Pseudoalignment facilitates assignment of error-prone Ultima Genomics reads

Abstract

We analyze single-cell RNA-seq data sequenced with Ultima Genomics technology and find high error rates in and near homopolymers. To compensate for these errors, we explore the use of pseudoalignment for read assignment, and find that it can perform better than standard read alignment. Our pseudoalignment read assignment for Ultima Genomics data is available as part of the kallisto-bustools kb-python package available at https://github.com/pachterlab/kb_python.