fit.ewas: Function to fit ewas to a data.frame or an expressionSet

In isglobal-brge/nlOmicAssoc: non-linar omic association

Description

For each variable (for each row in data) a natural cubic spline with 3 knots is fitted using splines package. Only significative variables obtained after multiple comparison are then included in the multivariate model. Models are fitted using lm and simple least squares. Model is compared to linear model using AIC

Usage

fit.ewas(data = data_m, vars_df = vars_df, df = 3L, fdr = 0.05,
  cores = 1L, verbose = FALSE, ...)

Arguments

data	ExpressionSet or data.frame with samples as columns and observations as rows could be an ExpressionSet(modif!)
vars_df	data.frame with sampes as rows and variables as columns
df	degrees of freedom to apply to model (number of knots in the natural cubic splines)
fdr	threshold to select linear significative vars in terms of FDR adjusted p.value
cores	cores in case of parallelization (no windows)
verbose	logical to verbose (comment) the steps of the function, default(FALSE)

isglobal-brge/nlOmicAssoc documentation built on May 9, 2019, 8:16 a.m.