeutic studies.
The plan is to improve upon the existing techniques by developing new models and algorithms for the study of RNA binding to proteins and RNA conformational changes. The LRI bioinformatics group has experience in protein-protein docking scoring functions that could be extended to protein-nucleic acid complexes.
The combination of Voronoi models at a coarse-grained level and powerful machine learning techniques allows the accurate scoring of protein-protein complexes. By adapting these approaches to protein-RNA, we would have a fast and efficient technique for scoring large protein-RNA complexes where conformational changes are involved.
Context
The function of RNA molecules depends on their interaction with one or many partners. Upon interaction, RNA molecules often undergo large conformation changes. Understanding how these molecules interact with proteins would allow better targeting for therapeutic studies. The CAPRI (Critical Assessment of PRediction of Interactions) challenge1 has shown that classical docking procedures largely fail when large conformation change occurs and when RNA is involved 1. This is especially true for RNA molecules, whose large-scale dynamics remain often unknown. Modeling RNA conformational changes is made hard by the inherent flexible nature of their structure but also by the electrostatics involved. These are hard to model and often lead to computationally expensive simulations. Even if for small RNA molecules, molecular dynamics can be used, such simulations are hard to extend to larger molecules and protein-RNA complexes.
Objectives
This PhD project contains two objectives. The first objective is related to computational structural biology: developing new approaches for modeling RNA interactions and flexibility . The second is related to bioinformatics: conceiving new machine learning techniques to learn models from very imperfect data using multiple representations of an object. One of the specificity of this project is that it implies multi-scale modeling. We would start from a set of coarse-grained conformations, select the most promising subset of conformations that should be refined and generate new conformations from this subset at higher resolution. In this new approach, machine learning and the conformational generation have to be closely linked.
Work program
The first step will be to be able to efficiently model RNA: i.e. generate models that are closed enough to the known biological solution in a relatively computationally inexpensive manner.
The second step will deal with the improvement of the existing machine learning approaches in order to evaluate the efficiency of the modeling stage: can we score efficiently models we generated?
The final step will be to develop new machine learning approaches that may need to interact with the modeling stage in a multi-scale fashion. This would allow to gradually focus on the best model created by the modeling approach.
Extra information
Prerequisite
Good skills in computational biology and RNA modeling are required for this PhD. Skills in Machine Learning will also be appreciated.
Détails
Expected funding
Institutional funding
Status of funding
Expected
Candidates
Adrien Guilhot Gaudeffroy (M2 BIBS)
Utilisateur
Créé
Mercredi 01 juin 2011 17:03:55 CEST
dernière modif.
Jeudi 09 juin 2011 18:44:31 CEST
Fichiers joints
filename
créé
hits
filesize
sujet-bioinfo-ARN.pdf
09 Jun 2011 18:42
414
92.31 Kb
Connexion
Ecole Doctorale Informatique Paris-Sud
Directrice
Nicole Bidoit Assistante
Stéphanie Druetta Conseiller aux thèses
Dominique Gouyou-Beauchamps
ED 427 - Université Paris-Sud
UFR Sciences Orsay
Bat 650 - aile nord - 417
Tel : 01 69 15 63 19
Fax : 01 69 15 63 87
courriel: ed-info à lri.fr