B1413
Title: scLANE: single-cell linear adaptive negative-binomial expression testing
Authors: Rhonda Bacher - University of Florida (United States) [presenting]
Abstract: Single-cell RNA-sequencing (scRNA-seq) has advanced the ability to obtain high-resolution views of dynamic biological processes such as cellular differentiation and disease progression. Many methods have emerged that estimate a cell-level ordering from snapshot scRNA-seq samples by using the similarity of gene expression to place cells along a trajectory. With the goal of making biological inferences regarding gene expression across or between trajectories, researchers have typically turned to generalized additive models to capture complex and nonlinear trends. However, their flexibility comes at the cost of interpretability. To address this, single-cell linear adaptive negative-binomial expression (scLANE) testing is developed. The method balances the need for a nonlinear model to accurately characterize changes in expression while enabling direct biological interpretation. The method's accuracy and ability are demonstrated to draw meaningful comparisons on simulated data and case-study datasets having tens of thousands of cells and from multiple subjects.