rowTrendFuzzy: Trend Test for Fuzzy Genotype Calls

In scrime: Analysis of High-Dimensional Categorical Data Such as SNP Data

Description

rowTrendFuzzy performs the trend test proposed by Louis et al. (2010) based on fuzzy genotype calls, i.e. the weighted sums over the confidences for the three genotypes as they are determined by preprocessing algorithms (e.g., CRLMM) or imputation procedures.

Given the confidences and scores for all three genotypes, getMatFuzzy constructs a matrix containing the fuzzy genotype calls.

Usage

rowTrendFuzzy(score, probs, y, mat.fuzzy = NULL, 
    alternative = c("two.sided", "less", "greater"), 
    check = TRUE)
    
getMatFuzzy(score, probs, check = TRUE)

Arguments

score	either a numeric vector of length 2 or 3, or a character string. If the latter, score must be either "additive", "dominant", "recessive", or an abbreviation of these terms. If score = "additive", then score is set to c(0, 1, 2). If score = "dominant", then score is set to c(0, 1, 1). And if score = "recessive", score is set to c(0, 0, 1). If score is a numeric vector of length 2, then the first value must be the score for the heterozygous genotype, and the second value the score for the homozygous variant genotype. If score is of length, the first entry of this vector must be a zero (the score for the homozygous reference genotype), followed by the scores for the heterozygous and homozygous variant genotype.
probs	a list of length 2 or 3 consisting of matrices of the same size. Each matrix must contain the confidences for one of the three genotypes, where each row in the matrix represents a SNP and each column a sample (which must be in the same order in all matrices). The matrices in probs correspond to the scores in score. Thus, if probs has length 2, then the first matrix must contain the confidences for the heterozygous genotype, and the second matrix the confidences for the homozygous variant genotype. All elementwise sums of these two matrices must be smaller than or equal to 1. If probs has length 3, the first object must be a matrix containing the confidences for the homozygous reference genotype, followed by the two other matrices comprising the confidences for the heterozygous and homozygous variant genotype. All elementwise sums of the three matrices must be equal to 1.
y	a vector of zeros and ones specifying which of the samples in the matrices in probs are cases (1) and which are controls (0).
mat.fuzzy	a matrix containing the fuzzy genotype calls. If specified, score and probs are not allowed to be specified in rowTrendFuzzy. If NULL, mat.fuzzy is determined by employing getMatFuzzy, i.e.\ by multiplying the confidences in probs with the corresponding scores in score and computing the elementwise sums over the resulting matrices with the weighted confidences.
alternative	a character string specifying the alternative hypothesis. Must be one of "two.sided" (default), "greater", or "less". Abbreviations (e.g.\ the initial letter) for these choices are also allowed.
check	logical specifying whether the specified objects should be extensively checked. If FALSE, only some basic checks are done. It is highly recommended to use check = TRUE, although the checking takes much more computing time than the determination of the values of the trend test statistic.

Value

For getMatFuzzy, a matrix containing the fuzzy genotype calls. For rowTrendFuzzy, a list consisting of

stat	a vector containing the values of the trend test statistic for all SNPs comprised by probs,
rawp	a vector containing the unadjusted p-values computed for the values in stat based on a standard normal distribution and the specification of alternative.
theta	a vector containing estimates for the log odds ratios for risk corresponding to stat,
varTheta	a vector containing the variance estimates for theta.

References

Louis, T.A., Carvalho, B.S., Fallin, M.D., Irizarry, R.A., Li, Q., and Ruczinski, I. (2010). Association Tests that Accommodate Genotyping Errors. In Bernardo, J.M., Bayarri, M.J., Berger, J.O., Dawid, A.P., Heckerman, D., Smith, A.F.M., and West, M. (eds.), Bayesian Statistics 9, 393-420. Oxford University Press, Oxford, UK. With Discussion.

See Also

rowTrendStats, rowCATTs

scrime documentation built on May 2, 2019, 10:24 a.m.