Feature selection and dimension reduction for single-cell RNA-Seq based on a multinomial model

By F. W. Townes, Stephanie C. Hicks, Martin J. Aryee, Rafael A. Irizarry. 2019

457 citations recorded by the source corpus.

Single-cell RNA-Seq (scRNA-Seq) profiles gene expression of individual cells. Recent scRNA-Seq datasets have incorporated unique molecular identifiers (UMIs). Using negative controls, we show UMI counts follow multinomial sampling with no zero inflation. Current normalization procedures such as log of counts per million and feature selection by highly variable genes produce false variability in dimension reduction. We propose simple multinomial methods, including generalized principal component analysis (GLM-PCA) for non-normal distributions, and feature selection using deviance. These methods outperform the current practice in a downstream clustering assessment using ground truth datasets.

Subfields: Molecular Biology, Statistics and Probability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Molecular Biology

In Statistics and Probability

Citations in the corpus

The art of using t-SNE for single-cell transcriptomics (2018, methodology): “This is more appropriate for UMI counts because multiplying by 1 million can strongly distort the data after the subsequent log ð 1 þ x Þ transform 51 .”
Orchestrating single-cell analysis with Bioconductor (2019, methodology): “Alternate metrics to variance have also been proposed, such as selecting genes based on their deviance, a metric that quantifies how well each gene fits a null model of constant expression across cells 48 .”
Aitchison’s Compositional Data Analysis 40 Years on: A Reappraisal (2022, background): “As argued by a number of authors, for example [93], the counts of the unique molecular identifiers (UMI) in single-cell expression data can be considered multinomial.”
Probabilistic harmonization and annotation of single‐cell transcriptomics data with deep generative models (2019, unknown): “Briefly, we find that exclusion of zero inflation from the model results in approximately similar performance, which is consistent with findings in (Hafemeister & Satija, 2019; Townes et al , 2019).”
Best practices for single-cell analysis across modalities (2023, background): “Deviance identifies highly informative genes by fitting a gene-wise model that assumes constant expression across all cells and quantifying which genes violate this assumption 47 .”
Comparison and evaluation of statistical error models for scRNA-seq (2021, unknown): “Our sctransform [9] approach utilizes the Pearson residuals from negative binomial regression as input to standard dimensional reduction techniques, while GLM-PCA [10] focuses on a generalized version of principal component analysis (PCA) for data with Poisson-distributed errors.”
Normalization and variance stabilization of single-cell RNA-seq data using regularized negative binomial regression (2019, methodology): “For example, Townes et al. [29] introduced GLM-PCA, a generalization of PCA for data exhibiting non-normal error distributions such as the negative binomial, that takes count data directly as input instead of relying on intermediate residuals.”