What question did this study set out to answer?

This research aims to extend methods for analyzing single-cell RNA sequencing data without unique molecular identifiers (UMIs).

June 29, 2026Open Access

Compound models and Pearson residuals for single-cell RNA-seq data without UMIs

JLJan LauseHertie Institute for Clinical Brain Research CZChristoph ZiegenhainScience for Life Laboratory LHLeonard HartmanisKarolinska Institutet

Key Points

This research aims to extend methods for analyzing single-cell RNA sequencing data without unique molecular identifiers (UMIs).
Developed a compound distribution model to account for RNA amplification processes.
Used Pearson residuals to analyze Smart-seq2 datasets.
Examined amplification distributions using a broken power law across various sequencing protocols.
The compound model effectively captures overdispersion patterns, enhancing data interpretation.
Demonstrated improved gene selection in Smart-seq2 datasets.
Identified zero-inflation phenomena in non-UMI scRNA-seq data.

Abstract

Abstract Recent work employed Pearson residuals from Poisson or negative binomial models to normalize UMI-based scRNA-seq data. To extend this approach to non-UMI data, we model the amplification step with a compound distribution: we assume that captured RNA molecules follow a negative binomial distribution and are replicated following an amplification distribution. This model leads to compound Pearson residuals, yielding meaningful gene selection and embeddings of Smart-seq2 datasets. Furthermore, we show that amplification distributions across several sequencing protocols can be described by a broken power law. The resulting compound model captures previously unexplained overdispersion and zero-inflation patterns in non-UMI data.

AI에게 질문

Bookmark

View Full Paper