hm-methods: Jointly analyse multiple genetic perturbation screens using a...

In cbg-ethz/knockout: Statistical Analysis of High-Throughput Genetic Perturbation Screens

Description

Analyse multiple different genetic perturbation screens at once using a hierarchical model. The model estimates general relative effect sizes for genes across all experiments. This could for instance be a pan-pathogenic host factor, i.e. a gene that decisively impacts the life-cycle of multiple pathogens.

Usage

hm(obj, formula = Readout ~ Condition + (1 | GeneSymbol) + (1 |
  Condition:GeneSymbol), drop = TRUE, weights = 1, bootstrap.cnt = 0)

## S4 method for signature 'PerturbationData'
hm(obj, formula = Readout ~ Condition + (1 |
  GeneSymbol) + (1 | Condition:GeneSymbol), drop = TRUE, weights = 1,
  bootstrap.cnt = 0)

Arguments

obj	an PerturbationData object
formula	a formula object that is used to model the readout of your data set. If no formula is provided, the formula 'Readout ~ Condition + (1|GeneSymbol) + (1|Condition:GeneSymbol)' is used. For other data sets with more variables, it might makes sense to use other fixed and random effects
drop	boolean if genes that are not found in every Condition should be dropped
weights	a numeric vector used as weights for the single perturbations
bootstrap.cnt	the number of bootstrap runs you want to do in order to estimate a significance level for the gene effects

Value

returns a HMAnalysedPerturbationData object

Examples

 data(rnaiscreen)
 res <- hm(rnaiscreen)

cbg-ethz/knockout documentation built on Feb. 13, 2020, 7:27 p.m.