| Date |
Sept 15, 2015 |
| Speaker |
Jiangning Song
|
| Title |
The Predictive Power of Machine Learning Techniques in Data-driven Biomedical Knowledge Discovery: My Structural Bioinformatics Journey
|
| Abstract |
Structural bioinformatics is the branch of bioinformatics which is concerned with the analysis and prediction of the three-dimensional structure of biological macromolecules on a genomic scale by developing computational methods. Recently, machine learning techniques based on statistical learning have provided efficient solutions to challenging problems that were previously considered difficult to address. In this talk, by combining my research experiences, I highlight some important recent developments in the prediction and analysis of functional residues or sites that are based on such methods. In particular, I focus on two tasks in structural bioinformatics, i.e. predicting protease-specific substrate cleavage sites and predicting enzyme catalytic sites from sequence and/or structural information. I dig deeper into the predictions, showing how machine learning methods can extract the predictive power and to what extent heterogeneous features derived at different levels (i.e. sequence, structure, and network) of the data samples can contribute to the model’s performance. Some of the existing difficult issues/problems will also be discussed in the talk.
|
|
アブストラクト