Posts classified under: Biomedical Data Science Seminar Series

HomeStanford - Department of Biomedical Data Science  > Pro Event Calendar  > Biomedical Data Science Seminar Series

We used RL but…. Did it work?!

Thursday, May 20, 2021, 2:30-3:50pm, virtual access only

Susan Murphy

Professor of Statistics and of Computer Science, and Radcliffe Alumnae Professor at the Radcliffe Institute, Harvard University

Seminar Title: We used RL  but…. Did it work?!

Abstract: Digital Healthcare is a growing area of importance in modern healthcare due to its potential in helping individuals improve their behaviors so as to better manage chronic health challenges such as hypertension, mental health, cancer and so on. Digital apps and wearables, observe the user’s state via sensors/self-report, deliver treatment actions (reminders, motivational messages, suggestions, social outreach,…) and observe rewards repeatedly on the user across time. This area is seeing increasing interest by reinforcement learning (RL) researchers with the goal of including in the digital app/wearable an RL algorithm that “personalizes” the treatments to the user. But after RL is run on a number of users, how do we know whether the RL algorithm actually personalized the sequential treatments to the user?  In this talk we report on our first efforts to address this question after our RL algorithm was deployed on individuals with hypertension.

Interrogating the Gut Microbiome: Estimation of Bacterial Growth Rates and Prediction of Biosynthetic Gene Clusters

Thursday, April 15, 2021, 2:30-3:50pm, virtual access only

Hongzhe Li (Lee)

Perelman Professor of Biostatistics, Epidemiology and Informatics, University of Pennsylvania School of Medicine

Seminar Title: Interrogating the Gut Microbiome: Estimation of Bacterial Growth Rates and Prediction of Biosynthetic Gene Clusters

Abstract: The gut microbiome plays an important role in maintenance of human health. High-throughput shotgun metagenomic sequencing of a large set of samples provides an important tool to interrogate the gut microbiome. Besides providing footprints of taxonomic community composition and genes, these data can be further explored to study the bacterial growth dynamics and metabolic potentials via generation of small molecules and secondary metabolites. In this talk, Dr. Lee will present several computational and statistical methods for estimating the bacterial growth rate for metagenome-assembled genomes (MAGs) and for predicting all biosynthetic gene clusters (BGCs) in bacterial genomes. The key statistical and computational tools used include optimal permutation recovery based on low-rank matrix projection and improved LSTM deep learning methods to improve prediction of BGCs. He will demonstrate the application of these methods using several ongoing microbiome studies of inflammatory bowel disease at the University of Pennsylvania.

Removing unwanted variation and quantifying cellular aging in single cell experiments

BIODS 260B
Nancy Zhang
Professor of Statistics and Data Science
Vice Dean of Wharton Doctoral Programs
3/16/2023
1:30 pm-2:50 pm
MSOB X303 (SEE ZOOM DETAILS BELOW)

Title: Removing unwanted variation and quantifying cellular aging in single cell experiments

Abstract: I will discuss two common problems in single cell experiments. The first is the pervasive issue of unwanted variation (batch effects).  These can be technical noise but can also be biological variation that is not of interest to a study. I will show that current paradigms for batch integration (also called “cell alignment/integration”) are overly aggressive, removing biologically meaningful variation.  I will describe a novel statistical procedure, cellanova, which makes use of a “pool-of-controls” design that is feasible in both clinical and laboratory settings, to separate unwanted variation from biological variation of interest.  This new framework is validated across diverse settings.  I will summarize the validations and illustrate using a few examples.

Next, I will shift gears and discuss another type of “noise” in single cell data: the intrinsic transcriptional  “noise” of cells, as defined by the classic experiments by Elowitz, Levine, Siggia and Swain (2002).  Commonly ignored in current analysis pipelines,  we show that intrinsic transcriptional noise is not only estimable from single cell data but also uniquely informative about cellular senescence.  Through analysis of data from in vitro cell irradiation experiments, from the Aging Mouse Atlas, and from a system of telomerase dysfunction-induced cellular aging, we show that while aging has varying effects across cell types, intrinsic transcriptional noise can serve as a universal and sensitive marker of cellular senescence.

The first part of this talk is joint work with Zhaojun Zhang and Zongming Ma, while the second part of this talk is joint work with Paul Hess and Bradley Johnson.

Website: https://statistics.wharton.upenn.edu/profile/nzh/

Zoom link: https://stanford.zoom.us/j/92124459914?pwd=cFpJYXVLOExUVjMzZkNsYXA0b0RxUT09&from=addon

Password: 840608

1 2