Serkan Apaydin
My research is about NMR protein backbone resonance assignments. My work builds on the Nuclear Vector Replacement (NVR) framework developed in the Donald Lab. In this framework, a homologous protein structure is used to facilitate the resonance assignments for a target protein. The resonance assignment accuracies are rather high when a very close homologous structure is provided.
My work involves extending the NVR framework such that it returns a high resonance assignment accuracy when such a close homologous structure is not available. In other words, my work tries to extend the circle of convergence of NVR such that it works well with a diverse set of structural templates.
Towards this goal, I use backbone flexibility tools, in order to perturb the backbone structure. My current aim is to find out of the perturbed structures one template which maximizes the resonance assignment accuracy. Some specific problems include determining a scoring function to pick the right structure, making the assignment scheme more robust, and returning multiple assignments that agree with the existing data.
One idea I am currently working on involves determining the confidence level of each of the data sources used in the framework in order to better weight their contribution to the scoring function.
References
- An Expectation/Maximization Nuclear Vector Replacement Algorithm for Automated NMR Resonance Assignments C. Langmead and B. Donald. Journal of Biomolecular NMR 2004; 29(2):111-138.
- High-Throughput 3D Structural Homology Detection via NMR Resonance Assignment. C. Langmead, B. Donald. The IEEE Computational Systems Bioinformatics Conference (CSB), Stanford CA, (August, 2004) pp. 278-289.