Bayesian Nonparametric Methods for the Experimental Design of Single Cell Studies: From Clustering to Sorting

• Activity Number: 286 - Advances in Bayesian Nonparametric Methods and Its Applications
• Type: Topic Contributed
• Date/Time: Tuesday, July 30, 2019 : 8:30 AM to 10:20 AM
• Sponsor: International Society for Bayesian Analysis (ISBA)
• Author(s): Bianca Dumitrascu* and Federico Ferrari and Stefano Favaro and Barbara Engelhardt
• Companies: Princeton University and Duke University and Universita di Torino and Princeton University
• Keywords: single-cell data; Thompson sampling; experimental design; Pitman-Yor; Good-Toulmin; Bandit algorithms

Abstract

Bandit algorithms are often the tool of choice for recommendation engines, and have recently seen applications in the context of medical health care data. Here, inspired by bandit ideas, we show a novel application to iterative experimental design in multi-tissue single-cell RNA-seq (scRNA-seq) data. We present two algorithms, a Good-Toulmin like estimator via Thompson sampling and an bayesian nonparametric extension involving a Hierarchical Pitman-Yor prior. Given a budget and modeling cell type information across tissues, they both estimate how many cells are required for sampling from each tissue with the goal of maximizing cell type discovery across samples from multiple iterations. In both real and simulated data, we demonstrate the advantages these algorithms provide in data collection planning when compared to a random strategy in the absence of experimental design.