model.edgeR.genewise: Modeling NB Genewise Dispersion with the Adjusted Profile...
In gu-mi/NBGOF: Goodness-of-Fit Tests and Model Diagnostics for Negative Binomial Regression of RNA Sequencing Data

Description Usage Arguments Details Value Author(s) References See Also

This function fits an NB regression model with genewise dispersions using the adjusted profile likelihood estimator. This function differs from the model.edgeR.tagcom and model.edgeR.tagtrd functions: the tagwise model shrinks the dispersion towards a common dispersion or a global dispersion trend, while the genewise model (implemented by this function) applies no shrinkage. This model is not recommended, and is used for diagnostics only. See details below. The output of this function will be passed to the main GOF function nb.gof.m.

1	model.edgeR.genewise(counts, x, lib.sizes=colSums(counts), min.n=min.n, method=method)

`counts`	an m-by-n count matrix of non-negative integers. For a typical RNA-Seq experiment, this is the read counts with m genes and n samples.
`x`	an n-by-p design matrix.
`lib.sizes`	library sizes of an RNA-Seq experiment. Default is the column sums of the `counts` matrix.
`min.n`	minimim number of genes in a bin. Default is 100. See `dispBinTrend` for details (lower-level function of `estimateGLMTrendedDisp`).
`method`	method for estimating the trended dispersion, including "auto", "bin.spline", "bin.loess", "power" and "spline". If NULL, then the "auto" method. Normally the number of tags analyzed is greater than 200, so the "bin.spline" method is used which calls the `dispBinTrend` function. See `estimateGLMTrendedDisp` for more details.

In this genewise dispersion model, the dispersion parameter φ_i is estimated by maximizing a penalized log-likelihood APL_g(φ_g) plus a weighted shared log-likelihood. The weight denoted by G_0 controls the level of shrinkage applied to purely genewise dispersions towards a common dispersion. In the paper of McCarthy et al. (2012), G_0=20/df performs well over real RNA-Seq datasets. Here the numerator 20 is the prior degrees of freedom, which can be specified by the argument prior.df in the lower-level function dispCoxReidInterpolateTagwise in edgeR. The denominator "df" is the residual degrees of freedom (number of libraries minus the number of distinct treatment groups). We set the value of prior.df equals 0, so that no shrinkage is applied. See the estimateGLMCommonDisp, estimateGLMTagwiseDisp, dispCoxReidInterpolateTagwise and glmFit functions in the edgeR package for more details.