global.test: A function to run a global test of the null hypothesis that...

In mnodzenski/epistasisGA: An R package to identify multi-snp effects in nuclear family studies using the GADGETS method

A function to run a global test of the null hypothesis that there are no SNP-disease associations across a range of chromosome sizes

Description

This function runs a global test of the null hypothesis that there are no SNP-disease associations across a range of chromosome sizes

Usage

global.test(results.list, n.top.scores = 10)

Arguments

results.list

A list of length d, where d is the number of chromosome sizes to be included in a global test. Each element of the list must itself be a list whose first element observed.data is a vector of fitness scores from combine.islands for the observed data for a given chromosome size. The second element permutation.list is a list containing vectors of all permutation results fitness scores, again using the results output by combine.islands for each permutation.

n.top.scores

The number of top scoring chromosomes, for each chromosome size, to be used in calculating the global test. Defaults to 10.

Value

A list containing the following:

obs.test.stat

The observed test statistic.

perm.test.stats

A vector of test statistics from permuted data.

pval

The p-value for the global test.

obs.marginal.test.stats

A vector of observed test statistics for each chromosome size.

perm.marginal.test.stats.mat

A matrix of test statistics for the permutation datasets, where rows correspond to permutations and columns correspond to chromosome sizes.

marginal.pvals

A vector containing marignal p-values for each chromosome size.

max.obs.fitness

A vector of the maximum fitness score for each chromosome size in the observed data.

max.perm.fitness

A list of vectors for each chromosome size of maximum observed fitness scores for each permutation.

max.order.pvals

A vector of p-values for the maximum observed order statistics for each chromosome size. P-values are the proportion of permutation based maximum order statistics that exceed the observed maximum fitness score.

boxplot.grob

A grob of a ggplot plot of the observed vs permuted fitness score densities for each chromosome size.

chrom.size.k

A vector indicating the number of top scores (k) from each chromosome size that the test used. This will be equal to n.top.scores unless GADGETS returns fewer than n.top.scores unique chromosomes for the observed data or any permute, in which case the chromosome size-specific value will be equal to the smallest number of unique chromosomes returned.

max.perm.95th.pctl

The 95th percentile of the permutation maximum order statistics for each chromosome size.

Examples

data(case)
data(dad)
data(mom)
case <- as.matrix(case)
dad <- as.matrix(dad)
mom <- as.matrix(mom)
data(snp.annotations)
set.seed(1400)

pp.list <- preprocess.genetic.data(case[, 1:10],
                               father.genetic.data = dad[ , 1:10],
                               mother.genetic.data = mom[ , 1:10],
                               ld.block.vec = c(10))
## run GA for observed data

#observed data chromosome size 2
run.gadgets(pp.list, n.chromosomes = 5, chromosome.size = 2,
       results.dir = 'tmp_2',
       cluster.type = 'interactive',
       registryargs = list(file.dir = 'tmp_reg', seed = 1500),
       generations = 2, n.islands = 2, island.cluster.size = 1,
       n.migrations = 0)
 combined.res2 <- combine.islands('tmp_2', snp.annotations[ 1:10, ],
                                  pp.list, 2)
 unlink('tmp_reg', recursive = TRUE)

 #observed data chromosome size 3
 run.gadgets(pp.list, n.chromosomes = 5, chromosome.size = 3,
       results.dir = 'tmp_3',
       cluster.type = 'interactive',
       registryargs = list(file.dir = 'tmp_reg', seed = 1500),
       generations = 2, n.islands = 2, island.cluster.size = 1,
       n.migrations = 0)
 combined.res3 <- combine.islands('tmp_3', snp.annotations[ 1:10, ],
                                  pp.list, 2)
 unlink('tmp_reg', recursive = TRUE)

# create three permuted datasets
set.seed(1400)
perm.data.list <- permute.dataset(pp.list, "perm_data",
                                  n.permutations = 3)

#pre-process permuted data
case.p1 <- readRDS("perm_data/case.permute1.rds")
comp.p1 <- readRDS("perm_data/complement.permute1.rds")
p1.list <- preprocess.genetic.data(case.p1,
                                   complement.genetic.data = comp.p1,
                                    ld.block.vec = c(10))

case.p2 <- readRDS("perm_data/case.permute2.rds")
comp.p2 <- readRDS("perm_data/complement.permute2.rds")
p2.list <- preprocess.genetic.data(case.p2,
                                   complement.genetic.data = comp.p2,
                                    ld.block.vec = c(10))

case.p3 <- readRDS("perm_data/case.permute3.rds")
comp.p3 <- readRDS("perm_data/complement.permute3.rds")
p3.list <- preprocess.genetic.data(case.p3,
                                   complement.genetic.data = comp.p3,
                                   ld.block.vec = c(10))

#permutation 1, chromosome size 2
run.gadgets(p1.list, n.chromosomes = 5, chromosome.size = 2,
       results.dir = 'p1_tmp_2',
       cluster.type = 'interactive',
       registryargs = list(file.dir = 'tmp_reg', seed = 1500),
       generations = 2, n.islands = 2, island.cluster.size = 1,
       n.migrations = 0)
 p1.combined.res2 <- combine.islands('p1_tmp_2', snp.annotations[ 1:10, ],
                                     p1.list, 2)
 unlink('tmp_reg', recursive = TRUE)

#permutation 1, chromosome size 3
run.gadgets(p1.list, n.chromosomes = 5, chromosome.size = 3,
       results.dir = 'p1_tmp_3',
       cluster.type = 'interactive',
       registryargs = list(file.dir = 'tmp_reg', seed = 1500),
       generations = 2, n.islands = 2, island.cluster.size = 1,
       n.migrations = 0)
 p1.combined.res3 <- combine.islands('p1_tmp_3', snp.annotations[ 1:10, ],
                                     p1.list, 2)
 unlink('tmp_reg', recursive = TRUE)

#permutation 2, chromosome size 2
run.gadgets(p2.list, n.chromosomes = 5, chromosome.size = 2,
       results.dir = 'p2_tmp_2',
       cluster.type = 'interactive',
       registryargs = list(file.dir = 'tmp_reg', seed = 1500),
       generations = 2, n.islands = 2, island.cluster.size = 1,
       n.migrations = 0)
 p2.combined.res2 <- combine.islands('p2_tmp_2', snp.annotations[ 1:10, ],
                                     p2.list, 2)
 unlink('tmp_reg', recursive = TRUE)

#permutation 2, chromosome size 3
run.gadgets(p2.list, n.chromosomes = 5, chromosome.size = 3,
       results.dir = 'p2_tmp_3',
       cluster.type = 'interactive',
       registryargs = list(file.dir = 'tmp_reg', seed = 1500),
       generations = 2, n.islands = 2, island.cluster.size = 1,
       n.migrations = 0)
 p2.combined.res3 <- combine.islands('p2_tmp_3', snp.annotations[ 1:10, ],
                                     p2.list, 2)
 unlink('tmp_reg', recursive = TRUE)

#permutation 3, chromosome size 2
run.gadgets(p3.list, n.chromosomes = 5, chromosome.size = 2,
       results.dir = 'p3_tmp_2',
       cluster.type = 'interactive',
       registryargs = list(file.dir = 'tmp_reg', seed = 1500),
       generations = 2, n.islands = 2, island.cluster.size = 1,
       n.migrations = 0)
 p3.combined.res2 <- combine.islands('p3_tmp_2', snp.annotations[ 1:10, ],
                                     p3.list, 2)
 unlink('tmp_reg', recursive = TRUE)

#permutation 3, chromosome size 3
run.gadgets(p3.list, n.chromosomes = 5, chromosome.size = 3,
       results.dir = 'p3_tmp_3',
       cluster.type = 'interactive',
       registryargs = list(file.dir = 'tmp_reg', seed = 1500),
       generations = 2, n.islands = 2, island.cluster.size = 1,
       n.migrations = 0)
 p3.combined.res3 <- combine.islands('p3_tmp_3', snp.annotations[ 1:10, ],
                                     p3.list, 2)
 unlink('tmp_reg', recursive = TRUE)

## create list of results

# chromosome size 2 results
chrom2.list <- list(
    observed.data = combined.res2$fitness.score,
    permutation.list = list(
        p1.combined.res2$fitness.score,
        p2.combined.res2$fitness.score,
        p3.combined.res2$fitness.score
    )
)

# chromosome size 3 results
chrom3.list <- list(
    observed.data = combined.res3$fitness.score,
    permutation.list = list(
        p1.combined.res3$fitness.score,
        p2.combined.res3$fitness.score,
        p3.combined.res3$fitness.score
    )
)

final.results <- list(chrom2.list, chrom3.list)

 lapply(c('tmp_2', 'tmp_3', 'p1_tmp_2', 'p2_tmp_2', 'p3_tmp_2',
          'p1_tmp_3', 'p2_tmp_3', 'p3_tmp_3', 'perm_data'), unlink,
          recursive = TRUE)

mnodzenski/epistasisGA documentation built on Jan. 17, 2023, 7:07 p.m.