Join us for our upcoming SAFARI Live Seminar.
Speaker: Ercument Cicek, Bilkent University (currently on sabbatical at EPFL)
Date: Friday, June 2 2023, 11:00 Zurich time (CEST)
Where: ETZ E81 & Livestream on YouTube (Link)
Title: Detecting Copy Number Variants on Exome Sequencing Data via Deep Learning
Accurate and efficient detection of copy number variants (CNVs) is of critical importance due to their significant association with complex genetic diseases. Although algorithms that use whole genome sequencing (WGS) data provide stable results with mostly-valid statistical assumptions, CNV detection on whole exome sequencing (WES) data shows comparatively lower accuracy. This is unfortunate as WES data is cost efficient, compact and is relatively ubiquitous. The bottleneck is primarily due to non-contiguous nature of the targeted capture: biases in targeted genomic hybridization, GC content, targeting probes, and sample batching during sequencing. In this talk, I will talk about deep learning based techniques that use matched WGS and WES samples to learn making accurate CNV predictions using the WES read depth signal. Our techniques can call both germline and somatic variants, can be fine-tuned to adjust their decision-making according to different ground truth call sets and finally, achieve state-of-the-art precision and recall in mimicking human expert calls.
Ercument Cicek is a Visiting Professor at EPFL on his sabbatical leave from the Computer Engineering Department of Bilkent University where he is working as an Associate Professor. He is also an Adjunct Faculty member at the Computational Biology Department of Carnegie Mellon University. His research is focused on using machine learning techniques to uncover the genetic basis of complex disorders. To achieve this goal, he works on a variety of research questions such as detecting structural variations, assessing gene risks, developing generative models for drug design, and implementing privacy-preserving distributed methods for sharing biomedical data and training models. Ercument‘s research has been funded by the NIH, Simons Foundation, and TUBITAK, and he has received research incentive grants from several institutions including the Turkish Academy of Sciences, Health Institutes of Turkey, the Science Academy, and IEEE Turkey.