PyData, AI & Single-Cell Genomics (Berkeley, Yale Speakers)

We are delighted to announce the first PyData Meetup focused on ML methods in Single-cell Genomics. In single-cell genomics, a terrabyte of data is generated from about a milliliter of blood. This kind of data is revolutionizing healthcare and drug development, and if you are wondering how to analyze this kind of data, this is your event!

The event will be held on Zoom Webinar platform and it will be recorded.

Link: https://us02web.zoom.us/webinar/register/WN_JV2dteswSvyuLwA3CT8bMg

Agenda (NY time on all):
10:00-10:30 a.m- Talk #1 By Drausin Wulsin
Reprogramming immunity with AI and single-cell multi-omics
10 minutes Q&A
10:40-11:15 a.m- talk #2 by Dan Burkhardt
Quantifying the effect of experimental perturbations at single-cell resolution
10 minutes Q&A
11:25-11:50 a.m- Talk #3 by Romain Lopez
Deep generative modeling for single-cell transcriptomics
10 minutes Q&A

Talk #1: Reprogramming immunity with AI and single-cell multiomics
Our ability to interrogate and decipher the immune system has dramatically improved over the last 5 years with major advances in single-cell multiomic technology, both in the wet lab and in silico. Immunai has built one of the largest centralized immune single-cell data assets in the world and is using it to expand the boundary of our understanding of core immune biology and how it translates to the clinical setting. But this massive data asset offers a unique challenge in how to understand individual cell types, patients, diseases and treatments in the context of all the others. Immunai tackles this problem with cutting-edge artificial intelligence coupled tightly with our functional genomics platform, which together identify core biological mechanisms that enable us to develop the next generation of immune system therapeutics.

Talk #2: Quantifying the effect of experimental perturbations at single-cell resolution
Current methods for comparing single-cell RNA sequencing datasets collected in multiple conditions focus on discrete regions of the transcriptional state space, such as clusters of cells. We will present a graph-based framework to quantify the effects of perturbations at the single-cell level. We first construct a joint cell similarity graph constructed from multiple samples to approximate an underlying manifold of cellular states. We then develop a kernel density estimate over this graph using the heat kernel. We use these density estimates to calculate the relative likelihood of observing each cell in each of the experimental conditions using graph signal processing. This likelihood estimate can be used to identify cell populations specifically affected by a perturbation. We show how this cell-level estimate provides a fine-grained resolution into the effect of experimental perturbations in single-cell datasets.

Talk #3: Deep generative modeling for single-cell transcriptomics
Single-cell omics measurements can reveal unexplored biological diversity, but they suffer from technical noise and bias that must be modeled to account for the resulting uncertainty in downstream analyses. Consequently, probabilistic models have demonstrated state-of-the-art performance for many single-cell omics data analysis tasks, including dimensionality reduction, clustering, differential expression, annotation, and removal of unwanted variation. As many of these methods exploit scalable stochastic inference techniques, they are also critically important in light of growing single-cell dataset sizes. Here, we present single-cell variational inference (scVI), a ready-to-use scalable framework for the probabilistic representation and analysis of gene expression in single cells. Then, we explore several other applications of deep generative modeling, and finally introduce scvi-tools (https://scvi-tools.org), a high-level interface to probabilistic programming especially designed for single-cell omics data.