evaluate: Calculates prognostic statistics based on m2eQTLs.

Description Usage Arguments Value

View source: R/m2efm.R

Description

Calculates prognostic statistics based on m2eQTLs.

Usage

1
2
3
4
5
6
7
evaluate(meth = NULL, exp = NULL, clin = NULL, eqtls = NULL,
  gene_type = NULL, covariates = c(), gene_sig = NULL, gene_seed = NULL,
  genes = NULL, num_genes = NULL, num_probes = NULL,
  intersect_data = FALSE, integrate_data = FALSE, single_data = "exp",
  time_to_event_col = NULL, event_col = NULL, rescale = TRUE,
  monte_carlo_size = 100, training_prop = 0.7, pam50risk = NULL,
  val = list(), val_clin = list(), val_pam50 = list(), alpha = 0)

Arguments

meth

a MethylationData object, or a data.frame with DNA methylation Beta-values

exp

an ExpressionData object, or a data.frame with gene expression values

clin

a data.frame of clinical data for both meth and exp samples

eqtls

object returned from get_m2eqtls()

gene_type

type of gene list being provided ("trans", "cis", "random", "pre") where trans and cis are those types of m2eQTLs and random is a random set of genes

covariates

a vector of covariates to include from the clinical data

gene_sig

file to store the gene list in (or load it from)

gene_seed

random seed to generate random gene list from

num_genes

number of genes to use for a random gene list

num_probes

number of probes to use for a random probe list

intersect_data

whether to intersect samples for meth and exp (automatically done if integrate_data=TRUE)

integrate_data

whether or not to integrate meth and exp values into a single data frame for the model

single_data

which single data type to use if integrate_data=FALSE, can be "exp" or "meth"

time_to_event_col

name of column with time to event

event_col

name of column with sample event status, or outcome for binomial models

monte_carlo_size

how many random splits of the training data to perform

training_prop

proportion of the data to use for training

val

a list of validation data sets (list of data.frame objects, with combined methylation and expression values or just one or the other)

val_clin

a list of validation clinical data, parallel to val

alpha

a value in [0,1] where 0 is Ridge penalty and 1 is LASSO

Value

a list of results


jeffreyat/M2EFM documentation built on July 11, 2018, 5:27 p.m.