Version française

Saturday 15 December 2007

BECKER Emmanuelle

Bioinformatic predictions of the properties of peptide recognition module.


University : Paris VI Pierre et Marie Curie

Jury :

  • Président : O. Lequin
  • Rapporteurs : A. Poupon, G. Labesse
  • Examinateur : R. Guérois
  • Directeur de thèse : JM. Neumann

Summary :

Proteins involved in signalling pathways are frequently activated and inactivated by weak affinity interactions. In particular, domains that bind specifically short protein fragments, often called Peptide Recognition Modules (PRMs), allow an efficient intra- and inter-molecular regulation of the catalytic domains to whom they are associated. This work focuses on two domain families, the FHA and the tandem BRCT, often involved in the cell responses to DNA damage. Given their major role in the signaling networks, the prediction of their binding properties by bioinformatics is of crucial interest. However, bioinformatic strategies are still limited by methodological problems associated with the intrinsic characteristics of PRMs. They typically harbour very divergent sequences and their affinity for their physiological target is relatively weak despite their high specificity. This work aims at developing predictive approaches for the study of PRMs. Three points have been considered successively : (i) the prediction of the three-dimensional structure of PRMs, (ii) the prediction of their binding sites when the partner is known, (iii) the prediction of the sequence motifs each PRM specifically recognizes based on its three-dimensional structure. PDF

Keywords :

DNA repair, interaction networks, signalling, bioinformatics, structure prediction, peptide recognition module, extended linear motifs, protein-protein interactions, binding