Significance of Knotted Structures for Function of Proteins and Nucleic Acids
Poster Session II
42 – POS
Board 14
Implementation and Evaluation of the New Protocol for Protein Homology Modeling
Marta Strumillo
, Dominik Gront.
LTB, Warsaw, Poland.
Template-based modelling of a protein structure is one of ubiquitous tasks of structural
bioinformatics. The methods can deliver model structures important for testing biological
hypotheses, virtual docking and drug design. The performance of these methods is evaluated
every two years during Critical Assessment of Protein Structure Prediction (CASP) experiment.
In this contribution we present a new automated protocol for template-based modelling, which
combines computational tools recently developed in the Laboratory of Theory of Biopolymers:
database of protein domain structured (BDDB) with one dimensional and three dimensional
threading. At the first stage, template domain structures are assigned to the a sequence by the
means of 1D Threading approach. Then the query-template pairs are subjected to a 3D threading
procedure where a Monte Carlo search scheme is used to search for lowest scoring alignments.
Finally, a number of model structures are build based on each of the alignments. Three different
methods are used for this purpose: Modeller, Rosetta and CABS to ensure proper handling of
various modelling scenarios that range from very easy to most difficult cases where only a
remote homology may be detected. The models are clustered including possibility of knotted
structures and the final solution is selected. The protocol is currently tested in ongoing CASP11
experiment.
Gniewek et al. BMC Bioinformatics 2014, 15:22 BioShell-Threading: versatile Monte Carlo
package for protein 3D threading
Gront et al. NAR, 2012, Vol. 40 BioShell Threader: protein homology detection based on
sequence profiles and secondary structure profiles
- 89 -
