coda_glmnet_longitudinal {coda4microbiome}

coda_glmnet_longitudinal

Description

Microbial signatures in longitudinal studies. Identification of a set of microbial taxa whose joint dynamics is associated with the phenotype of interest. The algorithm performs variable selection through penalized regression over the summary of the log-ratio trajectories (AUC). The result is expressed as the (weighted) balance between two groups of taxa.

Usage

coda_glmnet_longitudinal(
  x,
  y,
  x_time,
  subject_id,
  ini_time,
  end_time,
  covar = NULL,
  lambda = "lambda.1se",
  nvar = NULL,
  alpha = 0.9,
  nfolds = 10,
  showPlots = TRUE,
  coef_threshold = 0
)

Arguments

x

abundance matrix or data frame in long format (several rows per individual)

y

outcome (binary); data type: numeric, character or factor vector

x_time

observation times

subject_id

subject id

ini_time

initial time to be analyzed

end_time

end time to be analyzed

covar

data frame with covariates (default = NULL)

lambda

penalization parameter (default = "lambda.1se")

nvar

number of variables to use in the glmnet.fit function (default = NULL)

alpha

elastic net parameter (default = 0.9)

nfolds

number of folds (default = 10)

showPlots

if TRUE, shows the plots (default = FALSE)

coef_threshold

coefficient threshold, minimum absolute value of the coefficient for a variable to be included in the model (default =0)

Value

in case of binary outcome: list with "taxa.num","taxa.name","log-contrast coefficients","predictions","apparent AUC","mean cv-AUC","sd cv-AUC","predictions plot","signature plot","trajectories plot"

Author(s)

M. Calle - T. Susin

Examples

data(ecam_filtered, package = "coda4microbiome")   # load the data

ecam_results<-coda_glmnet_longitudinal (x=x_ecam[,(1:4)],y= metadata$diet,
x_time= metadata$day_of_life, subject_id = metadata$studyid, ini_time=0,
end_time=60,lambda="lambda.min",nfolds=4, showPlots=FALSE)

ecam_results$taxa.num