Dr. Shamith Samarajiwa

Group Leader - Computational Biology and Data Science

University of Cambridge, UK

The access to faster computers, the invention of powerful algorithms, and the creation of vast amounts of data have led to data science reshaping various aspects of our society. Medicine and biomedical sciences are no exception, with a tsunami of sequencing, imaging, and health data being generated due to high throughput methods being developed in Genomics and Molecular Cell Biology. This talk will provide a brief history of the emergence of data science and how the modern artificial intelligence field survived multiple AI winters to emerge as a powerful technological force.

The talk will focus on applications in medicine such as the use of image data for predicting cardiovascular risk factors and cancer diagnosis, early detection and treatment, and how algorithms such as AlphaFold solved a 50-year problem in protein structure prediction. Finally, some examples of how machine and deep learning algorithms and data science methods developed in laboratories such as Pangaea, Piranha, and MCMC are being used to predict genome architecture, mine biomedical text, and in cancer early detection. The importance of domain expertise in developing medical algorithms, the necessity for clinical trials of AI algorithms, and limitations and biases in AI-based methods and models will be also discussed.