Invited Talk - Burkhard Rost (Artificial Intelligence Deciphers the Code of Life Written in Proteins)

Burkhard Rost obtained his doctoral degree (Dr. rer. nat.) from Heidelberg U (Germany) in Theoretical Physics. His research began on spin glasses and brain-like artificial neural networks. A peace/arms control research project sketched non-intrusive sensor networks to monitor aircraft (1988-1990). He entered biology at the European Molecular Biology Laboratory (EMBL Heidelberg, 1990-1995), spent a year at the European Bioinformatics Institute (EBI Cambridgshire, 1995), returned to the EMBL (1996-1998), briefly joined the company LION Biosciences (1998), became faculty in the Medical School of Columbia U in 1998, and joined the TUM in Munich as one of the first Alexander von Humboldt professors in 2009. From 1998-2010, Rost led a research group of ~10 PhD students in the Dept Biochem & Mol Biophysics at Columbia Univ. From 2009-2014, he built a research group at the Technical Univ Munich (TUM). In this issues related to gender, diversity and equal opportunity have been important, as has been the incentive to benefit from the appeal of important science to energize young minds. One particular teaching success were two JavaScript seminars mainly taught by Guy Yachdav using machine learning and data mining to predict Who will die (2016) and Who will survive (2019) in Game Of Thrones; press reports of those each reached over 1 billion readers worldwide. The MOOC course Web app development with the power of Node.js together with Drs. Guy Yachdav & Tatyana Goldberg (originally March-June 2018) became the most successful MOOC from TUM. All major lectures from the Rostlab since 2010 have been published through online videos (since 2015 youtube) downloaded by many thousands of students. In 1992, Dr. Rost developed the first Internet server for protein prediction (PredictProtein) that began at the EMBL, moved to Columbia Univ. and to the TUM, and is now hosted at the LCSB in Luxembourg. PredictProtein is the oldest such service in molecular biology, having gone online before or at the same time as other database lookup services such as ExPasy, SWISS-PROT, or BLOCKS. Since, the group has contributed many highly used and original methods for the prediction of aspects of protein function and structure. His major contribution has been the combination of machine learning and evolutionary information. His academic research goal focuses on the development of top tools that can be applied in the context of analyzing entirely sequence organisms. After having changed the field of protein prediction through the marriage of evolutionary information and machine learning, the group is now seeking to build models for the language of life that allow to side-step the explicit generation of evolutionary information.