PPI-3D – Structure Meets Genomics

PPI-3D is an Associate Inria Team shared between Nano-D at Inria Grenoble and Kozakov Lab at Stony Brook University, USA. Team members include:

  • Kozakov lab :
    • Dima Kozakov, Assistant Professor
    • Bing Xia, fixed – term graduate student, started in 2011
      • expertise in computer science
    • Dzmitry Padhorny, fixed – term graduate student, started in 2014
      • expertise in the development of computational tools to predict protein-protein interactions

Record of activities :

  1. Emilie Neveu visited the University of Luxembourg to work on sampling with protein flexibility on 26–29 of August  2015.
  2. 4 weeks of visit Emilie Neveu to Stony Brook (Oct 26 – Nov 27)
  3. 2 weeks of visit Alexandre Hoffmann to Stony Brook (Nov 15 – Nov 27)
  4. 2 weeks of visit Dima Kozakov to Inria Grenoble (Jan 2016)
  5. 2 weeks of visit Dzmitry Padhorny to Inria Grenoble (Jan 2016)
  6. 2 weeks of visit Leonard Jaillet to Stony Brook (March 2016)
  7. 2 weeks of visit Dima Kozakov to Inria Grenoble (July 2016)
  8. The partners also met at international conferences and workshops,
    1. 3rd International Conference on Protein and RNA Structure Prediction (formerly Zing) – Punta Cana, Dominican Republic, Dec 14-18, 2015
    2. Nano-D 2016 workshop, Autrans Jan 20-22, 2016.
    3. The 6th CAPRI evaluation meeting, Tel-Aviv, Israel, APRIL 17-19, 2016
    4. The First International Conference on Computational Genomics and Proteomics, Costa Rica, October 18 – 22
  9. A master student working on structural databases of small molecules was hired for 6 months.

Scientific progress :

  1. Flexible docking predictions The project aims to develop new models of flexible proteins. We are exploring large-scale conformational changes using reduced representations of the conformational space. More precisely, we are using eigen-analysis of the stiffness interaction matrices with some constrained optimization problems.
  2. New sampling strategies  The project aims to develop new search strategies when doing docking predictions with flexible proteins. We focus here on the first step of the docking method, that is finding all promising conformations, without prior knowledge of the biding sites. For this step, we have to rely on a good sampling strategy not to miss the most promising conformations, i.e. the lowest energy conformations. We study  meta-heuristics, robotics-inspired sampling methods, hybrid FFT-based approaches, and combinations of these with detailed knowledge-based potentials.
  3. New scoring functions This project aims at developing new physics-based scoring functions for protein docking and protein folding applications. In particular, we are analyzing structural databases of proteins to identify some new functional forms of protein-protein interactions. We have already developed scoring functions for protein-protein and protein-ligand interactions and are currently working on the applications of these and on protein folding potentials.

Production :

  1. Sergei Grudinin, Maria Kadukova, Andreas Eisenbarth, Simon Marillet, Frederic Cazals. Predicting binding poses and affinities for protein-ligand complexes in the 2015 D3R Grand Challenge using a physical model with a statistical parameter estimation. Journal of Computer-Aided Molecular Design, Springer Verlag, 2016, 30 (9), pp.791-804.
  2. Marc F. Lensink, … , Sergei Grudinin, …, Dima Kozakov, et al. Prediction of homo- and hetero-protein complexes by protein docking and template-based modeling: a CASP-CAPRI experiment. Proteins – Structure, Function and Bioinformatics, Wiley, 2016.
  3. Andreas Eisenbarth defended his master thesis at the Kaiserslautern University of Technology.
  4. The corresponding software package along with a manuscript are currently in preparation.
  5. We extracted functional forms of physical interactions between small molecules by means of large-scale machine learning and knowledge extraction from 3D data. The corresponding manuscript is being currently reviewed at J. Chem. Inf. Mod.
  6. The corresponding software package that scores protein-ligand complexes has been posted on the teams website.
  7. The PRE (NMR paramagnetic relaxation enhancement) analysis module has been added into the SAMSON software toolbox.
  8. The two teams critically assessed the Cartesian-based FFT sampling schemes developed by the US team with the rotational-based FFT sampling schemes developed in Nano-D in the framework of CASP12/CAPRI blind docking experiment.

Comments are closed.