Skip to main content
Fig. 1 | BMC Bioinformatics

Fig. 1

From: Feature selection followed by a novel residuals-based normalization that includes variance stabilization simplifies and improves single-cell gene expression analysis

Fig. 1

Genes with low mean expression exhibit quasi-Poisson variance. In all the plots, each dot represents a gene and the color of the dots reflect the local point density, with brighter shades (yellow) indicating high density and darker shades (deep blue) indicating low density. A Variance (\(\sigma ^2\)) vs mean (\(\mu\)) log-log scatter plots for the Svensson 1 technical control (left panel), NIH/3T3 fibroblast cell line (center panel), and PBMC 33k (right panel) datasets. The solid black line corresponds to the Poisson model (\(\sigma ^2 = \mu\)). For genes with low mean expression levels, the variance can be adequately described by the Poisson model. B Quasi-Poisson dispersion coefficients (\(\alpha _{QP}\)) vs mean (\(\mu\)) log-log scatter plots for the Svensson 1 (left panel), NIH/3T3 (center panel), and PBMC 33k (right panel) datasets. \(\alpha _{QP}\) for each gene were estimated from the observed counts using a regression-based approach

Back to article page

BMC Bioinformatics

ISSN: 1471-2105

Contact us