Prof. Volkan Atalay will give a seminar on “Machine Learning Applications in Bioinformatics” on 29 November at 13:40. The abstract of the talk and a short bio is shared below.

Bio:
After graduating from Ankara Atatürk Anadolu Lisesi, Volkan Atalay obtained B.S. and M.S. in Electrical and Electronics Engineering from the Middle East Technical University (METU). He received his PhD in Computer Science from Université Descartes, Paris, France in 1993. He spent a year at the New Jersey Institute of Technology, USA as a visiting scholar and another year at the Virginia Bioinformatics Institute, Virginia Tech, USA as a visiting faculty member. He is a Professor in the Department of Computer Engineering, METU, where he has been since 1993. He served as Department Chair, Assistant to President and Vice President at METU. From 2010 to 2016 he was the Chair of Board of Directors of ODTÜ TEKNOKENT (Technology Park of METU). He served as a member of Arçelik R&D Advisory Board and Board of Governors of İzmir Biomedicine and Genome Center. He was the recipient of Parlar Foundation Research Incentive Award and METU Graduate School of Natural and Applied Sciences the Best Thesis Award as the supervisor. His main research interests include machine learning applications in bioinformatics, data stream analysis and machine learning applications.

Abstract:
Bioinformatics is an interdisciplinary field that develops methods and software tools for understanding, analyzing, and interpreting biological data while machine learning deals with the development of algorithms that learn how to make predictions based on data. Thanks to the advances in computing and the growth of available data, deep learning algorithms, have been successfully applied in several fields including bioinformatics and cheminformatics. Our research group (https://github.com/cansyl) focuses on developing and applying computational techniques for the analysis of biological data and modeling of biological processes at the molecular level. We develop tools, we participate in challenges and of course we publish articles. Our effort is concentrated in automated protein function prediction and virtual screening for early drug discovery. In the first topic, we have developed DEEPred which is a hierarchical stack of multi-task feed-forward DNN and ECPred for Enzyme Commission (EC) number prediction. In virtual screening for early drug discovery, in addition to drug target interaction (DTI) prediction, we are interested in the representation of compounds and proteins and transfer learning and low-shot learning. To this end, we have developed DEEPScreen (a large-scale DTI prediction system using deep CNN), MDeePred (a large-scale DTI regression system using deep neural networks, iBioProVis (an interactive tool which embeds and visualizes compounds on 2D space) and CROssBAR (a service to generate knowledge graphs.