Chargement Évènements

« Tous les Évènements

  • Cet évènement est passé

Séminaire bioinformatique structurale

février 23 @ 10:30 - 12:00

The CAPSID team organizes a seminar. The researcher Wim Vranken from the Université Libre de Bruxelles (VUB) and Graham Kemp from Gothenburg will give a talk
on Friday 23rd February at LORIA in room A006.

10:30    Wim Vranken (Vrije Universiteit Brussel)

Capturing the ambiguous behaviour of proteins from sequence-based predictions.

Protein conformations are not fixed in time, and conformational transitions are often essential for protein functionality. In addition, many of the as yet uncharacterised proteins in the ‘dark proteome’ likely exhibit conformational behaviour that cannot be captured by discrete conformational states. With our work, we aim to create a more probabilistic picture of protein behaviour that requires only their sequence. We predict per-amino acid residue biophysical characteristics such as dynamics, conformation and capacity for early folding, as influenced by the local sequence environment. The approach is applicable to any protein, and can be further combined with evolutionary information in an attempt to better understand the capacity proteins have for ambiguous and complex behaviour.

11:15    Graham Kemp (Chalmers University of Technology, Gothenburg)
Focusing the search when modelling proteins with Zipping and Assembly Mechanism (ZAM)
The zipping and assembly mechanism (ZAM) is a dynamic programming algorithm that can be used to sample the vast conformational search space when modelling a protein’s main chain.
Starting with a set of conformations for individual amino acid residues, these are combined to make longer fragments until, eventually, we obtain structural models for the entire protein main chain that are compatible with a given set of angle and distance constraints.
Examination of the results produced using a straight-forward implementation of ZAM suggested that constraints could be used to focus the search and avoid building fragments that are unlikely to be used in eventual solutions.
Focusing the search saves computational effort and reduces the amount of memory that is needed, without significantly affecting the quality of the resulting models.

Détails

Date :
février 23
Heure :
10:30 - 12:00
Catégorie d’Évènement:

En ce moment

Colloquium Loria 2018

Exposés précédents

Logo du CNRS
Logo Inria
Logo Université de Lorraine