FeatureSelector: A function used internally for selecting genes based on their...

In yasinkaymaz/HieRFIT: This package provides a utility for information transfer between various data collections using hierarchical random forest

Description

p: pca object, pcs: number of PCs to include, num: number of genes from top and bottom. Weighted gene picking depending on PC number: Initial PCs give more genes. For example; the number of genes are taken from PC i is calculated by (pcs-i+1)*(num*2)/(pcs*(pcs+1))

Usage

Predictors <- FeatureSelector(Data = as.matrix(SeuratObject@data), PCs = 10, num = 2000)

Arguments

Data	an data matrix storing gene expression as genes in rows and samples in columns.
ClassLabels	A list of class labels for cells/samples in the Data matrix. Same length as colnames(Data).
PC_n	the number of PCs to be looked at when selecting genes. Default is 40.
num	the number of Predictors (genes) in total to be included. Default is 2000.
...	parameters to be passed down to subfunctions such as f_n, tree_n, and PC_n, for "number of features per local classifier", "number of trees per local classsifier", and "number of PC space for feature search", respectively.

yasinkaymaz/HieRFIT documentation built on June 1, 2021, 9:37 a.m.