inittime <- Sys.time()
cat(paste("\n Starting FDF-small-fitness-specs", date(), "\n"))
## Testing fitness specs with missing genotypes and with letters too
## Also tests that are failed by version from October-December 2020 at least.
test_that("We can run and equal with letters", {
g1 <- data.frame(
Genotype = c("WT", "A", "A, B"),
Fitness = c("1",
"1.3 + 1.2 * n_1_2",
"1.1"),
stringsAsFactors = FALSE)
fg1 <- allFitnessEffects(genotFitness = g1,
frequencyDependentFitness = TRUE)
g11 <- data.frame(
Genotype = c("WT", "A", "A, B"),
Fitness = c("1",
"1.3 + 1.2 * n_A_B",
"1.1"),
stringsAsFactors = FALSE)
fg11 <- allFitnessEffects(genotFitness = g11,
frequencyDependentFitness = TRUE)
## FIXME
## So, what does spPopSizes refer to ?? that is ambiguous here.
## It seems it is the Genotypes in the original Genotype spec.
## but they are the ones left in fg1$Genotype
## Since we cannot now what was in g1, do not allow for this.
efg1 <- suppressWarnings(evalAllGenotypes(fg1, spPopSizes = c(9, 2, 6)))
efg11 <- suppressWarnings(evalAllGenotypes(fg11, spPopSizes = c(9, 2, 6)))
expect_identical(efg1, efg11)
## is that correct?
expect_identical(efg1$Fitness, c(1, 1.3 + 1.2 * 6, 0, 1.1))
g1b <- data.frame(
Genotype = c("WT", "A", "A, B", "B"),
Fitness = c("1",
"1.3 + 1.2 * n_1_2",
"1.1",
"0"),
stringsAsFactors = FALSE)
fg1b <- allFitnessEffects(genotFitness = g1b,
frequencyDependentFitness = TRUE)
## it thinks n_1_2 is 6. This is right
(ea1b <- suppressWarnings(evalAllGenotypes(fg1b,
spPopSizes = c(9, 2, 0, 6))))
## is it correct? Yes
stopifnot(identical(ea1b[, "Fitness"], c(1, 1.3 + 1.2 * 6, 0, 1.1)))
set.seed(1)
rfg1 <- oncoSimulIndiv(fg1, initMutant = "A", onlyCancer = FALSE,
finalTime = 500, seed = NULL)
set.seed(1)
rfg1b <- oncoSimulIndiv(fg1b, initMutant = "A", onlyCancer = FALSE,
finalTime = 500, seed = NULL)
expect_equivalent(rfg1, rfg1b)
g1c <- data.frame(
Genotype = c("WT", "A", "A, B", "B"),
Fitness = c("1",
"1.3 + 1.2 * n_A_B",
"1.1",
"0"),
stringsAsFactors = FALSE)
fg1c <- allFitnessEffects(genotFitness = g1c,
frequencyDependentFitness = TRUE)
set.seed(1)
rfg1c <- oncoSimulIndiv(fg1c, initMutant = "A", onlyCancer = FALSE,
finalTime = 500, seed = NULL)
expect_equivalent(rfg1, rfg1c)
set.seed(1)
oncoSimulIndiv(fg1, initMutant = "A, B", onlyCancer = FALSE,
finalTime = 500, seed = NULL)
} )
test_that("We can run and equal in different order" , {
## Actually, spPopSizes must be order as in the table of fitnessLandscape_df:
## spPopSizes are in the same order as given in "Genotype" in the output of
## evalAllGenotypes
g2 <- data.frame(
Genotype = c("WT", "A", "A, B", "B"),
Fitness = c("1",
"1 + 2 * n_1_2",
"1 + 2 * n_2",
"1 + 2 * n_1"),
stringsAsFactors = FALSE)
fg2 <- allFitnessEffects(genotFitness = g2,
frequencyDependentFitness = TRUE)
## identical, except order of genotypes changes. But the specification is the same
g2b <- data.frame(
Genotype = c("WT", "A", "B", "A, B"),
Fitness = c("1",
"1 + 2 * n_1_2",
"1 + 2 * n_1",
"1 + 2 * n_2"),
stringsAsFactors = FALSE)
fg2b <- allFitnessEffects(genotFitness = g2b,
frequencyDependentFitness = TRUE)
## spPopSizes are for genotypes AT, A, B, AB
ofg2 <- evalAllGenotypes(fg2, spPopSizes = c(9, 2, 6, 3))
ofg2b <- evalAllGenotypes(fg2b, spPopSizes = c(9, 2, 6, 3))
## Are they correct?
expect_identical(ofg2, ofg2b)
out_expec_ofg2 <- c(1, 1 + 2 * 3 , 1 + 2 * 2, 1 + 2 * 6)
expect_identical(ofg2[, "Fitness"], out_expec_ofg2)
})
test_that("Breaks as it should", {
df1 <- data.frame(Genotype = c("WT", "B", "C", "A", "B, A", "C, A"),
Fitness = c("1",
"n_1_3", ## BA
"n_2_3", ## CA
"n_1_2",
"n_1",
"n_2"
))
adf1 <- allFitnessEffects(genotFitness = df1,
frequencyDependentFitness = TRUE)
(adf1)
expect_error(evalAllGenotypes(adf1, spPopSizes = 1:6))
## Breaks in old and new: n_2_3
df1 <- data.frame(Genotype = c("WT", "B", "C", "A", "B, A", "C, A"),
Fitness = c("1",
"f_1_3", ## BA
"f_2_3", ## CA
"f_1_2",
"f_1",
"f_2"
))
adf1 <- allFitnessEffects(genotFitness = df1,
frequencyDependentFitness = TRUE)
(adf1)
expect_error(evalAllGenotypes(adf1, spPopSizes = 1:6)) ## Breaks in old
## and new: f_2_3
})
## Good example of how easy it was to screw up
test_that("eval fitness gives wrong answer, as misspecified", {
df2 <- data.frame(Genotype = c("WT", "B", "C", "A", "B, A", "C, A"),
Fitness = c("1",
"n_1_3", ## you want BA
"n_3",
"n_3",
"n_1",
"n_2"
))
adf2 <- allFitnessEffects(genotFitness = df2,
frequencyDependentFitness = TRUE)
(adf2)
gg <- evalAllGenotypes(adf2, spPopSizes = 1:6)
expect_equal(gg[gg$Genotype == "B", "Fitness"], 6)
## Wrong: gives for B fitness of using CA, not BA (both old and new versions)
})
test_that("Can deal with some user misorderig (and naming spPopSizes)" , {
## Older one could not
df3 <- data.frame(Genotype = c("WT", "B", "C", "A", "B, A", "C, A"),
Fitness = c("1",
"n_B_A", ## you want BA
"n_A",
"n_A",
"n_B",
"n_C"
))
adf3 <- allFitnessEffects(genotFitness = df3,
frequencyDependentFitness = TRUE)
(adf3)
spPopSizes <- 1:6
names(spPopSizes) <- c("WT", "B", "C", "A", "B, A", "C, A")
gg <- evalAllGenotypes(adf3, spPopSizes = spPopSizes)
expect_equal(gg[gg$Genotype == "B", "Fitness"], 5)
})
test_that("Correct values of evalAllGenotypes and substitutions", {
## Old one gives wrong result
df3 <- data.frame(Genotype = c("WT", "B", "C", "A", "B, A", "C, A"),
Fitness = c("1",
"f_A_B", ## you want BA
"f_A",
"f_A",
"f_B",
"f_C"
))
adf3 <- allFitnessEffects(genotFitness = df3,
frequencyDependentFitness = TRUE)
spPopSizes <- 1:6
names(spPopSizes) <- c("WT", "B", "C", "A", "B, A", "C, A")
## spPopSizes <- c(WT = 1, B = 2, C = 3, A = 4, AB=5, CA = 6)
gg <- evalAllGenotypes(adf3, spPopSizes = spPopSizes)
## should give
## c(1, c(A = 4, B = 5, C = 4, AB = 2, AC = 3, BC = 0, ABC = 0)/sum(spPopSizes))
expect_equal(gg[, "Fitness"],
c(1, c(4, 5, 4, 2, 3, 0, 0)/sum(spPopSizes))
)
r <- data.frame(rfitness(3))
r <- r[, c(2, 3, 1, 4)]
r <- r[c(1, 3, 4, 2, 5, 6), ]
r$Fitness <- c("1",
"f_A_B", ## you want BA
"f_A",
"f_A_C",
"f_B",
"f_C"
)
df3m <- r
df3m2 <- data.frame(rfitness(3))[, c(2, 3, 1, 4)][1:6, ]
df3m2$Fitness <- c("1",
"f_A_C",
"f_A_B", ## you want BA
"f_A",
"f_B",
"f_C"
)
## Note the columns are sorted as per message
adf3m <- allFitnessEffects(genotFitness = df3m,
frequencyDependentFitness = TRUE)
adf3m2 <- allFitnessEffects(genotFitness = df3m2,
frequencyDependentFitness = TRUE)
o1 <- evalAllGenotypes(adf3m, spPopSizes = spPopSizes)
o2 <- evalAllGenotypes(adf3m2, spPopSizes = spPopSizes)
expect_equal(o1, o2)
})
test_that("Fail if unequal names spPopSizes", {
r <- data.frame(rfitness(3))
r <- r[, c(2, 3, 1, 4)]
r <- r[c(1, 3, 4, 2, 5, 6), ]
r$Fitness <- c("1",
"f_A_B", ## you want BA
"f_A",
"f_A_C",
"f_B",
"f_C"
)
df3m <- r
df3m2 <- data.frame(rfitness(3))[, c(2, 3, 1, 4)][1:6, ]
df3m2$Fitness <- c("1",
"f_A_C",
"f_A_B", ## you want BA
"f_A",
"f_B",
"f_C"
)
## Note the columns are sorted as per message
adf3m <- allFitnessEffects(genotFitness = df3m,
frequencyDependentFitness = TRUE)
spPopSizes <- 1:6
names(spPopSizes) <- c("WT", "B", "C", "A", "B, A", "B, C")
expect_error(evalAllGenotypes(adf3m, spPopSizes = spPopSizes),
"Genotype names in spPopSizes differ ", fixed = TRUE)
})
test_that("Correct values of evalAllGenotypes and substitutions, some math and changed order genes in n expression", {
## Old one gives wrong result
df3 <- data.frame(Genotype = c("WT", "B", "C", "A", "B, A", "C, A"),
Fitness = c("1",
"log(n_A_B)", ## you want BA
"sqrt(n_A) + 3 * exp(n_C)",
"-1 * n_A + 7 * (n_C_A > 2)",
"2 - 1/n_B",
"min(n_C, n_B) - 0.1 * (n_B_A > 2)"))
adf3 <- allFitnessEffects(genotFitness = df3,
frequencyDependentFitness = TRUE)
spPopSizes <- 1:6
names(spPopSizes) <- c("WT", "B", "C", "A", "B, A", "C, A")
gg <- evalAllGenotypes(adf3, spPopSizes = spPopSizes)
## should give
expect_equal(gg[, "Fitness"],
c(1, c(-4 + 7,
log(5),
sqrt(4) + 3 * exp(3),
2 - 1/2,
2 - 0.1,
0,
0))
)
df3x <- data.frame(Genotype = c("WT", "B", "C", "A", "B, A", "C, A"),
Fitness = c("n_A + 0.1",
"log(n_A_B)", ## you want BA
"sqrt(n_A) + 3 * exp(n_C)",
"-1 * n_A + 7 * (n_C_A > 2)",
"2 - 0.1/n_B",
"min(n_C, n_B) - 1 * (n_B_A > 2)"))
adf3x <- allFitnessEffects(genotFitness = df3x,
frequencyDependentFitness = TRUE)
spPopSizes <- 1:6
names(spPopSizes) <- c("WT", "B", "C", "A", "B, A", "C, A")
ggx <- evalAllGenotypes(adf3x, spPopSizes = spPopSizes)
expect_equal(ggx[, "Fitness"],
c(4.1, c(-4 + 7,
log(5),
sqrt(4) + 3 * exp(3),
2 - 0.1/2,
2 - 1,
0,
0))
)
r <- data.frame(rfitness(3))
r <- r[, c(2, 3, 1, 4)]
r <- r[c(1, 3, 4, 2, 5, 6), ]
r$Fitness <- df3x$Fitness[c(1, 2, 3, 4, 5, 6)]
df3m <- r
df3m2 <- data.frame(rfitness(3))[, c(2, 3, 1, 4)][1:6, ]
df3m2$Fitness <- df3x$Fitness[c(1, 4, 2, 3, 5, 6)]
## Note the columns are sorted as per message
adf3m <- allFitnessEffects(genotFitness = df3m,
frequencyDependentFitness = TRUE)
adf3m2 <- allFitnessEffects(genotFitness = df3m2,
frequencyDependentFitness = TRUE)
o1 <- evalAllGenotypes(adf3m, spPopSizes = spPopSizes)
o2 <- evalAllGenotypes(adf3m2, spPopSizes = spPopSizes)
expect_equal(o1, o2)
expect_equal(o1, ggx)
})
test_that("should run", {
r <- data.frame(Genotype = c("WT", "A", "B", "A, B"),
Fitness = c("10*f_",
"10*f_1",
"50*f_2",
"200*(f_1 + f_2) + 50*f_1_2"))
afe <- allFitnessEffects(genotFitness = r,
frequencyDependentFitness = TRUE,
frequencyType = "rel")
evalAllGenotypes(afe,
spPopSizes = c(WT = 2500, A = 2000,
B = 5500, "A, B" = 700))
gffd0 <- data.frame(
Genotype = c(
"A", "A, B",
"C", "C, D", "C, E"),
Fitness = c(
"1.3",
"1.4",
"1.4",
"1.1 + 0.7*((f_1 + f_A_B) > 0.3)",
"1.2 + sqrt(f_A + f_C + f_C_D)"))
afd0 <- allFitnessEffects(genotFitness = gffd0,
frequencyDependentFitness = TRUE)
sp <- 1:5
names(sp) <- c("A", "C", "A, B", "C, D", "C, E")
eag0 <- evalAllGenotypes(afd0, spPopSizes = sp)
})
cat(paste("\n Ending FDF-small-fitness-specs", date(), "\n"))
cat(paste(" Took ", round(difftime(Sys.time(), inittime, units = "secs"), 2), "\n\n"))
rm(inittime)
