tests/testthat/test_fishhook_ops.R
In mskilab/fish.hook: R Package for performing Gamma-Poisson regression on somatic mutation count data
library(fishHook)
library(testthat)
library(zoo)
## Sys.setenv(DEFAULT_BSGENOME = 'http://mskilab.com/gUtils/human_g1k_v37.chrom.sizes')
Sys.setenv(DEFAULT_BSGENOME = system.file('extdata', 'human_g1k_v37.chrom.sizes', package = 'fishHook'))
## hypotheses BED
## hypothesesbed = "http://mskilab.com/fishHook/tests/targets.bed"
## hypothesesbed = system.file("extdata", "targets.bed", package = "fishHook")
hypothesesbed = testthat::test_path("data", "targets.bed")
## Sample Covariate #replication_timing = readRDS(gzcon(file("http://mskilab.com/fishHook/tests/covariate.rds")))
## replication_timing = readRDS(gzcon(file("http://mskilab.com/fishHook/tests/covariate.rds")))
replication_timing = readRDS(testthat::test_path("data", "covariate.rds"))
## Same Eligible Subset
## eligible = readRDS(gzcon(file("http://mskilab.com/fishHook/tests/eligible.rds")))
eligible = readRDS(testthat::test_path("data", "eligible.rds"))
## Sample Hypotheses
#hypotheses = readRDS(gzcon(file("http://mskilab.com/fishHook/tests/targets.rds")))
hypotheses = readRDS(testthat::test_path("data", "targets.rds"))
hypotheses.2 = readRDS(testthat::test_path("data", "hypotheses.rds"))
## indexed pathways
#indexed_pathways = readRDS(gzcon(file("http://mskilab.com/fishHook/tests/indexed_pathways.rds")))
## indexed_pathways = readRDS(gzcon(file("http://mskilab.com/fishHook/tests/indexed_pathways.rds")))
indexed_pathways = readRDS(testthat::test_path("data", "indexed_pathways.rds"))
annotated_cov = readRDS(testthat::test_path("data", "annotated_cov.rds"))
## segs
## segs = readRDS(gzcon(file("http://mskilab.com/fishHook/tests/jabba_segs_11517.rds")))
segs = readRDS(testthat::test_path("data", "jabba_segs_11517.rds"))
#segs = readRDS(gzcon(file("http://mskilab.com/fishHook/tests/jabba_segs_11517.rds")))
## eligible
## eligible = readRDS(gzcon(file("http://mskilab.com/fishHook/tests/eligible.rds")))
# Sample annotate
#anno = readRDS(gzcon(file("http://mskilab.com/fishHook/tests/anno.rds")))
## anno = readRDS(gzcon(file( "http://mskilab.com/fishHook/tests/anno.rds")))
anno = readRDS(testthat::test_path("data", "anno.rds"))
context('unit testing fishhook operations')
test_that('annotate.hypotheses', {
## default args
t1 = hypotheses[1]
e1 = events[1]
el1 = eligible[1]
r1 = replication_timing[1]
expect_error({annotate.hypotheses(hypotheses = hypotheses[1], events = events[1], idcol = 'hello')})
## c1 = Covariate$new(data = readRDS(gzcon(file("http://mskilab.com/fishHook/tests/covariate.rds"))), type = 'numeric', name = 'c1', pad = 10)
c1 = Covariate$new(data = replication_timing, type = 'numeric', name = 'c1', pad = 10)
c2 = Covariate$new(data = replication_timing, type = 'interval', name = 'c2', pad = 10, na.rm = TRUE)
## c2 = Covariate$new(data = readRDS(gzcon(file("http://mskilab.com/fishHook/tests/covariate.rds"))), type = 'interval', name = 'c2', pad = 10, na.rm = TRUE)
c3 = Covariate$new(data = list(RleList()), type = 'interval', name = 'c3', pad = 10, na.rm = TRUE)
anno = annotate.hypotheses(hypotheses = t1, events = e1, covariates = c1$toList())
anno2 = annotate.hypotheses(hypotheses = t1, events = e1, covariates = c2$toList())
#anno3 = annotate.hypotheses(hypotheses = t1, events = e1, covariates = c3$toList())
anno4 = annotate.hypotheses(hypotheses = e1, events = e1, eligible = e1)
expect_true(is(annotate.hypotheses(hypotheses), 'GRanges')) ## annotate.hypotheses(hypotheses, weightEvents=TRUE)
expect_equal(length(annotate.hypotheses(hypotheses)), 19688)
foo = annotate.hypotheses(hypotheses)
expect_equal(max(foo$query.id), 19688)
expect_equal(max(foo$eligible), 2304638)
expect_equal(max(foo$count), 0)
## covered
## events
## mc.cores
expect_true(is(annotate.hypotheses(hypotheses, mc.cores=2), 'GRanges'))
expect_equal(length(annotate.hypotheses(hypotheses, mc.cores=2)), 19688)
expect_equal(max(annotate.hypotheses(hypotheses, mc.cores=2)$query.id), 19688)
expect_equal(max(annotate.hypotheses(hypotheses, mc.cores=2)$eligible), 2304638)
expect_equal(max(annotate.hypotheses(hypotheses, mc.cores=2)$count), 0)
## na.rm
## pad
## verbose
expect_true(is(annotate.hypotheses(hypotheses, verbose=FALSE), 'GRanges'))
expect_equal(length(annotate.hypotheses(hypotheses, verbose=FALSE)), 19688)
expect_equal(max(annotate.hypotheses(hypotheses, verbose=FALSE)$query.id), 19688)
expect_equal(max(annotate.hypotheses(hypotheses, verbose=FALSE)$eligible), 2304638)
expect_equal(max(annotate.hypotheses(hypotheses, verbose=FALSE)$count), 0)
## max.slice
## annotate.hypotheses(hypotheses, max.slice=1e8)
## ff.chunk
##
## max.chunk
## out.path
## covariates
## maxpatientpergene
## ptidcol
## weightEvents
expect_equal(is.na(all(annotate.hypotheses(hypotheses, weightEvents=TRUE)$count)), TRUE)
## if (is.character(hypotheses)){
## expect_equal(length(annotate.hypotheses('/home/travis/build/mskilab/fishHook/data/hypotheses.rds')), 19688)
## if (length(hypotheses)==0){
expect_error(annotate.hypotheses(GRanges()))
## if (!is.null(out.path)){
## expect_equal(length(annotate.hypotheses('/home/travis/build/mskilab/fishHook/data/hypotheses.rds', out.path = '/home/travis/build/mskilab/fishHook/data/output.RDS')), 19688)
## if (!all(cov.types %in% COV.TYPES) & !(all(cov.classes %in% COV.CLASSES))){
expect_error(annotate.hypotheses(hypotheses, covariates = grl2))
## if events != NULL
expect_equal(max(annotate.hypotheses(hypotheses, events=events)$count), 9750)
##
## annotate.hypotheses(readRDS(gzcon(file("http://mskilab.com/fishHook/tests/targets.rds"))))
annotate.hypotheses(hypotheses)
expect_true(is(annotate.hypotheses(hypotheses), 'GRanges'))
##Interval Covariates
int = Covariate$new(data = replication_timing[1], name = 'int', type = 'interval', pad = 10)
fish3 = FishHook$new(hypotheses = hypotheses, events = events[1], covariates = int)
fish3$score()
##
## else if (grepl('(\\.bed$)', hypotheses[1])){
expect_equal(length(annotate.hypotheses(hypothesesbed)), 19688)
})
## aggregate.hypotheses
test_that('aggregate.hypotheses', {
## expect_error(aggregate.hypotheses(readRDS(gzcon(file("http://mskilab.com/fishHook/tests/targets.rds"))), rolling = 1))
expect_error(aggregate.hypotheses(hypotheses, rolling = 1))
## tmp.anno = c(readRDS(gzcon(file("http://mskilab.com/fishHook/tests/anno.rds"))) , readRDS(gzcon(file("http://mskilab.com/fishHook/tests/anno.rds"))))
tmp.anno = c(anno , anno)
tmp.anno$eligible = tmp.anno$coverage
agg = aggregate.hypotheses(tmp.anno, rolling = 20, na.rm = T)
expect_error({agg = aggregate.hypotheses(anno, rolling = -1)})
expect_error(aggregate.hypotheses(hypotheses)) ## Error: argument "by" must be specified and same length as hypotheses or "rolling" must be non NULL
foo = aggregate.hypotheses(hypotheses, by='gene_name')
expect_equal(length(foo$gene_name), 16352)
## if (is.null(by) & is.character(hypotheses)){
## expect_error(aggregate.hypotheses('/home/travis/build/mskilab/fishHook/data/hypotheses.rds'))
## annotate hypotheses
annotated = annotate.hypotheses(hypotheses, events=events)
## by
expect_equal(length(aggregate.hypotheses(annotated, by = 'gene_name')[[1]]), 16352)
## fields
expect_equal(length(aggregate.hypotheses(annotated, fields = 'count', rolling = 200)), 15032)
## rolling
#### Error: Malformed input, only NA ranges produced. Reduce value of running
expect_error(aggregate.hypotheses(annotated, rolling = 5e5))
#### Column 1 of result for group 2 is type 'logical' but expecting type 'double'. Column types must be consistent for each group.
expect_error(aggregate.hypotheses(annotated, rolling = 1e3))
#### works
expect_equal(length(aggregate.hypotheses(annotated, rolling = 200)), 15032)
## disjoint
expect_equal(length(aggregate.hypotheses(annotated, rolling = 200, disjoint = FALSE)), 15032)
## na.rm
expect_equal(length(aggregate.hypotheses(annotated, rolling = 200, na.rm = TRUE)), 15032)
## FUN
## verbose
expect_equal(length(aggregate.hypotheses(annotated, by = 'gene_name', verbose = FALSE)[[1]]), 16352)
##
## if (is.null(by) & is.character(hypotheses)){
expect_error(aggregate.hypotheses(hypotheses.2)) ## Coverage missing for input hypotheses
## if (is.null(by) & is.character(hypotheses)){ (continued)
expect_error(aggregate.hypotheses(annotated_cov))
##testing rolling aggregation
start = c(1,1001,2001,3001,4001,5001)
end = c(1000,2000,3000,4000,5000,6000)
chr = 1
test = dt2gr(as.data.table(cbind(start,end,chr)))
mcols(test) = NULL
fishAgg = FishHook$new(hypotheses = test, events = test)
fishAgg$aggregate(rolling = 3)
expect_error(fishAgg$aggregate(rolling = 0))
expect_error(fishAgg$aggregate(rolling = 4.5))
expect_equal(length(fishAgg$aggregated), 4)
expect_equal(width(fishAgg$aggregated), rep(3000,4))
})
## score.hypotheses
test_that('score.hypotheses', {
annotated = annotate.hypotheses(hypotheses, events=events)
## hypotheses
## covariates = names(values(hypotheses))
## model = NULL
## return.model = FALSE
expect_equal(as.integer(round(score.hypotheses(annotated, return.model = TRUE)$coefficients)), as.integer(-5))
## nb = TRUE
expect_equal(round(max(score.hypotheses(annotated, nb=TRUE)$effectsize)), 5)
expect_equal(round(max(score.hypotheses(annotated, nb=FALSE)$effectsize)), 7)
## verbose = TRUE
expect_equal(dim(as.data.table(score.hypotheses(annotated, verbose=FALSE)))[1], 19688)
expect_equal(dim(as.data.table(score.hypotheses(annotated, verbose=FALSE)))[2], 16)
## iter = 200
expect_equal(dim(as.data.table(score.hypotheses(annotated, iter=4000)))[1], 19688)
expect_equal(dim(as.data.table(score.hypotheses(annotated, iter=4000)))[2], 16)
## subsample = 1e5
expect_equal(dim(as.data.table(score.hypotheses(annotated, subsample = 5)))[1], 19688)
expect_equal(dim(as.data.table(score.hypotheses(annotated, subsample = 5)))[2], 16)
## seed = 42
## p.randomized = TRUE
expect_equal(dim(as.data.table(score.hypotheses(annotated, p.randomized = FALSE)))[1], 19688)
## classReturn = FALSE
expect_equal(dim(as.data.table(score.hypotheses(annotated, classReturn = TRUE)$res))[1], 19688)
## MISC testing
anno = annotated
expect_error({s1 = score.hypotheses(anno, covariates = c('hello'))})
anno$cov2 = c(1:length(anno))
anno$cov2 = as.character(anno$cov2)
expect_error({s2 = score.hypotheses(anno)})
anno3 = anno
anno3$count = NULL
anno3$cov2 = NULL
expect_error({s3 = score.hypotheses(anno3)})
anno3$count = 1
expect_error({s4 = score.hypotheses(anno3)})
anno3$dummy = NA
anno3$dummy = as.numeric(anno3$dummy)
anno3$count = c(1:length(anno3))
expect_error({s5 = score.hypotheses(anno3)})
anno$cov2 = NULL
anno$count = round(runif(length(anno))*5)
s6 = score.hypotheses(anno, subsample = 0.7)
anno$cov = NULL
anno$p = NULL
m1 = score.hypotheses(anno, return.model = TRUE)
s8 = score.hypotheses(anno[1:1000], model = m1)
anno$fac = c(1:2)
anno$fac = as.factor(anno$fac)
anno$cov = as.numeric(c(1:length(anno)))
anno4 = anno[1:1000]
s9 = score.hypotheses(anno4, verbose = TRUE, covariates = c('cov', 'fac'))
})
## Covariate
test_that('Covariate', {
## initialize = function(..., name = NULL, data = NULL, pad = NULL, type = NULL, signature = NULL, field = NULL, na.rm = NULL, grep = NULL)
## default
foobar = Covariate$new(data = replication_timing, type='numeric')
foobar2 = Covariate$new(data=replication_timing[50:100], type='numeric')
foobar3 = Covariate$new(data=list(replication_timing,replication_timing), type = c('numeric','numeric'), name = c('1','1'), na.rm = F)
## merge
#### check it runs
##expect_error(foobar$merge(foobar2), NA)
## chr
## expect_false(foobar$chr())
## seqlevels
expect_equal(length(foobar$seqlevels()[[1]]), 25)
## subset
foobar3$subset(1)
expect_equal(length(foobar3$data), 1)
## toList
##expect_equal(length(foobar$toList()[[1]]$track), 2385966)
##expect_equal(foobar$toList()[[1]]$type, 'numeric')
##expect_equal(foobar$toList()[[1]]$signature, NA)
##expect_equal(foobar$toList()[[1]]$pad, NA)
##expect_equal(foobar$toList()[[1]]$na.rm, NA)
##expect_equal(foobar$toList()[[1]]$field, NA)
##expect_equal(foobar$toList()[[1]]$grep, NA)
## print
##expect_equal(foobar$print()[[1]], NULL)
foobar4 = Covariate$new(data = list(replication_timing, replication_timing), type = c('numeric','numeric'))
expect_equal(any(foobar4$chr()), FALSE)
expect_equal(length(foobar4$toList()), 2)
expect_equal(foobar4$toList()[[2]]$type, 'numeric')
expect_output(foobar4$print())
foobar5 = c(foobar4,foobar4)
expect_equal(length(foobar5$data), 4)
foobar6 = foobar5[c(1,3)]
expect_equal(length(foobar6$data), 2)
##Testing print empty cov_arr
empty = Covariate$new()
##Testing active binding errors in cov_arr
##Names
expect_error({foobar3$names = 1})
expect_error({foobar3$names = c('1','2')})
##Type
foobar3 = c(foobar3, foobar3, foobar3, foobar3)
expect_error({foobar3$type = 1})
foobar3$type = c('sequence')
expect_error({foobar3$type = 'hello'})
foobar3$type = 'interval'
expect_error({foobar3$type = rep('interval', 10)})
##pad
expect_error({foobar3$pad = '1'})
foobar3$pad = 1
expect_equal(foobar3$pad[3], 1)
expect_error({foobar3$pad = c(1,1,1,1,1,1,1,1,1,1,1,1)})
##na.rm
expect_error({foobar3$na.rm = '1'})
expect_error({foobar3$na.rm = c('1','2')})
expect_error({foobar3$na.rm = c(T,T,T,T,T,T,T,T,T,T,T,T)})
##grep
expect_error({foobar3$grep = 1})
expect_error({foobar3$grep = c('1','2','1','2','3','4','5','6')})
##Sequence
#expect_error({foobar3$signature = 1})
##Field
expect_error({foobar3$field = 1})
foobar3$field = '1'
expect_error({foobar3$field = c('1','1','1','1','1','1','1','1','1')})
expect_equal(foobar3$field[1], '1')
##Covariate concatentation error on non cov_arr inputs
expect_error({foobar3$signature = c(1,2,3,4,5,6,7,8)})
##covariate merging
foobar7 = foobar5$merge(foobar5)
expect_equal(length(foobar7$data), 8)
##chr function
foobar8 = Covariate$new()
expect_equal(foobar8$chr(), NULL)
foobar9 = expect_error(Covariate$new(data = list(replication_timing, 'hello'), name = c('r1','r2')))
##seqlevels function
expect_equal(foobar8$seqlevels(), NULL)
##empty output print
expect_equal(foobar8$print(), NULL)
##c with item that is not Covariate
expect_error({foobar10 = c(foobar2, 1)})
})
## FishHook
## initialize = function(hypotheses = NULL, out.path = NULL, eligible = NULL, ... ,events = NULL, covariates = NULL,
## use_local_mut_density = FALSE, local_mut_density_bin = 1e6, genome = 'BSgenome.Hsapiens.UCSC.hg19::Hsapiens',
## mc.cores = 1, na.rm = TRUE, pad = 0, verbose = TRUE, max.slice = 1e3, ff.chunk = 1e6, max.chunk = 1e11, ptidcol = NULL,
## maxpatientpergene = Inf, weightEvents = FALSE, nb = TRUE)
test_that('FishHook', {
##FishHook initialization warnings:
##Mismatched seqlevels
set.seed(42)
events = sample(events, 1e5)
r1 = replication_timing[1]
r2 = replication_timing[1]
r2 = gr2dt(r2)
r2$seqnames = 'chr1'
r2 = dt2gr(r2)
c1 = Covariate$new(data = r1, type = 'interval', name = 'c1')
c2 = Covariate$new(data = r2, type = 'interval', name = 'c2')
t1 = hypotheses[1]
t2 = t1
t2 = gr2dt(t2)
t2$seqnames = 'chr1'
t2 = dt2gr(t2)
expect_warning({fish = FishHook$new(hypotheses = t1, events = t1, covariates = c(c1,c2))})
expect_warning({fish = FishHook$new(hypotheses = t1, events = t1, covariates = c(c2))})
#expect_error({fish = FishHook$new(hypotheses = t1, events = t1, eligible = t2)})
#expect_error({fish = FishHook$new(hypotheses = t1, events = t2)})
##Improper Cov class
expect_error({fish = FishHook$new(hypotheses = t1, events = t1, covariates = t1)})
t1 = hypotheses[1]
expect_error({fish = FishHook$new(hypotheses = 'hello', events = t1)})
expect_error({fish = FishHook$new(hypotheses = t1, events = 'hello')})
expect_error({fish = FishHook$new(hypotheses = t1, events = t1, eligible = 'hello')})
##local mut_density with covariates present
r1 = replication_timing[1]
c1 = Covariate$new(data = r1, type = 'interval', name = 'c1')
fish = FishHook$new(hypotheses = hypotheses, events = events, eligible = eligible, covariates = c(c1), use_local_mut_density = T)
##Trying to score after we have already scored
fish = FishHook$new(hypotheses = hypotheses, events = events, eligible = eligible)
fish$score()
fish$aggregate(rolling = 4)
expect_error({fish$aggregated = fish$anno})
fish$score()
##Printing functions, all regions eligible and no covs
t1 = hypotheses[1]
fish = FishHook$new(hypotheses = t1, events = t1)
expect_output(fish$print())
##QQ_plot
fish = FishHook$new(hypotheses = hypotheses, events = events, eligible = eligible)
fish$score()
p = fish$qqp(columns = c('gene_name', 'strand'))
expect_equal(grepl('gene_name', p$x[[3]][[1]]$text[1]) && grepl('strand', p$x[[3]][[1]]$text[1]), T)
## default
fish1 = FishHook$new(hypotheses = hypotheses, events = events, eligible =eligible, use_local_mut_density = T)
expect_equal(length(fish1$toList(fish1$covariates)), 1)
expect_output(fish1$print())
expect_true(fish1$nb)
expect_equal(fish1$state, 'Annotated')
expect_false(fish1$weightEvents)
anno = fish1$state
print('created anno')
## toList
## > fish1$toList()
## Error in fish1$toList() : argument "x" is missing, with no default
## print
fish1$print()
## annotate
expect_equal(anno, "Annotated")
## aggregate
## > agg = fish1$aggregate(mc.cores=2)
## Error in fish1$aggregate(mc.cores = 2) : unused argument (mc.cores = 2)
expect_error(fish1$aggregate())
agg = fish1$aggregate(by='gene_name')
## score
##
## Error in score.hypotheses(targ, covariates = names(values(private$panno)), :
## Error: "score.hypotheses" input malformed --> count does not vary!
## qq_plot
expect_error(fish1$qqp())
## clear
## with eligible
## foobar = FishHook$new(hypotheses = hypotheses, events = events, eligible = eligible)
fish2 = FishHook$new(hypotheses = hypotheses, events = events,
covariates = Covariate$new(data = replication_timing,
name = 'rept', type = 'numeric',
field = NA, pad = 10))
fish3 = fish2$clone()
replication_timing$bla = runif(length(replication_timing))
fish3$merge(Covariate$new(data = replication_timing[1:1000], field = 'bla', name = 'c.new'))
fish3$merge(fish2)
## test set analysis
expect_equal(dim(fish3)[2],3)
set.seed(42)
fish2$sets = split(sample(length(fish2), 100), 1:20)
expect_equal(nrow(values(fish2$setres)), 20)
res = fishHook::score(fish2[1:200,], fish2[1:200, ])
## test active bindings
## data
expect_equal(length(fish2$covariates), 1)
## eligible
expect_equal(fish2$eligible, NULL)
anno2 = fish2$state
print('created anno2')
## anno
expect_equal(max(fish2$data$count), 466)
## hypotheses
expect_equal(length(fish2$hypotheses), 19688)
##Scoring
fish2$score()
#expect_error(fish2$score())
expect_equal(ncol(as.data.table(fish2$res)), 17)
##Clearing
fish2$clear('Annotated')
expect_equal(fish2$state, 'Annotated')
expect_error(fish2$res)
fish2$score()
plot = fish2$qqp()
expect_equal(length(plot[[1]]), 7)
##testing rolling aggregation
start = c(1,1001,2001,3001,4001,5001)
end = c(1000,2000,3000,4000,5000,6000)
chr = 1
test = dt2gr(as.data.table(cbind(start,end,chr)))
mcols(test) = NULL
fishAgg = FishHook$new(hypotheses = test, events = test)
fishAgg$aggregate(rolling = 3)
fishAgg$clear('')
expect_error(fishAgg$aggregate(rolling = 0))
expect_error(fishAgg$aggregate(rolling = 4.5))
expect_equal(length(fishAgg$aggregated), 4)
expect_equal(any(!width(fishAgg$aggregated) == rep(3000,4)), FALSE)
##Interval Covariates
int = Covariate$new(data = replication_timing[1], name = 'int', type = 'interval', pad = 10)
fish3 = FishHook$new(hypotheses = hypotheses, events = events[1], covariates = int)
fish3$score()
##Mismatched warning
mis_eve = gr2dt(events[1])
mis_eve$seqnames = 'f'
mis_eve = dt2gr(mis_eve)
#expect_error({fish4 = FishHook$new(hypotheses = hypotheses, events = mis_eve, eligible = hypotheses[1])})
#expect_error({fish4 = FishHook$new(hypotheses = hypotheses, events = events[1],
# eligible = mis_eve,
# covariates = Covariate$new(csv = c(mis_eve,events[1]) ,
# type = c('interval'), name = c('mis','eve')))})
#expect_error({fish5 = FishHook$new(hypotheses = hypotheses, events = events[1], eligible = mis_eve)})
##Testing FishHook subsetting
fish2$eligible = eligible[1:100]
expect_equal(fish2$state, 'Annotated')
fish2$covariates = c(int,int,int,int)
expect_equal(fish2$state, 'Annotated')
fish5 = fish2
fish6 = fish5[c(1:10000), c(1,2)]
expect_equal(length(fish6$hypotheses), 10000)
expect_equal(length(fish6$covariates$data), 2)
##Testing fishHook active Bindings
#data
expect_error({fish2$data = 'Hello World'})
#eligible
expect_error({fish2$eligible = 'Hello World'})
#hypotheses
expect_error({fish2$hypotheses = 42})
##Need to fix issue with loading bed files
#fish2$hypotheses = '~/git/fishHook/data/hypotheses.bed'
##Events
expect_error({fish2$events = 'Hello World'})
expect_equal(length(fish2$events), 1e5)
##Out path
expect_error({fish2$out.path = 12})
# fish2$out.path = '/home/travis/build/mskilab/fishHook/data/out'
# expect_error({fish2$out.path = '~/git/fishHook/data/out'})
# expect_equal(fish2$out.path, '/home/travis/build/mskilab/fishHook/data/out')
##anno
expect_error({fish2$anno = 'hello world'})
expect_warning({fish2$data = fish2$data[, 1:4]})
##scores
expect_error({fish2$scores = 'Hello World'})
x = as.data.table('a')
##Model
expect_equal(fish2$model, NULL)
expect_error({fish2$model = 'hello'})
##mc.cores
expect_error({fish2$mc.cores = 'hello'})
fish2$mc.cores = 2
expect_equal(fish2$mc.cores, 2)
##na.rm
expect_error({fish2$na.rm = 'hello'})
fish2$na.rm = T
expect_equal(fish2$na.rm, T)
##pad
expect_error({fish2$pad = 'hello'})
fish2$pad = 2
expect_equal(fish2$pad, 2)
##verbose
expect_error({fish2$verbose = 'hello'})
fish2$verbose = T
expect_equal(fish2$verbose, T)
##max.slice
expect_error({fish2$max.slice = 'hello'})
fish2$max.slice = 2
expect_equal(fish2$max.slice, 2)
##ff.chunk
expect_error({fish2$ff.chunk = 'hello'})
fish2$ff.chunk = 2
expect_equal(fish2$ff.chunk, 2)
##max.chunk
expect_error({fish2$max.chunk = 'hello'})
fish2$max.chunk = 2
expect_equal(fish2$max.chunk, 2)
##ptidcol
expect_error({fish2$idcol = 1})
expect_error({fish2$idcol = "hello"})
fish2$max.chunk = 1e11
fish2$max.slice = 100000
fish2$idcol = 'id'
expect_equal(fish2$idcol, "id")
##maxpatientpergene
expect_error({fish2$idcap = 'hello'})
fish2$idcap = 2
expect_equal(fish2$idcap, 2)
##weightEvents
expect_error({fish2$weightEvents = 'hello'})
fish2$ff.chunk = 1e7
fish2 = fish2[1:1000, ]
fish2$weightEvents = T
expect_equal(fish2$weightEvents, T)
##nb
expect_error({fish2$nb = 'hello'})
fish2$nb = T
expect_equal(fish2$nb, T)
##all
expect_error({fish2$all = 'hello'})
##state
expect_error({fish2$state = 'hello'})
##agg
expect_error({fish2$aggregated = 'hello'})
grl = GRangesList(events[1], events[1], events[1])
expect_error({fish2$aggregated = grl})
})
### Annotate
## anno= Annotated$new(hypotheses=hypotheses, events=events, covariates=replication_timing)
## SLOW
test_that('qqp', {
pvals = c(0.0001, 0.0001, 0.032, 0.005, 0.9, 0.15)
## check default params
pdf('test.pdf')
foo = qqp(pvals, plotly = FALSE)
dev.off()
system('rm test.pdf')
## expect_match(names(foo)[1], 'rect')
## expect_match(names(foo)[2], 'text')
## expect_equal(round(foo$rect$w, 2), round(0.8597488, 2)) ## rounding; otherwise, 'double' types not exactly the same
## expect_equal(round(foo$rect$h, 2), round(0.4891113, 2))
## expect_equal(round(foo$rect$left, 2), round(3.820251, 2))
## expect_equal(round(foo$rect$top, 2), round(0.3091113, 2))
## expect_equal(round(foo$text$x, 2), round(4.073376, 2))
## expect_equal(round(foo$text$y, 2), round(0.008372093, 2))
## exp: Error: length of exp must be = length(obs
expect_error(qqp(pvals, exp=(c(1, 2, 3))))
## not sure how to test 'lwd'
## or these other args
foobar = qqp(pvals, plotly=TRUE)
expect_match(names(foobar)[1], 'x')
expect_match(names(foobar)[2], 'width')
expect_match(names(foobar)[3], 'height')
dist = rnorm(1000, mean = 10)/1000
p1 = qqp(dist)
p2 = qqp(dist, exp = rnorm(1000, mean = 8)/1000)
p3 = qqp(dist, exp = rnorm(1000, mean = 8)/1000, highlight = rep(c(T,F),500))
p4 = qqp(dist, exp = rnorm(1000, mean = 8)/1000, highlight = rep(c(T,F),500), max.x = 10, max.y = 10)
p4 = qqp(dist, exp = rnorm(1000, mean = 8)/1000, highlight = rep(c(T,F),500), max.x = 10, max.y = 10, subsample = 100)
p5 = qqp(plotly = T, obs = dist, exp = rnorm(1000, mean = 8)/1000, highlight = rep(c(T,F),500))
p6 = qqp(plotly = T, obs = dist, exp = rnorm(1000, mean = 8)/1000, highlight = rep(c(T,F),500), gradient = list('arb' = c(1:1000)))
p7 = qqp(plotly = T, obs = rnorm(10000,mean = 10)/1000, exp = rnorm(10000, mean = 8)/10000, highlight = rep(c(T,F),5000), gradient = list('arb' = c(1:10000)))
p8 = qqp(plotly = T, obs = rnorm(10000,mean = 10)/1000, exp = rnorm(10000, mean = 8)/10000, highlight = rep(c(T,F),5000))
p9 = qqp(plotly = T, obs = rnorm(10000,mean = 10)/1000, exp = rnorm(10000, mean = 8)/10000, highlight = 'hello')
expect_error({p10 = qqp(plotly = T, obs = rnorm(10000,mean = 10)/1000, exp = rnorm(10000, mean = 8)/10000, highlight = rep(c(T,F), 1000000))})
p11 = qqp(obs = rnorm(10000,mean = 10)/1000, exp = rnorm(10000, mean = 8)/10000, label = c(1:1000))
p12 = qqp(obs = rnorm(10020,mean = 10)/1000, exp = rnorm(10020, mean = 8)/10020)
expect_error({p13 = qqp(obs = rnorm(10000,mean = 10)/1000, exp = rnorm(10000, mean = 8)/10000, highlight = rep(c(T,F), 100000000))})
p14 = qqp(obs = rnorm(10000,mean = 10)/1000, exp = rnorm(10000, mean = 8)/10000, highlight = c(1:5000))
})
mskilab/fish.hook documentation built on Jan. 12, 2023, 8:35 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
library(fishHook)
library(testthat)
library(zoo)
## Sys.setenv(DEFAULT_BSGENOME = 'http://mskilab.com/gUtils/human_g1k_v37.chrom.sizes')
Sys.setenv(DEFAULT_BSGENOME = system.file('extdata', 'human_g1k_v37.chrom.sizes', package = 'fishHook'))
## hypotheses BED
## hypothesesbed = "http://mskilab.com/fishHook/tests/targets.bed"
## hypothesesbed = system.file("extdata", "targets.bed", package = "fishHook")
hypothesesbed = testthat::test_path("data", "targets.bed")
## Sample Covariate #replication_timing = readRDS(gzcon(file("http://mskilab.com/fishHook/tests/covariate.rds")))
## replication_timing = readRDS(gzcon(file("http://mskilab.com/fishHook/tests/covariate.rds")))
replication_timing = readRDS(testthat::test_path("data", "covariate.rds"))
## Same Eligible Subset
## eligible = readRDS(gzcon(file("http://mskilab.com/fishHook/tests/eligible.rds")))
eligible = readRDS(testthat::test_path("data", "eligible.rds"))
## Sample Hypotheses
#hypotheses = readRDS(gzcon(file("http://mskilab.com/fishHook/tests/targets.rds")))
hypotheses = readRDS(testthat::test_path("data", "targets.rds"))
hypotheses.2 = readRDS(testthat::test_path("data", "hypotheses.rds"))
## indexed pathways
#indexed_pathways = readRDS(gzcon(file("http://mskilab.com/fishHook/tests/indexed_pathways.rds")))
## indexed_pathways = readRDS(gzcon(file("http://mskilab.com/fishHook/tests/indexed_pathways.rds")))
indexed_pathways = readRDS(testthat::test_path("data", "indexed_pathways.rds"))
annotated_cov = readRDS(testthat::test_path("data", "annotated_cov.rds"))
## segs
## segs = readRDS(gzcon(file("http://mskilab.com/fishHook/tests/jabba_segs_11517.rds")))
segs = readRDS(testthat::test_path("data", "jabba_segs_11517.rds"))
#segs = readRDS(gzcon(file("http://mskilab.com/fishHook/tests/jabba_segs_11517.rds")))
## eligible
## eligible = readRDS(gzcon(file("http://mskilab.com/fishHook/tests/eligible.rds")))
# Sample annotate
#anno = readRDS(gzcon(file("http://mskilab.com/fishHook/tests/anno.rds")))
## anno = readRDS(gzcon(file( "http://mskilab.com/fishHook/tests/anno.rds")))
anno = readRDS(testthat::test_path("data", "anno.rds"))
context('unit testing fishhook operations')
test_that('annotate.hypotheses', {
## default args
t1 = hypotheses[1]
e1 = events[1]
el1 = eligible[1]
r1 = replication_timing[1]
expect_error({annotate.hypotheses(hypotheses = hypotheses[1], events = events[1], idcol = 'hello')})
## c1 = Covariate$new(data = readRDS(gzcon(file("http://mskilab.com/fishHook/tests/covariate.rds"))), type = 'numeric', name = 'c1', pad = 10)
c1 = Covariate$new(data = replication_timing, type = 'numeric', name = 'c1', pad = 10)
c2 = Covariate$new(data = replication_timing, type = 'interval', name = 'c2', pad = 10, na.rm = TRUE)
## c2 = Covariate$new(data = readRDS(gzcon(file("http://mskilab.com/fishHook/tests/covariate.rds"))), type = 'interval', name = 'c2', pad = 10, na.rm = TRUE)
c3 = Covariate$new(data = list(RleList()), type = 'interval', name = 'c3', pad = 10, na.rm = TRUE)
anno = annotate.hypotheses(hypotheses = t1, events = e1, covariates = c1$toList())
anno2 = annotate.hypotheses(hypotheses = t1, events = e1, covariates = c2$toList())
#anno3 = annotate.hypotheses(hypotheses = t1, events = e1, covariates = c3$toList())
anno4 = annotate.hypotheses(hypotheses = e1, events = e1, eligible = e1)
expect_true(is(annotate.hypotheses(hypotheses), 'GRanges')) ## annotate.hypotheses(hypotheses, weightEvents=TRUE)
expect_equal(length(annotate.hypotheses(hypotheses)), 19688)
foo = annotate.hypotheses(hypotheses)
expect_equal(max(foo$query.id), 19688)
expect_equal(max(foo$eligible), 2304638)
expect_equal(max(foo$count), 0)
## covered
## events
## mc.cores
expect_true(is(annotate.hypotheses(hypotheses, mc.cores=2), 'GRanges'))
expect_equal(length(annotate.hypotheses(hypotheses, mc.cores=2)), 19688)
expect_equal(max(annotate.hypotheses(hypotheses, mc.cores=2)$query.id), 19688)
expect_equal(max(annotate.hypotheses(hypotheses, mc.cores=2)$eligible), 2304638)
expect_equal(max(annotate.hypotheses(hypotheses, mc.cores=2)$count), 0)
## na.rm
## pad
## verbose
expect_true(is(annotate.hypotheses(hypotheses, verbose=FALSE), 'GRanges'))
expect_equal(length(annotate.hypotheses(hypotheses, verbose=FALSE)), 19688)
expect_equal(max(annotate.hypotheses(hypotheses, verbose=FALSE)$query.id), 19688)
expect_equal(max(annotate.hypotheses(hypotheses, verbose=FALSE)$eligible), 2304638)
expect_equal(max(annotate.hypotheses(hypotheses, verbose=FALSE)$count), 0)
## max.slice
## annotate.hypotheses(hypotheses, max.slice=1e8)
## ff.chunk
##
## max.chunk
## out.path
## covariates
## maxpatientpergene
## ptidcol
## weightEvents
expect_equal(is.na(all(annotate.hypotheses(hypotheses, weightEvents=TRUE)$count)), TRUE)
## if (is.character(hypotheses)){
## expect_equal(length(annotate.hypotheses('/home/travis/build/mskilab/fishHook/data/hypotheses.rds')), 19688)
## if (length(hypotheses)==0){
expect_error(annotate.hypotheses(GRanges()))
## if (!is.null(out.path)){
## expect_equal(length(annotate.hypotheses('/home/travis/build/mskilab/fishHook/data/hypotheses.rds', out.path = '/home/travis/build/mskilab/fishHook/data/output.RDS')), 19688)
## if (!all(cov.types %in% COV.TYPES) & !(all(cov.classes %in% COV.CLASSES))){
expect_error(annotate.hypotheses(hypotheses, covariates = grl2))
## if events != NULL
expect_equal(max(annotate.hypotheses(hypotheses, events=events)$count), 9750)
##
## annotate.hypotheses(readRDS(gzcon(file("http://mskilab.com/fishHook/tests/targets.rds"))))
annotate.hypotheses(hypotheses)
expect_true(is(annotate.hypotheses(hypotheses), 'GRanges'))
##Interval Covariates
int = Covariate$new(data = replication_timing[1], name = 'int', type = 'interval', pad = 10)
fish3 = FishHook$new(hypotheses = hypotheses, events = events[1], covariates = int)
fish3$score()
##
## else if (grepl('(\\.bed$)', hypotheses[1])){
expect_equal(length(annotate.hypotheses(hypothesesbed)), 19688)
})
## aggregate.hypotheses
test_that('aggregate.hypotheses', {
## expect_error(aggregate.hypotheses(readRDS(gzcon(file("http://mskilab.com/fishHook/tests/targets.rds"))), rolling = 1))
expect_error(aggregate.hypotheses(hypotheses, rolling = 1))
## tmp.anno = c(readRDS(gzcon(file("http://mskilab.com/fishHook/tests/anno.rds"))) , readRDS(gzcon(file("http://mskilab.com/fishHook/tests/anno.rds"))))
tmp.anno = c(anno , anno)
tmp.anno$eligible = tmp.anno$coverage
agg = aggregate.hypotheses(tmp.anno, rolling = 20, na.rm = T)
expect_error({agg = aggregate.hypotheses(anno, rolling = -1)})
expect_error(aggregate.hypotheses(hypotheses)) ## Error: argument "by" must be specified and same length as hypotheses or "rolling" must be non NULL
foo = aggregate.hypotheses(hypotheses, by='gene_name')
expect_equal(length(foo$gene_name), 16352)
## if (is.null(by) & is.character(hypotheses)){
## expect_error(aggregate.hypotheses('/home/travis/build/mskilab/fishHook/data/hypotheses.rds'))
## annotate hypotheses
annotated = annotate.hypotheses(hypotheses, events=events)
## by
expect_equal(length(aggregate.hypotheses(annotated, by = 'gene_name')[[1]]), 16352)
## fields
expect_equal(length(aggregate.hypotheses(annotated, fields = 'count', rolling = 200)), 15032)
## rolling
#### Error: Malformed input, only NA ranges produced. Reduce value of running
expect_error(aggregate.hypotheses(annotated, rolling = 5e5))
#### Column 1 of result for group 2 is type 'logical' but expecting type 'double'. Column types must be consistent for each group.
expect_error(aggregate.hypotheses(annotated, rolling = 1e3))
#### works
expect_equal(length(aggregate.hypotheses(annotated, rolling = 200)), 15032)
## disjoint
expect_equal(length(aggregate.hypotheses(annotated, rolling = 200, disjoint = FALSE)), 15032)
## na.rm
expect_equal(length(aggregate.hypotheses(annotated, rolling = 200, na.rm = TRUE)), 15032)
## FUN
## verbose
expect_equal(length(aggregate.hypotheses(annotated, by = 'gene_name', verbose = FALSE)[[1]]), 16352)
##
## if (is.null(by) & is.character(hypotheses)){
expect_error(aggregate.hypotheses(hypotheses.2)) ## Coverage missing for input hypotheses
## if (is.null(by) & is.character(hypotheses)){ (continued)
expect_error(aggregate.hypotheses(annotated_cov))
##testing rolling aggregation
start = c(1,1001,2001,3001,4001,5001)
end = c(1000,2000,3000,4000,5000,6000)
chr = 1
test = dt2gr(as.data.table(cbind(start,end,chr)))
mcols(test) = NULL
fishAgg = FishHook$new(hypotheses = test, events = test)
fishAgg$aggregate(rolling = 3)
expect_error(fishAgg$aggregate(rolling = 0))
expect_error(fishAgg$aggregate(rolling = 4.5))
expect_equal(length(fishAgg$aggregated), 4)
expect_equal(width(fishAgg$aggregated), rep(3000,4))
})
## score.hypotheses
test_that('score.hypotheses', {
annotated = annotate.hypotheses(hypotheses, events=events)
## hypotheses
## covariates = names(values(hypotheses))
## model = NULL
## return.model = FALSE
expect_equal(as.integer(round(score.hypotheses(annotated, return.model = TRUE)$coefficients)), as.integer(-5))
## nb = TRUE
expect_equal(round(max(score.hypotheses(annotated, nb=TRUE)$effectsize)), 5)
expect_equal(round(max(score.hypotheses(annotated, nb=FALSE)$effectsize)), 7)
## verbose = TRUE
expect_equal(dim(as.data.table(score.hypotheses(annotated, verbose=FALSE)))[1], 19688)
expect_equal(dim(as.data.table(score.hypotheses(annotated, verbose=FALSE)))[2], 16)
## iter = 200
expect_equal(dim(as.data.table(score.hypotheses(annotated, iter=4000)))[1], 19688)
expect_equal(dim(as.data.table(score.hypotheses(annotated, iter=4000)))[2], 16)
## subsample = 1e5
expect_equal(dim(as.data.table(score.hypotheses(annotated, subsample = 5)))[1], 19688)
expect_equal(dim(as.data.table(score.hypotheses(annotated, subsample = 5)))[2], 16)
## seed = 42
## p.randomized = TRUE
expect_equal(dim(as.data.table(score.hypotheses(annotated, p.randomized = FALSE)))[1], 19688)
## classReturn = FALSE
expect_equal(dim(as.data.table(score.hypotheses(annotated, classReturn = TRUE)$res))[1], 19688)
## MISC testing
anno = annotated
expect_error({s1 = score.hypotheses(anno, covariates = c('hello'))})
anno$cov2 = c(1:length(anno))
anno$cov2 = as.character(anno$cov2)
expect_error({s2 = score.hypotheses(anno)})
anno3 = anno
anno3$count = NULL
anno3$cov2 = NULL
expect_error({s3 = score.hypotheses(anno3)})
anno3$count = 1
expect_error({s4 = score.hypotheses(anno3)})
anno3$dummy = NA
anno3$dummy = as.numeric(anno3$dummy)
anno3$count = c(1:length(anno3))
expect_error({s5 = score.hypotheses(anno3)})
anno$cov2 = NULL
anno$count = round(runif(length(anno))*5)
s6 = score.hypotheses(anno, subsample = 0.7)
anno$cov = NULL
anno$p = NULL
m1 = score.hypotheses(anno, return.model = TRUE)
s8 = score.hypotheses(anno[1:1000], model = m1)
anno$fac = c(1:2)
anno$fac = as.factor(anno$fac)
anno$cov = as.numeric(c(1:length(anno)))
anno4 = anno[1:1000]
s9 = score.hypotheses(anno4, verbose = TRUE, covariates = c('cov', 'fac'))
})
## Covariate
test_that('Covariate', {
## initialize = function(..., name = NULL, data = NULL, pad = NULL, type = NULL, signature = NULL, field = NULL, na.rm = NULL, grep = NULL)
## default
foobar = Covariate$new(data = replication_timing, type='numeric')
foobar2 = Covariate$new(data=replication_timing[50:100], type='numeric')
foobar3 = Covariate$new(data=list(replication_timing,replication_timing), type = c('numeric','numeric'), name = c('1','1'), na.rm = F)
## merge
#### check it runs
##expect_error(foobar$merge(foobar2), NA)
## chr
## expect_false(foobar$chr())
## seqlevels
expect_equal(length(foobar$seqlevels()[[1]]), 25)
## subset
foobar3$subset(1)
expect_equal(length(foobar3$data), 1)
## toList
##expect_equal(length(foobar$toList()[[1]]$track), 2385966)
##expect_equal(foobar$toList()[[1]]$type, 'numeric')
##expect_equal(foobar$toList()[[1]]$signature, NA)
##expect_equal(foobar$toList()[[1]]$pad, NA)
##expect_equal(foobar$toList()[[1]]$na.rm, NA)
##expect_equal(foobar$toList()[[1]]$field, NA)
##expect_equal(foobar$toList()[[1]]$grep, NA)
## print
##expect_equal(foobar$print()[[1]], NULL)
foobar4 = Covariate$new(data = list(replication_timing, replication_timing), type = c('numeric','numeric'))
expect_equal(any(foobar4$chr()), FALSE)
expect_equal(length(foobar4$toList()), 2)
expect_equal(foobar4$toList()[[2]]$type, 'numeric')
expect_output(foobar4$print())
foobar5 = c(foobar4,foobar4)
expect_equal(length(foobar5$data), 4)
foobar6 = foobar5[c(1,3)]
expect_equal(length(foobar6$data), 2)
##Testing print empty cov_arr
empty = Covariate$new()
##Testing active binding errors in cov_arr
##Names
expect_error({foobar3$names = 1})
expect_error({foobar3$names = c('1','2')})
##Type
foobar3 = c(foobar3, foobar3, foobar3, foobar3)
expect_error({foobar3$type = 1})
foobar3$type = c('sequence')
expect_error({foobar3$type = 'hello'})
foobar3$type = 'interval'
expect_error({foobar3$type = rep('interval', 10)})
##pad
expect_error({foobar3$pad = '1'})
foobar3$pad = 1
expect_equal(foobar3$pad[3], 1)
expect_error({foobar3$pad = c(1,1,1,1,1,1,1,1,1,1,1,1)})
##na.rm
expect_error({foobar3$na.rm = '1'})
expect_error({foobar3$na.rm = c('1','2')})
expect_error({foobar3$na.rm = c(T,T,T,T,T,T,T,T,T,T,T,T)})
##grep
expect_error({foobar3$grep = 1})
expect_error({foobar3$grep = c('1','2','1','2','3','4','5','6')})
##Sequence
#expect_error({foobar3$signature = 1})
##Field
expect_error({foobar3$field = 1})
foobar3$field = '1'
expect_error({foobar3$field = c('1','1','1','1','1','1','1','1','1')})
expect_equal(foobar3$field[1], '1')
##Covariate concatentation error on non cov_arr inputs
expect_error({foobar3$signature = c(1,2,3,4,5,6,7,8)})
##covariate merging
foobar7 = foobar5$merge(foobar5)
expect_equal(length(foobar7$data), 8)
##chr function
foobar8 = Covariate$new()
expect_equal(foobar8$chr(), NULL)
foobar9 = expect_error(Covariate$new(data = list(replication_timing, 'hello'), name = c('r1','r2')))
##seqlevels function
expect_equal(foobar8$seqlevels(), NULL)
##empty output print
expect_equal(foobar8$print(), NULL)
##c with item that is not Covariate
expect_error({foobar10 = c(foobar2, 1)})
})
## FishHook
## initialize = function(hypotheses = NULL, out.path = NULL, eligible = NULL, ... ,events = NULL, covariates = NULL,
## use_local_mut_density = FALSE, local_mut_density_bin = 1e6, genome = 'BSgenome.Hsapiens.UCSC.hg19::Hsapiens',
## mc.cores = 1, na.rm = TRUE, pad = 0, verbose = TRUE, max.slice = 1e3, ff.chunk = 1e6, max.chunk = 1e11, ptidcol = NULL,
## maxpatientpergene = Inf, weightEvents = FALSE, nb = TRUE)
test_that('FishHook', {
##FishHook initialization warnings:
##Mismatched seqlevels
set.seed(42)
events = sample(events, 1e5)
r1 = replication_timing[1]
r2 = replication_timing[1]
r2 = gr2dt(r2)
r2$seqnames = 'chr1'
r2 = dt2gr(r2)
c1 = Covariate$new(data = r1, type = 'interval', name = 'c1')
c2 = Covariate$new(data = r2, type = 'interval', name = 'c2')
t1 = hypotheses[1]
t2 = t1
t2 = gr2dt(t2)
t2$seqnames = 'chr1'
t2 = dt2gr(t2)
expect_warning({fish = FishHook$new(hypotheses = t1, events = t1, covariates = c(c1,c2))})
expect_warning({fish = FishHook$new(hypotheses = t1, events = t1, covariates = c(c2))})
#expect_error({fish = FishHook$new(hypotheses = t1, events = t1, eligible = t2)})
#expect_error({fish = FishHook$new(hypotheses = t1, events = t2)})
##Improper Cov class
expect_error({fish = FishHook$new(hypotheses = t1, events = t1, covariates = t1)})
t1 = hypotheses[1]
expect_error({fish = FishHook$new(hypotheses = 'hello', events = t1)})
expect_error({fish = FishHook$new(hypotheses = t1, events = 'hello')})
expect_error({fish = FishHook$new(hypotheses = t1, events = t1, eligible = 'hello')})
##local mut_density with covariates present
r1 = replication_timing[1]
c1 = Covariate$new(data = r1, type = 'interval', name = 'c1')
fish = FishHook$new(hypotheses = hypotheses, events = events, eligible = eligible, covariates = c(c1), use_local_mut_density = T)
##Trying to score after we have already scored
fish = FishHook$new(hypotheses = hypotheses, events = events, eligible = eligible)
fish$score()
fish$aggregate(rolling = 4)
expect_error({fish$aggregated = fish$anno})
fish$score()
##Printing functions, all regions eligible and no covs
t1 = hypotheses[1]
fish = FishHook$new(hypotheses = t1, events = t1)
expect_output(fish$print())
##QQ_plot
fish = FishHook$new(hypotheses = hypotheses, events = events, eligible = eligible)
fish$score()
p = fish$qqp(columns = c('gene_name', 'strand'))
expect_equal(grepl('gene_name', p$x[[3]][[1]]$text[1]) && grepl('strand', p$x[[3]][[1]]$text[1]), T)
## default
fish1 = FishHook$new(hypotheses = hypotheses, events = events, eligible =eligible, use_local_mut_density = T)
expect_equal(length(fish1$toList(fish1$covariates)), 1)
expect_output(fish1$print())
expect_true(fish1$nb)
expect_equal(fish1$state, 'Annotated')
expect_false(fish1$weightEvents)
anno = fish1$state
print('created anno')
## toList
## > fish1$toList()
## Error in fish1$toList() : argument "x" is missing, with no default
## print
fish1$print()
## annotate
expect_equal(anno, "Annotated")
## aggregate
## > agg = fish1$aggregate(mc.cores=2)
## Error in fish1$aggregate(mc.cores = 2) : unused argument (mc.cores = 2)
expect_error(fish1$aggregate())
agg = fish1$aggregate(by='gene_name')
## score
##
## Error in score.hypotheses(targ, covariates = names(values(private$panno)), :
## Error: "score.hypotheses" input malformed --> count does not vary!
## qq_plot
expect_error(fish1$qqp())
## clear
## with eligible
## foobar = FishHook$new(hypotheses = hypotheses, events = events, eligible = eligible)
fish2 = FishHook$new(hypotheses = hypotheses, events = events,
covariates = Covariate$new(data = replication_timing,
name = 'rept', type = 'numeric',
field = NA, pad = 10))
fish3 = fish2$clone()
replication_timing$bla = runif(length(replication_timing))
fish3$merge(Covariate$new(data = replication_timing[1:1000], field = 'bla', name = 'c.new'))
fish3$merge(fish2)
## test set analysis
expect_equal(dim(fish3)[2],3)
set.seed(42)
fish2$sets = split(sample(length(fish2), 100), 1:20)
expect_equal(nrow(values(fish2$setres)), 20)
res = fishHook::score(fish2[1:200,], fish2[1:200, ])
## test active bindings
## data
expect_equal(length(fish2$covariates), 1)
## eligible
expect_equal(fish2$eligible, NULL)
anno2 = fish2$state
print('created anno2')
## anno
expect_equal(max(fish2$data$count), 466)
## hypotheses
expect_equal(length(fish2$hypotheses), 19688)
##Scoring
fish2$score()
#expect_error(fish2$score())
expect_equal(ncol(as.data.table(fish2$res)), 17)
##Clearing
fish2$clear('Annotated')
expect_equal(fish2$state, 'Annotated')
expect_error(fish2$res)
fish2$score()
plot = fish2$qqp()
expect_equal(length(plot[[1]]), 7)
##testing rolling aggregation
start = c(1,1001,2001,3001,4001,5001)
end = c(1000,2000,3000,4000,5000,6000)
chr = 1
test = dt2gr(as.data.table(cbind(start,end,chr)))
mcols(test) = NULL
fishAgg = FishHook$new(hypotheses = test, events = test)
fishAgg$aggregate(rolling = 3)
fishAgg$clear('')
expect_error(fishAgg$aggregate(rolling = 0))
expect_error(fishAgg$aggregate(rolling = 4.5))
expect_equal(length(fishAgg$aggregated), 4)
expect_equal(any(!width(fishAgg$aggregated) == rep(3000,4)), FALSE)
##Interval Covariates
int = Covariate$new(data = replication_timing[1], name = 'int', type = 'interval', pad = 10)
fish3 = FishHook$new(hypotheses = hypotheses, events = events[1], covariates = int)
fish3$score()
##Mismatched warning
mis_eve = gr2dt(events[1])
mis_eve$seqnames = 'f'
mis_eve = dt2gr(mis_eve)
#expect_error({fish4 = FishHook$new(hypotheses = hypotheses, events = mis_eve, eligible = hypotheses[1])})
#expect_error({fish4 = FishHook$new(hypotheses = hypotheses, events = events[1],
# eligible = mis_eve,
# covariates = Covariate$new(csv = c(mis_eve,events[1]) ,
# type = c('interval'), name = c('mis','eve')))})
#expect_error({fish5 = FishHook$new(hypotheses = hypotheses, events = events[1], eligible = mis_eve)})
##Testing FishHook subsetting
fish2$eligible = eligible[1:100]
expect_equal(fish2$state, 'Annotated')
fish2$covariates = c(int,int,int,int)
expect_equal(fish2$state, 'Annotated')
fish5 = fish2
fish6 = fish5[c(1:10000), c(1,2)]
expect_equal(length(fish6$hypotheses), 10000)
expect_equal(length(fish6$covariates$data), 2)
##Testing fishHook active Bindings
#data
expect_error({fish2$data = 'Hello World'})
#eligible
expect_error({fish2$eligible = 'Hello World'})
#hypotheses
expect_error({fish2$hypotheses = 42})
##Need to fix issue with loading bed files
#fish2$hypotheses = '~/git/fishHook/data/hypotheses.bed'
##Events
expect_error({fish2$events = 'Hello World'})
expect_equal(length(fish2$events), 1e5)
##Out path
expect_error({fish2$out.path = 12})
# fish2$out.path = '/home/travis/build/mskilab/fishHook/data/out'
# expect_error({fish2$out.path = '~/git/fishHook/data/out'})
# expect_equal(fish2$out.path, '/home/travis/build/mskilab/fishHook/data/out')
##anno
expect_error({fish2$anno = 'hello world'})
expect_warning({fish2$data = fish2$data[, 1:4]})
##scores
expect_error({fish2$scores = 'Hello World'})
x = as.data.table('a')
##Model
expect_equal(fish2$model, NULL)
expect_error({fish2$model = 'hello'})
##mc.cores
expect_error({fish2$mc.cores = 'hello'})
fish2$mc.cores = 2
expect_equal(fish2$mc.cores, 2)
##na.rm
expect_error({fish2$na.rm = 'hello'})
fish2$na.rm = T
expect_equal(fish2$na.rm, T)
##pad
expect_error({fish2$pad = 'hello'})
fish2$pad = 2
expect_equal(fish2$pad, 2)
##verbose
expect_error({fish2$verbose = 'hello'})
fish2$verbose = T
expect_equal(fish2$verbose, T)
##max.slice
expect_error({fish2$max.slice = 'hello'})
fish2$max.slice = 2
expect_equal(fish2$max.slice, 2)
##ff.chunk
expect_error({fish2$ff.chunk = 'hello'})
fish2$ff.chunk = 2
expect_equal(fish2$ff.chunk, 2)
##max.chunk
expect_error({fish2$max.chunk = 'hello'})
fish2$max.chunk = 2
expect_equal(fish2$max.chunk, 2)
##ptidcol
expect_error({fish2$idcol = 1})
expect_error({fish2$idcol = "hello"})
fish2$max.chunk = 1e11
fish2$max.slice = 100000
fish2$idcol = 'id'
expect_equal(fish2$idcol, "id")
##maxpatientpergene
expect_error({fish2$idcap = 'hello'})
fish2$idcap = 2
expect_equal(fish2$idcap, 2)
##weightEvents
expect_error({fish2$weightEvents = 'hello'})
fish2$ff.chunk = 1e7
fish2 = fish2[1:1000, ]
fish2$weightEvents = T
expect_equal(fish2$weightEvents, T)
##nb
expect_error({fish2$nb = 'hello'})
fish2$nb = T
expect_equal(fish2$nb, T)
##all
expect_error({fish2$all = 'hello'})
##state
expect_error({fish2$state = 'hello'})
##agg
expect_error({fish2$aggregated = 'hello'})
grl = GRangesList(events[1], events[1], events[1])
expect_error({fish2$aggregated = grl})
})
### Annotate
## anno= Annotated$new(hypotheses=hypotheses, events=events, covariates=replication_timing)
## SLOW
test_that('qqp', {
pvals = c(0.0001, 0.0001, 0.032, 0.005, 0.9, 0.15)
## check default params
pdf('test.pdf')
foo = qqp(pvals, plotly = FALSE)
dev.off()
system('rm test.pdf')
## expect_match(names(foo)[1], 'rect')
## expect_match(names(foo)[2], 'text')
## expect_equal(round(foo$rect$w, 2), round(0.8597488, 2)) ## rounding; otherwise, 'double' types not exactly the same
## expect_equal(round(foo$rect$h, 2), round(0.4891113, 2))
## expect_equal(round(foo$rect$left, 2), round(3.820251, 2))
## expect_equal(round(foo$rect$top, 2), round(0.3091113, 2))
## expect_equal(round(foo$text$x, 2), round(4.073376, 2))
## expect_equal(round(foo$text$y, 2), round(0.008372093, 2))
## exp: Error: length of exp must be = length(obs
expect_error(qqp(pvals, exp=(c(1, 2, 3))))
## not sure how to test 'lwd'
## or these other args
foobar = qqp(pvals, plotly=TRUE)
expect_match(names(foobar)[1], 'x')
expect_match(names(foobar)[2], 'width')
expect_match(names(foobar)[3], 'height')
dist = rnorm(1000, mean = 10)/1000
p1 = qqp(dist)
p2 = qqp(dist, exp = rnorm(1000, mean = 8)/1000)
p3 = qqp(dist, exp = rnorm(1000, mean = 8)/1000, highlight = rep(c(T,F),500))
p4 = qqp(dist, exp = rnorm(1000, mean = 8)/1000, highlight = rep(c(T,F),500), max.x = 10, max.y = 10)
p4 = qqp(dist, exp = rnorm(1000, mean = 8)/1000, highlight = rep(c(T,F),500), max.x = 10, max.y = 10, subsample = 100)
p5 = qqp(plotly = T, obs = dist, exp = rnorm(1000, mean = 8)/1000, highlight = rep(c(T,F),500))
p6 = qqp(plotly = T, obs = dist, exp = rnorm(1000, mean = 8)/1000, highlight = rep(c(T,F),500), gradient = list('arb' = c(1:1000)))
p7 = qqp(plotly = T, obs = rnorm(10000,mean = 10)/1000, exp = rnorm(10000, mean = 8)/10000, highlight = rep(c(T,F),5000), gradient = list('arb' = c(1:10000)))
p8 = qqp(plotly = T, obs = rnorm(10000,mean = 10)/1000, exp = rnorm(10000, mean = 8)/10000, highlight = rep(c(T,F),5000))
p9 = qqp(plotly = T, obs = rnorm(10000,mean = 10)/1000, exp = rnorm(10000, mean = 8)/10000, highlight = 'hello')
expect_error({p10 = qqp(plotly = T, obs = rnorm(10000,mean = 10)/1000, exp = rnorm(10000, mean = 8)/10000, highlight = rep(c(T,F), 1000000))})
p11 = qqp(obs = rnorm(10000,mean = 10)/1000, exp = rnorm(10000, mean = 8)/10000, label = c(1:1000))
p12 = qqp(obs = rnorm(10020,mean = 10)/1000, exp = rnorm(10020, mean = 8)/10020)
expect_error({p13 = qqp(obs = rnorm(10000,mean = 10)/1000, exp = rnorm(10000, mean = 8)/10000, highlight = rep(c(T,F), 100000000))})
p14 = qqp(obs = rnorm(10000,mean = 10)/1000, exp = rnorm(10000, mean = 8)/10000, highlight = c(1:5000))
})
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.