Fit a Regression Model Under the Null Hypothesis

Share:

Description

fitNullReg fits a regression model. The output of fitNullReg can be passed to assocTestSeq or assocTestSeqWindow for the purpose of genetic association testing.

Usage

1
2
fitNullReg(scanData, outcome, covars = NULL, scan.include = NULL,
            family = gaussian, verbose = TRUE)

Arguments

scanData

An object of class ScanAnnotationDataFrame from the package GWASTools, AnnotatedDataFrame, or class data.frame containing the outcome and covariate data for the samples to be used for the analysis. scanData must have a column scanID containing unique IDs for all samples.

outcome

A character string specifying the name of the outcome variable in scanData.

covars

A vector of character strings specifying the names of the fixed effect covariates in scanData; an intercept term is automatically included. If NULL (default) the only fixed effect covariate is the intercept term.

scan.include

A vector of scanIDs for samples to include in the analysis. If NULL, all samples in scanData are included.

family

A description of the error distribution to be used in the model. The default "gaussian" fits a linear model; see family for further options.

verbose

Logical indicator of whether updates from the function should be printed to the console; the default is TRUE.

Value

A list including:

fixef

A data.frame with effect size estimates (betas), standard errors, chi-squared test statistics, and p-values for each of the fixed effect covariates specified in covars.

betaCov

The estimated covariance matrix of the effect size estimates (betas) of the fixed effect covariates. This can be used for hypothesis tests regarding the fixed effects.

resid.response

The residuals from the model.

logLik

The log-likelihood value.

AIC

The Akaike Information Criterion value.

workingY

The "working" outcome vector. When family is gaussian, this is just the original outcome vector. When family is not gaussian, this is the PQL linearization of the outcome vector. This is used by assocTestSeq for genetic association testing.

model.matrix

The design matrix for the fixed effect covariates used in the model.

aliased

Coefficients removed from the model.

sigma

Variance of the model.

scanID

A vector of scanIDs for the samples used in the analysis.

family

A character string specifying the family used in the analysis.

Author(s)

Matthew P. Conomos