coxphParallel: Standard regression functions in R enabled for parallel...

In kevinblighe/RegParallel: Standard regression functions in R enabled for parallel processing over large data-frames

Standard regression functions in R enabled for parallel processing over large data-frames - Cox proportional hazards regression.

Description

This is a non-user function that is managed by RegParallel, the primary function.

Usage

coxphParallel(
  data,
  formula.list,
  FUN,
  variables,
  terms,
  startIndex,
  blocksize,
  blocks,
  APPLYFUN,
  conflevel,
  excludeTerms)

Arguments

data	A data-frame that contains all model terms to be tested. Variables that have all zeros will, automatically, be removed. REQUIRED.
formula.list	A list containing formulae that can be coerced to formula class via as.formula(). REQUIRED.
FUN	Regression function. Must be of form, for example: function(formula, data) glm(formula = formula, family = binomial, data = data). REQUIRED.
variables	Vector of variable names in data to be tested independently. Each variable will have its own formula in formula.list. REQUIRED.
terms	Vector of terms used in the formulae in formula.list, excluding the primary variable of interest. REQUIRED.
startIndex	Starting column index in data object from which processing can commence. REQUIRED.
blocksize	Number of variables to test in each foreach loop. REQUIRED.
blocks	Total number of blocks required to complete analysis. REQUIRED.
APPLYFUN	The apply function to be used within each block during processing. Will be one of: 'mclapply(...)', system=linux/mac and nestedParallel=TRUE; 'parLapply(cl, ...)', system=windows and nestedParallel=TRUE; 'lapply(...)', nestedParallel=FALSE. REQUIRED.
conflevel	Confidence level for calculating odds or hazard ratios. REQUIRED.
excludeTerms	Remove these terms from the final output. These will simply be grepped out. REQUIRED.

Details

This is a non-user function that is managed by RegParallel, the primary function.

Value

A data.table object.

Author(s)

Kevin Blighe <[e-mail]>

Examples

  options(scipen=10)
  options(digits=6)

  col <- 20000
  row <- 20
  mat <- matrix(
    rexp(col*row, rate = .1),
    ncol = col)
  colnames(mat) <- paste0('gene', 1:ncol(mat))
  rownames(mat) <- paste0('sample', 1:nrow(mat))

  modelling <- data.frame(
    cell = rep(c('B', 'T'), nrow(mat) / 2),
    group = c(rep(c('treatment'), nrow(mat) / 2), rep(c('control'), nrow(mat) / 2)),
    dosage = t(data.frame(matrix(rexp(row, rate = 1), ncol = row))),
    mat,
    row.names = rownames(mat))

  require(survival)
  data <- modelling[,1:800]
  variables <- colnames(data)[4:ncol(data)]
  data$time <- c(100,200,400,300,200,250,600,1000,886,450,
    c(100,200,400,300,200,250,600,1000,886,450)*1.5)
  data$alive <- c(0,0,0,0,0,0,0,0,1,1,1,0,0,1,1,1,1,1,1,1)
  res4 <- RegParallel(
    data = data,
    formula = 'Surv(time, as.integer(alive)) ~ group * [*] + cell',
    FUN = function(formula, data)
      coxph(formula = formula,
        data = data,
        ties = 'breslow',
        singular.ok = TRUE),
    FUNtype = 'coxph',
    variables = variables,
    blocksize = 399,
    cores = 2,
    nestedParallel = FALSE,
    p.adjust = "none",
    conflevel = 97.5,
    excludeTerms = c('group', 'cell'),
    excludeIntercept = FALSE
  )

  # spot checks
  m <- coxph(formula = Surv(time, as.integer(factor(alive))) ~ group * gene12 + cell, data = data, ties = 'breslow', singular.ok = TRUE)
  summary(m)
  exp(cbind("Odds ratio" = coef(m), confint.default(m, level = 0.975)))
  res4[which(res4$Variable == 'gene12'),]

  m <- coxph(formula = Surv(time, as.integer(factor(alive))) ~ group * gene267 + cell, data = data, ties = 'breslow', singular.ok = TRUE)
  summary(m)
  exp(cbind("Odds ratio" = coef(m), confint.default(m, level = 0.975)))
  res4[which(res4$Variable == 'gene267'),]

kevinblighe/RegParallel documentation built on Oct. 2, 2023, 2:55 p.m.