run.simpleregression: Linear modeling of drug to a binary(or categorical)...

In graeberlab/general:

Description

Linear modeling of drug to a binary(or categorical) covariate. With ability to control for one other covariate, output correlation of continuous value with drug sensitivity as log10 signed pvals.

writes out to a file and returns a data frame of the results.

Usage

run.simpleregression(GeneNumber, GeneName, exp_matrix, Predictor,
  Transpose)

Arguments

drug.frame	columns are samples drugs are rows. Data frame with 'drug' in first column header, and list of drug names underneath.Other columns have sample names as column headers and drug values in the frame itself
categorical.frame	columns are samples continous values are rows. data frame of expression or dependency etc data, 'gene' is first column header with list of gene names or feature names, and list of sample names. Other columns have sample names as column names and values (eg. expression) in the frame itself
drug.name	name of the drug you will be comparing to i.e. column name to extract from the drug frame
type.frame	Default null. If you want to correct for a covariate create a dataframe with headers 'sample' 'type' . sample names in first column , cancer type in second column (or whatever covariate you want to correct for)
output.file	where to write the output file. default is current working directory with name categorical.with.type.signedlog10pvals.txt (nmuts will not make sense if you are not using mutation data (i.e. 1s and 0s))
percent.zeros	remove rows that have more than this percent of zeros 0 to 1 scale. 0.98 (98 percent) is default
keep.na	keep genes with NA in the output. default T
reverse.sign	reverse sign of the output pval table, default F

graeberlab/general documentation built on Nov. 4, 2019, 1:21 p.m.