evalLogLikZINB: Compute loglikelihood of zero-inflated negative binomial...

In LineagePulse: Differential expression analysis and model fitting for single-cell RNA-seq data

Description

This liklihood function is appropriate for sequencing data with high drop out rate, commonly observed in single cell data (e.g. scRNA-seq). This is the core function used for every likelihood evaluation in LineagePulse, such as maximum likelihood-based estimation. It operates on a vector of counts, such as observations of a gene. Note that for the sake of numerical stability, lower bounds on loglikelihood terms are implemented.

Usage

evalLogLikZINB(vecCounts, vecMu, vecDisp, vecPi, vecidxNotZero, vecidxZero)

Arguments

vecCounts	(count vector number of samples) Observed read counts, not observed are NA.
vecMu	(vector number of samples) Negative binomial mean parameter.
vecDisp	(scalar vector number of samples) Negative binomial dispersion parameters.
vecPi	(probability vector number of samples) Drop-out rate estimates.
vecidxNotZero	(bool vector number of samples) Whether observation is larger than zero.
vecidxZero	(bool vector number of samples) Whether observation is zero.

Value

scaLogLik (scalar) Likelihood under zero-inflated negative binomial model.

Author(s)

David Sebastian Fischer

LineagePulse documentation built on Nov. 8, 2020, 7:01 p.m.