This is the source file for fitting the linear quadratic normal family

isoCorrect(train, data, cycles = 5000, long = FALSE)

Arguments

train

Long data.frame to train model.

data

Long data.frame to correct abundance.

cycles

Number of cycles to reach convergency.

long

Boolean if input is in long format instead of standard wide format (rows:miRNAs, columns:samples).

Value

data.frame with corrected expression

Details

Methods adapted from:

Argyropoulos, Christos, et al. "Modeling bias and variation in the stochastic processes of small RNA sequencing." Nucleic Acids Research (2017).

Examples

options(warn = -1) # this is only for tiny example data(mirTritation) ma <- isoCorrect(mirTritation[mirTritation$class=="train",], mirTritation[mirTritation$class=="test",],cycles=5,long=TRUE)
#> GAMLSS-RS iteration 1: Global Deviance = 39000 #> GAMLSS-RS iteration 2: Global Deviance = 38564 #> GAMLSS-RS iteration 3: Global Deviance = 37919 #> GAMLSS-RS iteration 4: Global Deviance = 36918 #> GAMLSS-RS iteration 5: Global Deviance = 35390 #> GAMLSS-RS iteration 1: Global Deviance = 9211 #> GAMLSS-RS iteration 2: Global Deviance = 8003 #> GAMLSS-RS iteration 3: Global Deviance = 7217 #> GAMLSS-RS iteration 4: Global Deviance = 6800 #> GAMLSS-RS iteration 5: Global Deviance = 6580
library(ggplot2) ggplot(ma,aes(y=log2(reads), x=Dilution)) + geom_jitter()
ggplot(ma,aes(y=m, x=Dilution)) + geom_jitter()