M3D_Dropout_Models: Fit functions to the dropouts vs expression distribution.

In M3Drop: Michaelis-Menten Modelling of Dropouts in single-cell RNASeq

Description

Fits the modified Michaelis-Menten equation (MM), a logistic regession (logistic), or a double exponential (ZIFA) function to the relationship between mean expression and dropout-rate (proportion of zero values).

Usage

M3DropDropoutModels(expr_mat, xlim=NA, suppress.plot=FALSE)

Arguments

expr_mat	a numeric matrix of normalized (not log-transformed) expression values, columns = samples, rows = genes.
xlim	limits for x-axis of plot.
suppress.plot	logical, whether to plot fit curves or not.

Details

Plots the dropout-rate (P) vs average gene expression (S) for all genes. Fits three different models and adds the fitted curves to the plot. The three models are: MMfit : the Michaelis-Menten function

P = 1 - S/(K+S)

(see: [1]). LogiFit : a logistic regression between P and log base 10 of S (used by [2]). ExpoFit : a double exponential

P = e^(-lambda*S^2)

(used by [3]).

Value

Invisibly, a list of output from each fit (MMfit, LogiFit, ExpoFit).

References

[1] Keener, J.; Sneyd, J. (2008). Mathematical Physiology: I: Cellular Physiology (2 ed.). Springer. ISBN 978-0-387-75846-6 [2] Kharchenko, PV; Silberstein, L; Scadden, DT. (2014) Bayesian approach to single-cell differential expression analysis. Nature Methods. 11:740-742 [3] Pierson, E; Yau, C. (2015) ZIFA: Dimensionality reduction for zero-inflated single-cell gene expression analysis. Genome Biology. 16:241 doi:10.1186/s13059-015-0805-z

Examples

  library(M3DExampleData)
  norm <- M3DropConvertData(Mmus_example_list$data, is.counts=TRUE)
  M3DropDropoutModels(norm)

Example output

Loading required package: numDeriv
[1] "Removing  18 undetected genes."

M3Drop documentation built on Nov. 8, 2020, 5:06 p.m.