Algorithms for accurate preprocessing and analysis of single-cell RNA-seq

Sina Booeshaghi

Monday June, 7, 2021
4:00-5:00 PM Eastern
Zoom link https://mit.zoom.us/j/97776274563
Calendar invite

Sina Booeshaghi
Department of Mechanical Engineering
California Institute of Technology

Abstract: We describe algorithms that are used in a new suite of tools, kallisto bustools (Melsted, Booeshaghi, et al.), for accurate and efficient preprocessing of single-cell RNA-seq. Some novel algorithmic developments include stream-wise processing of sequence records for constant-memory barcode error correction and UMI counting, UMI collapsing that leverages sequence diversity for fast count assignment, and nascent/mature mRNA read identification for RNA velocity. We demonstrate these tools on a large-scale spatially-resolved single-cell isoform atlas of the mouse primary motor cortex and briefly discuss extensions of these tools to COVID-19 sequencing-based diagnostics.