Learn R Programming

RVtests (version 1.2)

LASSO: LASSO for Rare Variant Tests

Description

Use LASSO for selecting significant variants and testing the variants associated with disease traits.

Usage

LASSO(x, y, family = c("gaussian", "binomial", "poisson", "multinomial", "cox"), alpha = 1, nlambda = 100, lambda.min.ratio, standardize = TRUE, size.max, a = 2, npermutation = 0, npermutation.max, min.nonsignificant.counts)

Arguments

x
Genotype matrix, each row as an individual and each column as a snp
y
Phenotype vector
family
Family: gaussian, binomial, poisson, multinomial, and cox
alpha
alpha = 1 for LASSO, see glmnet
nlambda
see glmnet
lambda.min.ratio
see glmnet
standardize
see glmnet
size.max
Maximum number of variants included
a
Penalty parameter for information criterion, a=2 for AIC.
npermutation
Number of permutation, if less than 1, the permutation will not be run.
npermutation.max
Maximum permutation
min.nonsignificant.counts
Minimum nonsignificant counts

Value

nonsignificant.counts
Counts of permuted data that have a higher score than unpermuted data.
pvalue.empirical
Empirical pvalue via permutation
pvalue.nominal
Not availabe
vs
The selected variants
total.permutation
Total permutation
family
Family

Details

Use glmnet package to implement LASSO and an information criterion (AIC, BIC, or GIC) to select a set of variants.

References

Xu C, Ladouceur M, Dastani Z, Richards JB, Ciampi A, Greenwood CMT. (2012) Multiple Regression Methods Show Great Potential for Rare Variant Association Tests. PLoS ONE 7(8): e41694. doi:10.1371/journal.pone.0041694

See Also

SPLS, glmnet