#' Title
#'
#' @param nets
#' @param metrics
#' @param key
#' @param plotFolder
#' @param noFlag
#'
#' @return
#' @export
#'
#' @examples
plotPercentiles <-
function(nets, metrics, key, plotFolder, noFlag = T) {
pat <- ifelse(noFlag, "_ALLnoFlag.csv", "_ALL.csv")
df <- sapply(nets, function(net) {
setwd(randRaw)
setwd(net)
setwd("CompiledData")
allDat <- read_csv(paste0(net, pat), col_types = cols()) %>%
mutate(Net = Network) %>% mutate(Network = net)
gsDat <-
read_csv(paste0(net, "_TeamStrengths.csv"), col_types = cols()) %>%
select(Network, Gs) %>% set_names(c("Net", "Gs"))
allDat <- merge(allDat, gsDat, by = "Net", all = T) %>%
select(Network, Net, all_of(metrics))
wtDat <- allDat %>% filter(Net == "wild")
percentiles <- sapply(metrics, function(x) {
sum(allDat[[x]] < wtDat[[x]], na.rm = T)
})
setwd("..")
percentiles * 100 / nrow(allDat)
})
if (length(metrics) > 1)
df <- t(df)
df <-
df %>% data.frame %>% set_names(labelshorts[metrics] %>% str_replace_all(" ", "\n")) %>%
mutate(Network = netNameKey[nets] %>% factor(levels = netNameKey[nets])) %>%
gather(key = "Metric", value = "Value",-Network) %>%
mutate(Metric = factor(Metric, levels = labelshorts[metrics] %>% str_replace_all(" ", "\n")))
ggplot(df, aes(x = Metric, y = Network, fill = Value)) +
geom_tile() + scale_fill_viridis_c(limits = c(0, 100)) + theme_Publication() +
theme(
legend.position = "right",
legend.direction = "vertical",
legend.key.height = unit(0.8, "cm")
) + labs(x = "", y = "", fill = "Percentile")
DirectoryNav(plotFolder)
ggsave(
paste0(key, "Percentiles.png"),
width = 2.3 + length(metrics),
height = 5
)
}
#' Title
#'
#' @param nets
#' @param metric
#' @param plotFolder
#' @param noFlag
#' @param w
#' @param h
#'
#' @return
#' @export
#'
#' @examples
distribution <- function(nets,
metric,
plotFolder,
noFlag = T,
w = NULL,
h = NULL) {
pat <- ifelse(noFlag, "_ALLnoFlag.csv", "_ALL.csv")
if (metric == "Gs")
pat <- "_TeamStrengths.csv"
df <- lapply(nets, function(net) {
setwd(randRaw)
setwd(net)
setwd("CompiledData")
allDat <- read_csv(paste0(net, pat), col_types = cols()) %>%
select(Network, all_of(metric)) %>%
mutate(Net = Network) %>% mutate(Network = net)
wtDat <- allDat %>% filter(Net == "wild")
ggplot(allDat, aes_string(x = metric)) +
geom_histogram(aes(y = ..count.. / sum(..count..))) +
geom_vline(
data = wtDat,
aes_string(xintercept = metric),
color = "red",
size = 1.5
) +
labs(x = labelvals[metric],
y = "Frequency",
title = netNameKey[net]) +
theme_Publication()
ggsave(
paste0(plotFolder, "/", net, "_", metric, ".png"),
width = 5.5,
height = 5
)
allDat
}) %>% reduce(rbind.data.frame) %>%
mutate(Network = netNameKey[Network] %>% factor(levels = netNameKey[nets]))
wtDf <- df %>% filter(Net == "wild")
ggplot(df, aes_string(x = metric)) +
geom_histogram(aes(y = (..count..) / tapply(..count.., ..PANEL.., sum)[..PANEL..])) +
geom_vline(
data = wtDf,
aes_string(xintercept = metric),
color = "red",
size = 1.5
) +
facet_wrap( ~ Network, scales = "free", nrow = 1) + theme_Publication() +
labs(x = labelvals[metric], y = "Frequency")
h1 <- 5
w1 <- 3 + 2.5 * length(nets)
if (!is.null(h))
h1 <- h
if (!is.null(w))
w1 <- w
ggsave(
paste0(
plotFolder,
"/",
paste0(nets, collapse = "_"),
"_",
metric,
".png"
),
height = h1,
width = w1
)
}
#' Title
#'
#' @param nets
#' @param metric
#' @param plotFolder
#' @param noFlag
#' @param h
#' @param w
#'
#' @return
#' @export
#'
#' @examples
teamStrengthPlots <-
function(nets,
metric,
plotFolder,
noFlag = F,
h = NULL,
w = NULL) {
phen <- paste0(metric, "Phen")
pat <- ifelse(noFlag, "_ALLnoFlag.csv", "_ALL.csv")
# browser()
df <- lapply(nets, function(net) {
# browser()
setwd(randRaw)
setwd(net)
setwd("CompiledData")
allDat <- read_csv(paste0(net, pat), col_types = cols()) %>%
select(Network, all_of(c(metric))) %>%
mutate(Net = Network) %>% mutate(Network = net)
GsDat <-
read_csv(paste0(net, "_teamStrengths.csv")) %>% select(Network, Gs) %>%
set_names(c("Net", "Gs"))
allDat <-
merge(allDat, GsDat, by = "Net", all = T) %>% drop_na
bins <- c(0.2, 0.4, 0.6, 0.8, 1)
bins <- bins[1:which(bins > max(allDat$Gs))[1]]
yVal <-
max(allDat[[metric]]) + 0.1 * (max(allDat[[metric]]) - min(allDat[[metric]]))
cv <- sapply(bins, function(b) {
d <- allDat %>% filter(Gs < b & Gs >= (b - 0.2))
round(sd(d[[metric]], na.rm = T) / mean(d[[metric]], na.rm = T), 2)
}) %>% data.frame(xVal = bins - 0.1,
yVal = yVal,
text = .)
# allDat[[phen]] <- ifelse(allDat[[phen]] == "H", "Hybrid", "Terminal")
c1 <- cor(allDat$Gs, allDat[[metric]])
ypos <- ifelse(c1 < -0.1, 0.9, 0.1)
wtDat <- allDat %>% filter(Net == "wild")
grob <- correlGrob(allDat, metric, "Gs", yPos = ypos)
lab <- "WT"
p <- ggplot(allDat, aes_string(x = "Gs", y = metric)) +
geom_point(size = 1.5) +
annotation_custom(grob) +
geom_text(data = cv, aes(
x = xVal,
y = yVal,
label = text
))
sapply(bins, function(b) {
p <<-
p + geom_vline(
xintercept = b,
color = "red",
linetype = "dashed"
)
})
p <- p +
theme_Publication() +
labs(
x = "Team Strength (Ts)",
y = labelvals[metric],
title = netNameKey[net],
color = "Phenotype"
) +
theme(legend.position = "top") +
geom_label_repel(data = wtDat,
label = lab,
show.legend = F)
ggsave(
paste0(plotFolder, "/", net, "_", metric, "_Ts.png"),
width = 5.5,
height = 5
)
allDat <-
allDat %>% mutate(Gsbinned = round(Gs, 1) %>% as.character)
df <-
allDat %>% select(all_of(metric), Gsbinned) %>% set_names(c("metric", "Gs")) %>%
filter(is.finite(metric),!is.na(metric),!is.nan(metric))
aovDat <- aov(metric ~ Gs , data = df)
pVal <- summary(aovDat)[[1]][["Pr(>F)"]][[1]]
if (pVal < 0.0001)
pValKey <- "P-value < 0.0001"
else if (pVal < 0.001)
pValKey <- "P-value < 0.001"
else if (pVal < 0.01)
pValKey <- "P-value < 0.01"
else if (pVal < 0.05)
pValKey <- "P-value < 0.05"
else
pValKey <- "n.s."
allDat$pVal <- pValKey
# yVal <- min(allDat[[metric]]) + 0.1*(max(allDat[[metric]]) - min(allDat[[metric]]))
# if (pValKey == "n.s.")
yVal <-
max(allDat[[metric]]) + 0.1 * (max(allDat[[metric]]) - min(allDat[[metric]]))
xVal <- unique(allDat$Gsbinned) %>% sort
xVal <- xVal[round((length(xVal) + 1) / 2)]
ggplot(allDat, aes_string(x = "Gsbinned", y = metric)) +
geom_violin() + theme_Publication() +
geom_text(
aes(x, y, label = caption),
data = data.frame(
x = xVal,
y = yVal,
caption = unique(allDat$pVal)
),
size = 5.5,
fontface = "bold"
) +
labs(x = "Team Strength (Ts)",
y = labelvals[metric],
title = netNameKey[net])
ggsave(
paste0(plotFolder, "/", net, "_", metric, "_TsViolin.png"),
width = 5.5,
height = 5
)
corr <-
cor.test(allDat[[metric]], allDat[["Gs"]], method = "spearman")
pVal <- ifelse(corr$p.value < 0.05, "*", "")
allDat %>% select(all_of(c("Gs", metric)), Net, Network, Gsbinned, pVal) %>%
mutate(xVal = xVal, yVal = yVal)
}) %>% reduce(rbind.data.frame) %>%
mutate(Network = netNameKey[Network] %>% factor(levels = netNameKey[nets]))
df$yVal <-
max(df[[metric]]) + 0.1 * (max(df[[metric]]) - min(df[[metric]]))
h1 <- 5
w1 <- 3 + 2.5 * length(nets)
if (!is.null(h))
h1 <- h
if (!is.null(w))
w1 <- w
labelDf <- df %>% select(Network, xVal, yVal, pVal) %>% unique
ggplot(df, aes_string(x = "Gsbinned", y = metric)) +
geom_violin() + theme_Publication() +
geom_text(
aes(x = xVal, y = yVal, label = pVal),
data = labelDf,
size = 5.5,
fontface = "bold"
) +
facet_wrap( ~ Network, nrow = 1, scales = "free_x") +
labs(x = "Team Strength (Ts)", y = labelvals[metric])
ggsave(
paste0(
plotFolder,
"/",
paste0(nets, collapse = "_"),
"_",
metric,
"_Ts.png"
),
height = h1,
width = w1
)
}
#' Title
#'
#' @param nets
#' @param metrics
#' @param metric2
#' @param plotFolder
#' @param noFlag
#'
#' @return
#' @export
#'
#' @examples
correlationPlots <-
function(nets,
metrics,
metric2,
plotFolder,
noFlag = F) {
# phen <- paste0(metric, "Phen")
pat <- ifelse(noFlag, "_ALLnoFlag.csv", "_ALL.csv")
df <- lapply(nets, function(net) {
print(net)
setwd(randRaw)
setwd(net)
setwd("CompiledData")
allDat <- read_csv(paste0(net, pat), col_types = cols()) %>%
# select(Network, all_of(c(metric, phen))) %>%
mutate(Net = Network) %>% mutate(Network = net)
GsDat <-
read_csv(paste0(net, "_teamStrengths.csv")) %>% select(Network, Gs) %>%
set_names(c("Net", "Gs"))
allDat <- merge(allDat, GsDat, by = "Net", all = T) %>%
select(Network, Net, all_of(c(metrics, metric2))) %>% drop_na %>%
gather(key = "Metric",
value = "Value",
-all_of(metric2),
-Network,
-Net) %>%
mutate(Metric = labelshorts[Metric] %>% str_replace_all(" ", "\n"))
multiFactorCorrelation(allDat, "Metric", "Value", "Gs", label = F) %>%
mutate(Network = netNameKey[net])
}) %>% reduce(rbind.data.frame) %>%
mutate(
Label = ifelse(pValue < 0.05, "", "X"),
Factors = factor(Factors, levels = labelshorts[metrics] %>% str_replace_all(" ", "\n"))
)
ggplot(df, aes(x = Factors, y = Network, fill = Correlation)) +
geom_tile() + geom_text(aes(label = Label)) +
theme_Publication() +
scale_x_discrete(expand = c(0, 0)) +
theme(
#axis.text.x = element_text(angle = 60, hjust = 1),
legend.position = "right",
legend.direction = "vertical",
legend.key.height = unit(0.8, "cm")
) +
scale_fill_gradient2(low = "blue",
high = "red",
limits = c(-1, 1)) +
labs(
x = "",
y = "",
fill = paste0("Correlation\n", labelshorts[metric2], "\n")
)
ggsave(
paste0(
plotFolder,
"/",
paste0(metrics, collapse = "_"),
"_",
metric2,
".png"
),
width = 2.3 + length(metrics),
height = 5
)
}
#' Title
#'
#' @param nets
#' @param metric
#' @param minMetric
#' @param maxMetric
#' @param plotFolder
#' @param noFlag
#'
#' @return
#' @export
#'
#' @examples
minMaxViolins <-
function(nets,
metric,
minMetric,
maxMetric,
plotFolder,
noFlag = F) {
dfAll <- lapply(nets, function(net) {
# browser()
setwd(randRaw)
setwd(net)
key <- ifelse(noFlag, "noFlag", "")
randDf <-
paste0("CompiledData/", net, "_ALL", key, ".csv") %>%
read_csv(col_types = cols()) %>%
select(all_of(c(minMetric, maxMetric))) %>%
gather(key = "Network", value = "Metric") %>%
mutate(Network = paste0("Rand\n", labelshorts[Network] %>% str_replace_all(" ", "\n")))
wtDf <- read_csv("wild_finFlagFreq.csv", col_types = cols())
if (noFlag)
wtDf <- wtDf %>% filter(flag == 1)
wtDf <-
wtDf %>% mutate(Network = "WT") %>% select(Network, all_of(metric)) %>%
set_names(c("Network", "Metric"))
df <- rbind.data.frame(wtDf, randDf) %>%
mutate(Network = factor(
Network,
levels = c(
paste0("Rand ", labelshorts[minMetric]),
"WT",
paste0("Rand ", labelshorts[maxMetric])
) %>%
str_replace_all(" ", "\n")
))
p <-
ggplot(df, aes(x = Network, y = Metric)) + geom_violin() + theme_Publication() +
labs(x = "",
y = labelshorts[metric],
title = netNameKey[net])
if (metric == "Avg0")
p <- p + scale_y_log10()
setwd(plotFolder)
ggsave(
paste0(net, "_", labelshorts[metric], "_violin.png"),
width = 5.5,
height = 5
)
df %>% mutate(Net = netNameKey[net])
}) %>% reduce(rbind.data.frame)
p <-
ggplot(dfAll, aes(x = Network, y = Metric)) + geom_violin() + theme_Publication() +
labs(x = "", y = labelshorts[metric]) + facet_wrap( ~ Net, nrow = 1)
if (metric == "Avg0")
p <- p + scale_y_log10()
setwd(plotFolder)
ggsave(
paste0(paste0(c(
nets, labelshorts[metric]
), collapse = "_"), "_violin.png"),
width = 3 + 2.5 * length(nets),
height = 5
)
}
#' Title
#'
#' @param net
#'
#' @return
#' @export
#'
#' @examples
singleCoherenceRand <- function(net) {
setwd(paste0(randRaw, "/", net))
dfTeams <-
read_csv(paste0("CompiledData/", net, "_TeamStrengths.csv"))
setwd("signalCoherence")
fils <- list.files(".", "_signalCoherence.csv")
netz <- fils %>% str_remove("_signal.*")
df <- lapply(fils, function(x) {
Net <- x %>% str_remove("_signalCoherence.csv")
d <- read_csv(x)
groupLabels <-
readLines("../wild_nodes.txt") %>% str_split(",")
eNodes <- groupLabels[[1]] %>% length
mNodes <- groupLabels[[2]] %>% length
d %>% mutate(
Core = (Epithelial * eNodes + Mesenchymal * mNodes) / (eNodes + mNodes),
Network = str_remove(x, "_sign.*")
)
}) %>% reduce(rbind.data.frame)
df <-
merge(df,
dfTeams %>% select(Network, Gs),
by = "Network",
all.x = T)
df2 <- df %>%
group_by(Network, initState, initPhen, signal) %>%
summarise(
Epithelial = sum(Freq * Epithelial),
Mesenchymal = sum(Freq * Mesenchymal),
Gs = unique(Gs),
Core = sum(Freq * Core),
.groups = "drop"
) %>%
group_by(Network, initPhen, signal) %>%
summarise(
Epithelial = mean(Epithelial),
Mesenchymal = mean(Mesenchymal),
Gs = unique(Gs),
Core = mean(Core),
.groups = "drop"
)
}
