Package 'geneSLOPE'

Title: Genome-Wide Association Study with SLOPE
Description: Genome-wide association study (GWAS) performed with SLOPE, short for Sorted L-One Penalized Estimation, a method for estimating the vector of coefficients in linear model. In the first step of GWAS, SNPs are clumped according to their correlations and distances. Then, SLOPE is performed on data where each clump has one representative.
Authors: Damian Brzyski [aut], Christine Peterson [aut], Emmanuel J. Candes [aut], Malgorzata Bogdan [aut], Chiara Sabatti [aut], Piotr Sobczyk [cre, aut]
Maintainer: Piotr Sobczyk <[email protected]>
License: GPL-3
Version: 0.38.1
Built: 2025-03-09 03:56:51 UTC
Source: https://github.com/psobczyk/geneslope

Help Index

Clumping procedure for SLOPE

Description

Clumping procedure performed on SNPs, columns of matrix X, from object of class screeningResult, which is an output of function screen_snps. SNPs are clustered based on their correlations. For details see package vignette.

Usage

clump_snps(screenResult, rho = 0.5, pValues = NULL, verbose = TRUE)

Arguments

screenResult

object of class screeningResult
rho

numeric, minimal correlation between two SNPs to be assigned to one clump
pValues

numeric vector, p-values for SNPs computed outside geneSLOPE, eg. with EMMAX
verbose

logical, if TRUE (default) progress bar is shown

Value

object of class clumpingResult

clumpingResult class

Description

A result of procedure for snp clumping produced by clump_snps

Details

Always a named list of eleven elements

  1. X numeric matrix, consists of one snp representative for each clump

  2. y numeric vector, phenotype

  3. SNPnumber numeric vector, which columns in SNP matrix X_all are related to clumps representatives

  4. SNPclumps list of numeric vectors, which columns in SNP matrix X_all are related to clump members

  5. X_info data.frame, mapping information about SNPs from .map file. Copied from the result of screening procedure.

  6. selectedSnpsNumbers numeric vector, which rows of X_info matrix are related to selected clump representatives

  7. X_all numeric matrix, all the snps that passed screening procedure

  8. numberOfSnps numeric, total number of SNPs before screening procedure

  9. selectedSnpsNumbersScreening numeric vector, which rows of X_info data.frame are related to snps that passed screening

  10. pVals numeric vector, p-values from marginal tests for each snp

  11. pValMax numeric, p-value used in screening procedure

See Also

screeningResult clump_snps

Lambda sequences for SLOPE

Description

Computes λ\lambda sequences for SLOPE according to several pre-defined methods.

Usage

create_lambda(n, p, fdr = 0.2, method = c("bhq", "gaussian"))

Arguments

n

number of observations
p

number of variables
fdr

target False Discovery Rate (FDR)
method

method to use for computing λ\lambda (see Details)

Details

The following methods for computing λ\lambda are supported:

  • bhq: Computes sequence inspired by Benjamini-Hochberg (BHq) procedure

  • gaussian: Computes modified BHq sequence inspired by Gaussian designs

Genome-Wide Association Study with SLOPE

Description

Package geneSLOPE performes genome-wide association study (GWAS) with SLOPE, short for Sorted L-One Penalized Estimation. SLOPE is a method for estimating the vector of coefficients in linear model. For details about it see references.

Details

GWAS is splitted into three steps.

  • In the first step data is read using bigmemory package and immediatly screened using marginal tests for each SNP

  • SNPs are clumped based on their correlations

  • SLOPE is performed on data where each clump has one representative (therefore we ensure that variables in linear model are not strognly correlated)

Version: 0.38.1

Author(s)

Malgorzata Bogdan, Damian Brzyski, Emmanuel J. Candes, Christine Peterson, Chiara Sabatti, Piotr Sobczyk

Maintainer: Piotr Sobczyk [email protected]

References

SLOPE – Adaptive Variable Selection via Convex Optimization, Malgorzata Bogdan, Ewout van den Berg, Chiara Sabatti, Weijie Su and Emmanuel Candes

Examples

famFile <- system.file("extdata", "plinkPhenotypeExample.fam", package = "geneSLOPE")
mapFile <- system.file("extdata", "plinkMapExample.map", package = "geneSLOPE")
snpsFile <- system.file("extdata", "plinkDataExample.raw", package = "geneSLOPE")
phe <- read_phenotype(filename = famFile)
screening.result <- screen_snps(snpsFile, mapFile, phe, pValMax = 0.05, chunkSize = 1e2)
clumping.result <- clump_snps(screening.result, rho = 0.3, verbose = TRUE)
slope.result <- select_snps(clumping.result, fdr=0.1)

## Not run: 
gui_geneSLOPE()

## End(Not run)

GUI for GWAS with SLOPE

Description

A graphical user interface for performing Genome-wide Association Study with SLOPE

Usage

gui_geneSLOPE()

Details

requires installing shiny package

Value

null

identify_clump

Description

identify_clump

Usage

identify_clump(x, ...)

Arguments

x

appropiate class object
...

other arguments

Details

Enable interactive selection of snps in plot. Return clump number.

Identify clump number in clumpingResult class plot

Description

Identify clump number in clumpingResult class plot

Usage

## S3 method for class 'clumpingResult'
identify_clump(x, ...)

Arguments

x

clumpingResult class object
...

Further arguments to be passed to or from other methods. They are ignored in this function.

Identify clump number in selectionResult class plot

Description

Identify clump number in selectionResult class plot

Usage

## S3 method for class 'selectionResult'
identify_clump(x, ...)

Arguments

x

selectionResult class object
...

Further arguments to be passed to or from other methods. They are ignored in this function.

phenotypeData class

Description

Phenotype data

Details

Always a named list of two elements

  1. y numeric vector, phenotype

  2. yInfo data.frame, additional information about observations provied in .fam file

See Also

read_phenotype

Plot selectionResult class object

Description

Plot selectionResult class object

Usage

## S3 method for class 'selectionResult'
plot(x, chromosomeNumber = NULL, clumpNumber = NULL, ...)

Arguments

x

selectionResult class object
chromosomeNumber

optional parameter, only selected chromosome will be plotted
clumpNumber

optional parameter, only SNPs from selected clump will be plotted
...

Further arguments to be passed to or from other methods. They are ignored in this function.

Print clumpingResult class object

Description

Print clumpingResult class object

Usage

## S3 method for class 'clumpingResult'
print(x, ...)

Arguments

x

clumpingResult class object
...

Further arguments to be passed to or from other methods. They are ignored in this function.

Print phenotypeData class object

Description

Print phenotypeData class object

Usage

## S3 method for class 'phenotypeData'
print(x, ...)

Arguments

x

phenotypeData class object
...

Further arguments to be passed to or from other methods. They are ignored in this function.

Print function for class screeningResult class

Description

Print function for class screeningResult class

Usage

## S3 method for class 'screeningResult'
print(x, ...)

Arguments

x

screeningResult class object
...

Further arguments to be passed to or from other methods. They are ignored in this function.

Print selectionResult class object

Description

Print selectionResult class object

Usage

## S3 method for class 'selectionResult'
print(x, ...)

Arguments

x

selectionResult class object
...

Further arguments to be passed to or from other methods. They are ignored in this function.

Value

Nothing.

Read phenotype from .fam file

Description

Reading phenotype data from file. It is assumed, that data is given in .fam file. In this format, first column is family id (FID), second is individual id (IID), third is Paternal individual ID (PAT), fourth is Maternal individual ID (MAT), fifth is SEX and sixth and last is PHENOTYPE. If file has only four columns, then it is assumed that PAT and MAT columns are missing. If there is only one column, then it is assumed that only phenotype is provided.

Usage

read_phenotype(filename, sep = " ", header = FALSE, stringAsFactors = FALSE)

Arguments

filename

character, name of file with phenotype
sep

character, field seperator in file
header

logical, does first row of file contain variables names
stringAsFactors

logical, should character vectors be converted to factors?

Value

object of class phenotypeData

Reading and screening SNPs from .raw file and

Description

Reading .raw file that was previously exported from PLINK - see details. Additional information about SNP mapping is read from .map file.

Usage

screen_snps(
  rawFile,
  mapFile = "",
  phenotype,
  pValMax = 0.05,
  chunkSize = 100,
  verbose = TRUE
)

Arguments

rawFile

character, name of .raw file
mapFile

character, name of .map file
phenotype

numeric vector or an object of class phenotypeData
pValMax

numeric, p-value threshold value used for screening
chunkSize

integer, number of snps that will be processed together. The bigger chunkSize is, the faster function works but computer might run out of RAM
verbose

if TRUE (default) information about progress is printed

Details

Exporting data from PLINK To import data to R, it needs to be exported from PLINK using the option "–recodeAD" The PLINK command should therefore look like plink --file input --recodeAD --out output. For more information, please refer to: http://pngu.mgh.harvard.edu/~purcell/plink/dataman.shtml

Value

object of class screeningResult

screeningResult class

Description

A result of procedure for snp clumping produced by screen_snps

Details

Always a named list of eight elements

  1. X numeric matrix, consists of snps that passed screening

  2. y numeric vector, phenotype

  3. X_info data.frame, SNP info from .map file

  4. pVals numeric vector, p-values from marginal tests for each snp

  5. numberOfSnps numeric, total number of SNPs in .raw file

  6. selectedSnpsNumbers numeric vector, which rows of X_info data.frame are related to snps that passed screening

  7. pValMax numeric, p-value used in screening procedure

  8. phenotypeInfo data.frame, additional information about observations provied in phenotypeData object

See Also

phenotypeData screen_snps

GWAS with SLOPE

Description

Performs GWAS with SLOPE on given snp matrix and phenotype. At first clumping procedure is performed. Highly correlated (that is stronger than parameter rho) snps are clustered. Then SLOPE is used on snp matrix which contains one representative for each clump.

Usage

select_snps(
  clumpingResult,
  fdr = 0.1,
  type = c("slope", "smt"),
  lambda = "gaussian",
  sigma = NULL,
  verbose = TRUE
)

Arguments

clumpingResult

clumpProcedure output
fdr

numeric, False Discovery Rate for SLOPE
type

method for snp selection. slope (default value) is SLOPE on clump representatives, smt is Benjamini-Hochberg procedure on single marker test p-values for clump representatives
lambda

lambda for SLOPE. See create_lambda
sigma

numeric, sigma for SLOPE
verbose

logical, if TRUE progress bar is printed

Value

object of class selectionResult

Examples

## Not run: 
slope.result <- select_snps(clumping.result, fdr=0.1)

## End(Not run)

selectionResult class

Description

A result of applying SLOPE to matrix of SNPs obtained by clumping produced. Result of function select_snps

Details

Always a named list of eighteen elements

  1. X numeric matrix, consists of one snp representative for each clump selected by SLOPE

  2. effects numeric vector, coefficients in linear model build on snps selected by SLOPE

  3. R2 numeric, value of R-squared in linear model build on snps selected by SLOPE

  4. selectedSNPs which columns in matrix X_all are related to snps selected by SLOPE

  5. y selectedClumps list of numeric vectors, which columns in SNP matrix X_all are related to clump members selected by SLOPE

  6. lambda numeric vector, lambda values used by SLOPE procedure

  7. y numeric vector, phenotype

  8. clumpRepresentatives numeric vector, which columns in SNP matrix X_all are related to clumps representatives

  9. clumps list of numeric vectors, which columns in SNP matrix X_all are related to clump members

  10. X_info data.frame, mapping information about SNPs from .map file. Copied from the result of clumping procedure

  11. X_clumps numeric matrix, consists of one snp representative for each clump

  12. X_all numeric matrix, all the snps that passed screening procedure

  13. selectedSnpsNumbers numeric vector, which rows of X_info data.frame are related to snps that were selected by SLOPE

  14. clumpingRepresentativesNumbers numeric vector, which rows of X_info data.frame are related to snps that are clump represenatives

  15. screenedSNPsNumbers numeric vector, which rows of X_info data.frame are related to snps that passed screening

  16. numberOfSnps numeric, total number of SNPs before screening procedure

  17. pValMax numeric, p-value used in screening procedure

  18. fdr numeric, false discovery rate used by SLOPE

See Also

screeningResult clumpingResult select_snps SLOPE

Summary clumpingResult class object

Description

Summary clumpingResult class object

Usage

## S3 method for class 'clumpingResult'
summary(object, ...)

Arguments

object

clumpingResult class object
...

Further arguments to be passed to or from other methods. They are ignored in this function.

Summary phenotypeData class object

Description

Summary phenotypeData class object

Usage

## S3 method for class 'phenotypeData'
summary(object, ...)

Arguments

object

phenotypeData class object
...

Further arguments to be passed to or from other methods. They are ignored in this function.

Summary function for class screeningResult

Description

Summary function for class screeningResult

Usage

## S3 method for class 'screeningResult'
summary(object, ...)

Arguments

object

screeningResult class object
...

Further arguments to be passed to or from other methods. They are ignored in this function.

Summary selectionResult class object

Description

Summary selectionResult class object

Usage

## S3 method for class 'selectionResult'
summary(object, clumpNumber = NULL, ...)

Arguments

object

selectionResult class object
clumpNumber

number of clump to be summarized
...

Further arguments to be passed to or from other methods. They are ignored in this function.