get_stats_multiallele: Perform linear regression on multi-allele markers.

View source: R/get_stats_multiallele.R

get_stats_multialleleR Documentation

Perform linear regression on multi-allele markers.

Description

This function performs the multi-allele version of linear regression. It is used in singlescan and one_singlescanDO. It performs marker-by-marker linear regressions for each trait and adjusts for covariates. It collects parameters from the linear models and returns for downstream use.

Usage

get_stats_multiallele(phenotype, genotype, covar_table, ph_family, ref_col)

Arguments

phenotype

A phenotype vector

genotype

genotypes of the marker being tested. If this is a vector, it will be converted into a one-column matrix. If it is a matrix, each column represents an allele, and each row represents an individual.

covar_table

The covariate matrix, with each covariate in a column, and each individual in a row.

ph_family

a character string for a description of the error distribution. Can be either "gaussian" or "binomial"

ref_col

The column belonging to the reference allele. The reference allele is removed from the genotype matrix so that the matrix is linearly independent. There are only n-1 degrees of freedom in the genotype matrix, where n is the number of alleles.

Value

a list with "stats", "pval", "score" stats is a matrix holding the t statistics and slopes (beta coefficients) from the linear model. pval holds the p value for the marker overall and score holds the test statistic for the marker overall.


TheJacksonLaboratory/CAPE documentation built on Feb. 12, 2024, 4:32 p.m.