regression: regression analysis

In metabolomicsR: Tools for Metabolomics Data

regression analysis

Description

Run regression models with adjusting for covariates. 'regression_each' is used for one outcome. In 'regression', several outcomes can be specified to run together.

Usage

regression(
  object,
  phenoData = NULL,
  model = NULL,
  outcome = NULL,
  covars = NULL,
  factors = NULL,
  feature_name = NULL,
  time = NULL,
  verbose = TRUE,
  ncpus = 1,
  p.adjust.method = "bonferroni",
  ...
)

regression_each(
  object,
  phenoData = NULL,
  model = NULL,
  formula = NULL,
  outcome = NULL,
  covars = NULL,
  factors = NULL,
  feature_name = NULL,
  time = NULL,
  verbose = TRUE,
  ncpus = 1,
  p.adjust.method = "bonferroni",
  ...
)

Arguments

object	A Metabolite object.
phenoData	A data.table with outcome and covariates. If 'phenoData' is NULL, '@sampleData' will be used.
model	Specify a regression model. See fit_lm for more details. 'auto' can be used to infer 'lm' or 'logistic' (with only 0, 1, and NA).
outcome	Column name of the outcome variable.
covars	Column names of covariates.
factors	Variables to be treated as factor.
feature_name	A vector of selected metabolites to run. If both feature_name and random_select are NULL, will run regression for all features.
time	Column name of survival time, used in cox regression, see coxph for more details.
verbose	Print log information.
ncpus	Number of CPUS for parallele job.
p.adjust.method	Adjust for P value method, see p.adjust.
...	Further arguments passed to regression model.
formula	A character or formula object to fit model (only used in 'regression_each')

Value

term estimate std.error statistic p.value n outcome p.value.adj.

Examples

data(df_plasma)
fit_lm <- regression(object = df_plasma, phenoData = NULL, model = "lm", 
outcome = "BMI", covars = c("AGE", "GENDER", "ETHNICITY"), factors = "ETHNICITY")

metabolomicsR documentation built on April 29, 2022, 9:05 a.m.