Genomic evolution is a process of change where a mutated genome is selected based on its fitness for survival. What can we say about the history of a genome from its current state? In other words, does a single genome have any evolutionary memory? If yes, can we computationally decode it? This talk will shed light on these questions from computing and information theoretic perspectives. The insights derived from our theoretical findings, augmented with data driven techniques lead us to potential applications in future cancer prediction, DNA storage, and viral evolution.

While working with medical/genomic datasets, we encountered data analysis challenges stemming from sampling bias and limited data. Realizing the generality of these problems for any data scientist, we develop theoretical frameworks to systematically analyze these issues. In particular, we develop (i) a technique to robustly correct sampling bias using CDFs and (ii) a framework of expert graphs to reuse the information available from “built models” (experts) to synthesize new models.

Siddharth Jain is a guest researcher in Electrical Engineering at Caltech. He also recently started serving as a Data Scientist for Rockley Photonics (a digital healthcare company). He obtained his PhD from Caltech in Electrical Engineering under the supervision of Prof. Jehoshua Bruck in 2019. His research interests include computing, information theory, data science and genome evolution. Siddharth received both his Bachelors and Masters degrees from Indian Institute of Technology (IIT) Kanpur, India in 2013 where he was awarded the proficiency medal for the best academic performance in Electrical Engineering. In his spare time, Siddharth enjoys playing cricket and spending time with his family.