cnvGWAS: Run the CNV-GWAS

In waldronlab/CNVRanger: Summarization and expression/phenotype association of CNV ranges

Run the CNV-GWAS

Description

Wraps all the necessary functions to run a CNV-GWAS using the output of setupCnvGWAS function.

(i) Produces the GDS file containing the genotype information (if produce.gds == TRUE), (ii) Produces the requested inputs for a PLINK analysis, (iii) run a CNV-GWAS analysis using a linear model (i.e. lm function), and (iv) export a QQ-plot displaying the adjusted p-values. In this release only the p-value for the copy number is available (i.e. 'P(CNP)').

Usage

cnvGWAS(
  phen.info,
  n.cor = 1,
  min.sim = 0.95,
  freq.cn = 0.01,
  snp.matrix = FALSE,
  method.m.test = "fdr",
  lo.phe = 1,
  chr.code.name = NULL,
  genotype.nodes = "CNVGenotype",
  coding.translate = "all",
  path.files = NULL,
  list.of.files = NULL,
  produce.gds = TRUE,
  run.lrr = FALSE,
  assign.probe = "min.pvalue",
  correct.inflation = FALSE,
  both.up.down = FALSE,
  verbose = FALSE
)

Arguments

phen.info	Returned by setupCnvGWAS
n.cor	Number of cores to be used
min.sim	Minimum CNV genotype distribution similarity among subsequent probes. Default is 0.95 (i.e. 95%)
freq.cn	Minimum CNV frequency where 1 (i.e. 100%), or all samples deviating from diploid state. Default 0.01 (i.e. 1%)
snp.matrix	Only FALSE implemented - If TRUE B allele frequencies (BAF) would be used to reconstruct CNV-SNP genotypes
method.m.test	Correction for multiple tests to be used. FDR is default, see p.adjust for other methods.
lo.phe	The phenotype to be analyzed in the PhenInfo$phenotypesSam data-frame
chr.code.name	A data-frame with the integer name in the first column and the original name for each chromosome
genotype.nodes	Expression data type. Nodes with CNV genotypes to be produced in the gds file.
coding.translate	For 'CNVgenotypeSNPlike'. If NULL or unrecognized string use only biallelic CNVs. If 'all' code multiallelic CNVs as 0 for loss; 1 for 2n and 2 for gain.
path.files	Folder containing the input CNV files used for the CNV calling (i.e. one text file with 5 collumns for each sample). Columns should contain (i) probe name, (ii) Chromosome, (iii) Position, (iv) LRR, and (v) BAF.
list.of.files	Data-frame with two columns where the (i) is the file name with signals and (ii) is the correspondent name of the sample in the gds file
produce.gds	logical. If TRUE produce a new gds, if FALSE use gds previously created
run.lrr	If TRUE use LRR values instead absolute copy numbers in the association
assign.probe	‘min.pvalue’ or ‘high.freq’ to represent the CNV segment
correct.inflation	logical. Estimate lambda from raw p-values and correct for genomic inflation. Use with argument method.m.test to generate strict p-values.
both.up.down	Check for CNV genotype similarity in both directions. Default is FALSE (i.e. only downstream)
verbose	Show progress in the analysis

Value

The CNV segments and the representative probes and their respective p-value

Author(s)

Vinicius Henrique da Silva

References

da Silva et al. (2016) Genome-wide detection of CNVs and their association with meat tenderness in Nelore cattle. PLoS One, 11(6):e0157711.

See Also

link{setupCnvGWAS} to setup files needed for the CNV-GWAS.

Examples

# Load phenotype-CNV information
data.dir <- system.file("extdata", package="CNVRanger")

phen.loc <- file.path(data.dir, "Pheno.txt")
cnv.out.loc <- file.path(data.dir, "CNVOut.txt")
map.loc <- file.path(data.dir, "MapPenn.txt")

phen.info <- setupCnvGWAS('Example', phen.loc, cnv.out.loc, map.loc)

# Define chr correspondence to numeric, if necessary
df <- '16 1A
25 4A
29 25LG1
30 25LG2
31 LGE22'

chr.code.name <- read.table(text=df, header=FALSE)
segs.pvalue.gr <- cnvGWAS(phen.info, chr.code.name=chr.code.name)
 

waldronlab/CNVRanger documentation built on Nov. 1, 2024, 7:11 a.m.