Moleküler Biyoloji ve Genetik Seminerleri: Dr. Thomas Wytock

Moleküler Biyoloji ve Genetik Bölümümüzün organize ettiği seminer serisinin yeni konuğu, Northwestern University’den Dr. Thomas Wytock olacak.

Dr. Wytock’ın “Inferring phenotype from ‘omics data using machine learning” başlıklı konuşmasını 17 Ekim Perşembe saat 16:00’dan itibaren Zoom’dan takip edebilirsiniz.

Etkinlik İngilizcedir.

Zoom ID: 839 7010 6851

Özet: Repositories of ‘omics data continue to expand at an exponential rate, raising the possibility of a comprehensive understanding of cellular function. Cellular phenotypes such as doubling time in single-cell organisms and cell type in humans arise from the collective interaction of myriad proteins, nucleic acids, and metabolites, but divining the specific interactions that mediate a phenotype of interest remains an open question. In this talk, I will describe our approach that uses datasets of thousands of ‘omics measurements to train a supervised k-nearest neighbors (KNN) model that predicts growth rate in microorganisms and cell type in humans with over 80% and 90% accuracy, respectively. Our approach learns which eigenvectors of the gene correlation matrix best describe the phenotype of interest, thereby taking advantage of the multi-scale structure of cellular function. Once trained, the KNN model maps an arbitrary gene expression state to a given phenotype, making it a tool suitable for metabolic engineering, drug discovery, and cell reprogramming.

Konuşmacı Hakkında: I am a Postdoctoral Fellow at Northwestern in Adilson Motter’s group, where I also completed my Ph. D. in Physics. My research interests include complex systems, nonlinear dynamics, bioinformatics, and computational biology. My research goals are (1) to take broad theoretical concepts from physics, network science, and nonlinear dynamics and use these to address problems of collective behavior in biology and (2) to develop methods that facilitate the holistic interpretation of ‘omics data to generate hypotheses.