findHVG: Find highly variable genes

Description Usage Arguments Value Author(s)

Description

It finds highly variable genes using the normalised data stored in the countsNorm slot within the RNAseq object provided. It follows Brennecke et al., Nature Methods, 2013. [email protected] vector is required. If any entry is true, it will take spike-ins to calculate the fit. If none is true, it will take biological genes to calculate the fit.

Usage

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
findHVG(object, signThres = 0.1, outputPlots = "./",
  plotting = "pdf", colVarGenes = "deeppink", UseSpike = TRUE,
  minQuantCv2 = 0.2, maxQuantCv2 = 0.8, minQuantMeans = 0.2,
  maxQuantMeans = 0.8)

## S4 method for signature 'RNAseq'
findHVG(object, signThres = 0.1, outputPlots = "./",
  plotting = "pdf", colVarGenes = "deeppink", UseSpike = TRUE,
  minQuantCv2 = 0.2, maxQuantCv2 = 0.8, minQuantMeans = 0.2,
  maxQuantMeans = 0.8)

Arguments

object

RNAseq object.

signThres

Threshold to adjust for multiple testing with the Benjamini-Hochberg method. Default: 0.1 (cut at 10 percent).

outputPlots

state directory where you want to output the plots. Default: Current directory.

plotting

It states whether plots are generated or not and you should specify the type. Options: "pdf", "tiff", "no". Default: "pdf". Default: TRUE

colVarGenes

Color for variable genes. Default: deeppink.

UseSpike

Logical vector stating whether to use Spike-ins to calculate the fitting curve or not. Default: NULL (it will depend on whether the SpikeIn slot has spike ins).

minQuantCv2

Lower quantile of cv2 values to discard to fit the data. Default: 0.2. This means that those genes with cv2 (across cells) in the lower 0.2 quantile will be excluded for the fit.

maxQuantCv2

Upper quantile of cv2 values to discard to fit the data. Default: 0.8. This means that those genes with cv2 (across cells) in the top 0.8 quantile will be discarded for the fit.

minQuantMeans

Lower quantile of means to discard to fit the data. Default: 0.2. This means that those genes with mean (across cells) in the 0.2 lower quantile will be discarded for the fit.

maxQuantMeans

Upper quantile of means to discard to fit the data. Default: 0.8. This means that those genes with mean (across cells) in the 0.8 upper quantile will be discarded for the fit.

Value

vector of highly variable genes. This should be stored in genesHVG slot within RNAseq object.

Author(s)

Blanca Pijuan Sala.


BPijuanSala/proSeq documentation built on Oct. 10, 2018, 7:20 p.m.