CIE-SF BME seminar: Immunosequencing

Using Machine Learning to Better Characterize and Understand the Clinical Prognostic Features and Landscape of Immunosequencing

Registration: https://www.eventbrite.com/e/cie-sf-bme-seminar-immunosequencing-tickets-73451093197

About this Event

Abstract:

In recent decades, immunotherapy has been demonstrated as a significant clinical activity in cancer studies. T cells and B cells represent a crucial component of the adaptive immune system and are thought to mediate anti-tumoral immunity. Antigen-specific recognition by T cells is via the T cell receptor (TCR) which is unique for each T cell, clonotype, while B cells can also take up antigens with their BCRs to act as antigen-presenting cells (APC). Next generation sequencing (NGS) of the TCRs/BCRs can be used as a platform to profile the T/B cell repertoire. Though there are a number of software tools available, most of them are either designed for processing repertoire data by mapping antigen receptor segments to sequencing reads and assembling the clonotypes, or only examining the diversity of the TCR/BCR repertoire, they all lack of the ability to track and examine the dynamic nature of the repertoire across serial time points (the treatment effect) or correlate with clinical characteristics and clinical outcomes. We aim to develop customized statistical methods with advanced machine learning techniques to investigate the different aspects of TCR/BCR repertoire data in a clinical context to shift into a new era of classification and pattern recognition of immunosequencing. I will present 1) a comparative study of different feature selection (chi-square filter, correlated-based filter, support vector machine wrapper SVM-RFE, random forest wrapper Boruta), classification (SVM, random forest, bagging, boosting, RBF neural nets, MLP neural nets) and clustering methods (K-means, expectation-maximization); 2) a visualization pipeline of BCR affinity maturation while solving the computational burden by social network analysis.

Speaker:

Li Zhang, Ph.D., is an Associate Professor of Department Medicine and Department of Epidemiology and Biostatistics at University of California San Francisco (UCSF). She is an associate member and principle statistician of UCSF Helen Diller Family Comprehensive Cancer Center (HDFCCC) and faculty statistician for UCSF Clinical Translational Science Institute.

She obtained her Ph.D. in Statistics from University of Florida in 2006. She received pre-doctoral training from Division of Cancer Epidemiology and Genetics at NIH/NCI. Before she joined UCSF in 2013, she was an Assistant Professor at Cleveland Clinic and Case Western Reserve University for about 7 years. She is currently the Visiting Professor or Capital Medical University in Beijing, China.

Prof. Zhang has over fifteen years of experience in applying statistical method in biomedical research including basic science studies, epidemiology studies, clinical research. Her statistical methodological research interests are cancer epidemiology and immunoinformatics. She received multiple research rewards as a principal investigator (PI) and co-Investigator (Co-I), for example, she received UCSF Global Oncology Pilot Project Award on Genome Wide Association Study (GWAS) of Esophageal Cancer in Tanzania and recently received her second UCSF Research Allocation Program Award of Using Machine Learning to Better Characterize and Understand the Clinical Prognostic Features and Landscape of Immunosequencing. She received two Prostate Cancer Foundation Challenge Awards as statistical Co-I. She has more than 80 peer-reviewed publications and has been listed on more than 20 federal or foundation grants as a statistician.

Print