Skip to main content

Table 4 High gene variance is not a good indicator of cell type heterogeneity under the alternative hypothesis of zero-inflated negative binomial, because the variance can be lower under the alternative hypothesis

From: Demystifying “drop-outs” in single-cell UMI data

Alternative hypothesis

Variance under H0

Variance under HA

Mixture of Poisson

\(\lambda = \sum _{k}\pi _{k}\lambda _{k}\)

\(\sum _{k} \pi _{k} (\lambda _{k}+\lambda _{k}^{2})-(\sum _{k}\pi _{k}\lambda _{k})^{2}\)

Negative Binomial

λ

\(\frac {\lambda ^{2}}{r} + \lambda \)

Zero-inflated negative binomial

λ

\((1-\pi _{0})^{2} \left (\frac {\lambda ^{2}}{r} + \lambda \right)\)