| dc.description.abstract | Finding wide tunnels in protein structures is an important problem in Structural Bioinformatics with applications in various areas such as drug design. Several algorithms have been proposed for finding wide tunnels in a fixed protein conformation. However, to the best of our knowledge, none of the existing work have considered widening the tunnel, i.e., finding a wider tunnel in an alternative conformation of the given structure. In this thesis we initiate this line of research by proposing a tunnel-widening algorithm which aims to make the tunnel wider by a slight local change in the structure of the protein.
Given a fixed conformation of a protein with a point located inside it, we first describe an algorithm to identify the widest tunnel from that point to the outside environment of the protein. Then we try to make the tunnel wider by considering various alternative conformations of the protein. We only consider conformations whose energies are not much higher than the energy of the initial conformation. Among these alternative conformations we select the one with the widest tunnel. However, the alternative conformation with the widest tunnel might not be accessible from the initial structure. Thus, in the next step we develop three algorithms for finding a feasible transition pathway from the initial structure to the alternative conformation, i.e., a sequence of intermediate conformations between the initial structure and the alternative conformation such that the energy values of all these intermediate conformations are close to the energy of the initial structure.
We evaluate our tunnel-finding and tunnel-widening algorithms on various proteins. Our experiments show that in most cases we can make the tunnel wider in an alternative conformation. However, there are cases in which we find a wider tunnel in an alternative conformation, but the energy value of the alternative conformation is much higher than the energy of the initial structure. We also implemented our three pathway-finding algorithms and tested them on various instances. Our experiments show that although in most cases we can find a feasible transition pathway, there are cases in which the alternative conformation has energy close to the initial structure, but our algorithms cannot find any feasible pathway from the initial structure to the alternative conformation. Furthermore, there is a trade-off between the running time and accuracy of the three pathway-finding algorithms. | en |
