sangerCoxSurv: Fit Cox survival to all variants in a .vcf.gz file from...

In suchestoncampbelllab/gwasurvivr: gwasurvivr: an R package for genome wide survival analysis

Description

Performs survival analysis using Cox proportional hazard models on imputed genetic data stored in compressed VCF files.

Usage

sangerCoxSurv(vcf.file, covariate.file, id.column, sample.ids = NULL,
  time.to.event, event, covariates, inter.term = NULL,
  print.covs = "only", out.file, maf.filter = 0.05,
  info.filter = NULL, chunk.size = 5000, verbose = TRUE,
  clusterObj = NULL)

Arguments

vcf.file	character(1) path to VCF file. This file needs to accompanied with an index file (.tbi file).
covariate.file	matrix(1) comprising phenotype (time, event) and additional covariate data.
id.column	character(1) providing exact match to sample ID column from covariate.file
sample.ids	character vector with sample ids to include in analysis
time.to.event	character(1) string that matches time column name in pheno.file
event	character(1) string that matches event column name in pheno.file
covariates	character vector with matching column names in pheno.file of covariates of interest
inter.term	character(1) string giving the column name of the covariate that will be added to the interaction term with SNP (e.g. term*SNP). See details.
print.covs	character(1) string of either "only", "all" or "some", defining which covariate statistics should be printed to the output. See details.
out.file	character(1) string with output name
maf.filter	integer(1) filter out minor allele frequency below threshold (i.e. 0.005 will filter MAF > 0.005)
info.filter	integer(1) of imputation quality score filter (i.e. 0.7 will filter info > 0.7)
chunk.size	integer(1) number of variants to process per thread
verbose	logical(1) for messages that describe which part of the analysis is currently being run
clusterObj	A cluster object that can be used with the parApply function. See details.

Details

Testing for SNP-covariate interactions: User can define the column name of the covariate that will be included in the interaction term. For example, for given covariates a and b, where c is defined as the inter.term the model will be: ~ a + b + c + SNP + c*SNP.

Printing results of other covariates: print.covs argument controls the number of covariates will be printed as output. The function is set to only by default and will only print the SNP or if an interaction term is given, the results of the interaction term (e.g. SNP*covariate). Whereas, all will print results (coef, se.coef, p.value etc) of all covariates included in the model. some is only applicable if an interaction term is given and will print the results for SNP, covariate tested for interaction and the interaction term. User should be mindful about using the all option, as it will likely slow down the analysis and will increase the output file size.

User defined parallelization: This function uses parApply from parallel package to fit models to SNPs in parallel. User is not required to set any options for the parallelization. However, advanced users who wish to optimize it, can provide a cluster object generated by makeCluster family of functions that suits their need and platform.

Value

Saves two text files directly to disk: .coxph extension containing CoxPH survival analysis results. .snps_removed extension containing SNPs that were removed due to low variance or user-defined thresholds.

Examples

vcf.file <- system.file(package="gwasurvivr",
                       "extdata",
                       "sanger.pbwt_reference_impute.vcf.gz")
pheno.fl <- system.file(package="gwasurvivr",
                        "extdata",
                     "simulated_pheno.txt")
pheno.file <- read.table(pheno.fl,
                         sep=" ",
                         header=TRUE,
                         stringsAsFactors = FALSE)
pheno.file$SexFemale <- ifelse(pheno.file$sex=="female", 1L, 0L)
sample.ids <- pheno.file[pheno.file$group=="experimental",]$ID_2
sangerCoxSurv(vcf.file=vcf.file,
              covariate.file=pheno.file,
              id.column="ID_2",
              sample.ids=sample.ids,
              time.to.event="time",
              event="event",
              covariates=c("age", "SexFemale", "DrugTxYes"),
              inter.term=NULL,
              print.covs="only",
              out.file="sanger_example",
              info.filter=0.3,
              maf.filter=0.005,
              chunk.size=50,
              verbose=TRUE,
              clusterObj=NULL)

suchestoncampbelllab/gwasurvivr documentation built on March 18, 2021, 5:50 p.m.