R/create_contour.R
In angy89/TinderMIX: TinderMIX: An R package to cluster gene expression by contour plots
Defines functions
plot3d
plot_contour_plot
create_contour
model_fitting_and_validation
Documented in
create_contour
plot3d
model_fitting_and_validation = function(Exp, pvalFitting= 0.05, modelSelection=1:3){
#TODO: if we decide to add different models, e.g. hill we can ask the user which method they want to use (e.g. with an extra parameter modelSelectionStrategy = c("anova","bic"))
# or we can still choose the optimal across the polynomials with the anova strategy and then compare this optimal one against the other by comparing their akaike criterion values
# fitting linear, poly2 and poly3 models
ModList = list()
null_model <- stats::lm(Exp ~ 1, data = Exp)
lin_model <- stats::lm(Exp ~ Dose + Time,data = Exp)
poly2_model <- stats::lm(Exp ~ I(Dose^2) + I(Time^2) + I(Dose * Time) + Dose + Time, data = Exp)
poly3_model <- stats::lm(Exp ~ I(Dose^3) + I(Dose^2 * Time) + I(Dose * Time^2) + I(Time^3) + I(Dose^2) + I(Time^2) + I(Dose * Time) + Dose + Time ,data = Exp)
ModList[["linear"]] = lin_model
ModList[["poly2"]] = poly2_model
ModList[["poly3"]] = poly3_model
#store fitted models in a list without names for do.call on anova
MML = list(lin_model, poly2_model, poly3_model)
# select only the ones in modelSelection parameter
ModList = ModList[modelSelection]
MML = MML[modelSelection]
# computing model statistics and p-value
SingleStats = c()
for(model in ModList){
f <- summary(model)$fstatistic
p <- stats::pf(f[1],f[2],f[3],lower.tail=F)
adj.r.square = summary(model)$adj.r.squared
RSS <- c(crossprod(model$residuals))
MSE <- RSS / length(model$residuals)
RMSE <- sqrt(MSE)
BICmod = stats::BIC(model)
AICmod = stats::AIC(model)
SingleStats = rbind(SingleStats, c(p, adj.r.square,RMSE, BICmod, AICmod))
}
colnames(SingleStats) = c("PValue","Adj.R.Square","RMSE","BIC","AIC")
rownames(SingleStats) = names(ModList)
# create a list including the null model and compute the anova between the null model, the linear, the poly2 and poly3 models
MMListN = c(list(null_model), MML)
anovaTest = do.call(anova, MMListN)
#take anova pvalues
anovaPVal = anovaTest$`Pr(>F)`
#assign to pvalues model names
names(anovaPVal) = c("Null",names(ModList))
#set to 1 the pvalue for the null model
anovaPVal[is.na(anovaPVal)] = 1
#identify the smallest pvalue
optIdx = which.min(anovaPVal)
# get the statistics for the model associated to the smallest pvalue
SS = SingleStats[(optIdx-1),]
# the model associated with the smallest pvalue is returned. We don't filter for sign pvalues here, since we will probably apply a correction after
optModel = ModList[[(optIdx-1)]]
SS["PValue"] = anovaPVal[optIdx] #N.B. the pvalue returned for the model is the one of the anova with respect to the previous model in the nested hierarchy
SS = c(SS, names(ModList)[(optIdx-1)])
names(SS)[length(SS)] = "OptMod"
toRet = list(optModel = optModel, stats = SS, modelsStats = SingleStats,anovaTest=anovaTest)
return(toRet)
# # if the model with smallest pvalue is significative with respect to his previous model in the anova and with respect to the null model, than that gene is selected as the otpimal one
# if(anovaPVal[optIdx] < pvalFitting & SS["PValue"]<pvalFitting){
# optModel = ModList[[(optIdx-1)]]
# SS["PValue"] = anovaPVal[optIdx] #N.B. the pvalue returned for the model is the one of the anova with respect to the previous model in the nested hierarchy
# SS = c(SS, names(ModList)[(optIdx-1)])
# names(SS)[length(SS)] = "OptMod"
# toRet = list(optModel = optModel, stats = SS)
# return(toRet)
# }else{
# #otherwise the gene is considered not fitted and the function return an empy list
# return(list(c(),SS*NA))
# }
}
#'
#' This function fits a 3D regression model for every gene in the dataset and creates an N x N contour plot
#'
#' @importFrom reshape melt
#' @importFrom utils txtProgressBar setTxtProgressBar
#' @importFrom stats predict lm pf
#' @param exp_data is the expression matrix with genes on the rows and samples on the columns
#' @param pheno_data is a dataframe with phenodata informations. Samples are on the rows. The columns should include the dose and time point information.
#' @param responsive_genes responsive_genes character vector with the genes statistically significant for the two-way anova
#' @param dose_index numeric value specifing the column of the phenodata table containing the doses
#' @param time_point_index numeric value specifing the column of the phenodata table containing the time points
#' @param gridSize numeric value specifing size of the z-grid
#' @param logScale boolean specifying if the fitting is performed by using the dose and time in log or linear scale
#' @param modelSelection is a vector of indices specifying which model to fit. 1:linear 2: poly2, 3: poly3
#' @return a list with list with estimated contour objects, 3D fitted objects, fitting statistics and feature values for time and dose
#' \item{GenesMap}{a matrix with the z-maps computed for each gene}
#' \item{RPGenes}{a list with the 3D fitted objects}
#' \item{Statis}{a matrix witht the fitting statistics: PValue,Adj.R.Square,RMSE}
#' \item{DFList}{a list with the data used for the fitting}
#' \item{ModList}{a list with the fitted models}
#' @export
#'
#modelType string showing the model to fit. Possible options: poly2 and loess
create_contour = function(exp_data, pheno_data, responsive_genes,dose_index, time_point_index, gridSize = 50,
pvalFitting.adj.method = "fdr",pvalFitting=0.05, logScale = FALSE,
modelSelection = c(1,2)){ #models is a vector of indices specifying which model to fit. 1:linear 2: poly2, 3: poly3
GenesMap = matrix(NA, ncol = length(responsive_genes), nrow = (gridSize*gridSize) )
colnames(GenesMap) = responsive_genes
RPGenes = list()
DFList = list()
ModList = list()
Stats = matrix(0, nrow = length(responsive_genes), ncol = 6)
rownames(Stats) = responsive_genes
colnames(Stats) = c("PValue","Adj.R.Square","RMSE","BIC","AIC", "OptMod")
index = 1
pb = txtProgressBar(min = 1, max = length(responsive_genes), style = 3)
for(g in responsive_genes){
Exp = reshape::melt(exp_data[g,])
Exp = cbind(rownames(Exp), Exp)
colnames(Exp)[1] = "variable"
Exp = cbind(Exp, pheno_data[as.character(Exp$variable),dose_index])
Exp = cbind(Exp, pheno_data[as.character(Exp$variable),time_point_index])
colnames(Exp) = c("Sample","Exp","Dose","Time")
Exp = as.data.frame(Exp)
Exp$Dose = as.numeric(as.vector(Exp$Dose))
Exp$Time = as.numeric(as.vector(Exp$Time))
if(logScale){
Exp$Dose = log(Exp$Dose)
Exp$Time = log(Exp$Time)
#fit three different models and store the results in a list
fitted_models = model_fitting_and_validation(Exp, modelSelection=modelSelection)
optModel = fitted_models$optModel
Stats[index,] = fitted_models$stats
x1 <-range(as.numeric(as.vector(exp(Exp$Dose))))
x1 <- seq(x1[1], x1[2], length.out=gridSize)
y1 <- range(as.numeric(as.vector(exp(Exp$Time))))
y1 <- seq(y1[1], y1[2], length.out=gridSize)
x <-range(as.numeric(as.vector(Exp$Dose)))
x <- seq(x[1], x[2], length.out=gridSize)
y <- range(as.numeric(as.vector(Exp$Time)))
y <- seq(y[1], y[2], length.out=gridSize)
z <- outer(x,y, function(Dose,Time) stats::predict(optModel, data.frame(Dose,Time)))
List3d=list(x,y,z)
#plot3d(toPlot = List3d,logScale = TRUE, DF = Exp)
}else{
fitted_models = model_fitting_and_validation(Exp, modelSelection=modelSelection)
optModel = fitted_models$optModel
Stats[index,] = fitted_models$stats
x <-range(as.numeric(as.vector(Exp$Dose)))
x <- seq(x[1], x[2], length.out=gridSize)
y <- range(as.numeric(as.vector(Exp$Time)))
y <- seq(y[1], y[2], length.out=gridSize)
#Z is a matrix dose X time Z[1,1] = first time and first dose
z <- outer(x,y, function(Dose,Time) stats::predict(optModel, data.frame(Dose,Time)))
List3d=list(x,y,z)
#plot3d(toPlot = List3d,logScale = FALSE, DF = Exp)
}
#GenesMap = cbind(GenesMap,as.vector(z)) # I want to compute the distance by the genes based on the mapping z
GenesMap[,g] = as.vector(z)
RPGenes[[g]] = List3d
DFList[[g]] = Exp
ModList[[g]] = optModel
# p = plot3d(toPlot = list(x,y,z),DF = Exp)
# p
setTxtProgressBar(pb,index)
index = index + 1
}
colnames(GenesMap) = responsive_genes#responsive_genes
close(pb)
SST = as.data.frame(Stats)
SST[,1] = as.numeric(as.vector(SST[,1]))
SST[,2] = as.numeric(as.vector(SST[,2]))
SST[,3] = as.numeric(as.vector(SST[,3]))
adj.pval = p.adjust(SST[,1],method = pvalFitting.adj.method)
SST = cbind(adj.pval,SST)
ggenes = rownames(SST)[SST[,1]<pvalFitting]
Stats = SST
return(list(GenesMap=GenesMap, RPGenes = RPGenes,Stats=Stats,DFList=DFList,ggenes=ggenes, ModList=ModList))
}
plot_contour_plot = function(immy, coord,geneName){
# geneName = "Hspbp1" #"Acmsd
# immy = contour_res$RPGenes[[geneName]][[3]]
#
# immy2 = clpr[[1]][[3]]
# coord = cbind(contour_res$RPGenes[[geneName]][[1]],contour_res$RPGenes[[geneName]][[2]])
res2 = compute_BMD_IC50(immy,coord, geneName,
activity_threshold = activity_threshold,
BMD_resonse_threhold = BMD_resonse_threhold,
mode = mode,
nTimeInt = nTimeInt,nDoseInt=nDoseInt,
timeLabels = timeLabels,
doseLabels = doseLabels, toPlot = TRUE, tosave = FALSE)
}
#'
#'This function plots the fitted 3d surface for the expression value of a gene
#'
#' @param toPlot is a list containing the predicted value for the x, y and z axis
#' @param DF is the data frame containing the information for the samples used in the fitting process
#' @return a plotly object
#'
#' @export
#'
#'
plot3d = function(toPlot = list(x,y,z), DF, logScale = FALSE){
if(logScale){
p = plot_ly(x=~toPlot[[1]], y=~toPlot[[2]], z=~t(toPlot[[3]]),
colors = c("#f5cb11", "#b31d83"),type="surface") %>%
add_trace(data=DF, x=exp(DF$Dose), y=exp(DF$Time), z=DF$Exp, mode="markers", type="scatter3d",
marker = list(opacity=0.7, symbol=105)) %>%
layout(scene = list(
aspectmode = "manual",
aspectratio = list(x=1, y=1, z=1),
xaxis = list(title = "Dose"),
yaxis = list(title = "Time"),
zaxis = list(title = "Exp")))
}else{
p = plot_ly(x=~toPlot[[1]], y=~toPlot[[2]], z=~t(toPlot[[3]]),
colors = c("#f5cb11", "#b31d83"),type="surface") %>%
add_trace(data=DF, x=DF$Dose, y=DF$Time, z=DF$Exp, mode="markers", type="scatter3d",
marker = list(opacity=0.7, symbol=105)) %>%
layout(scene = list(
aspectmode = "manual",
aspectratio = list(x=1, y=1, z=1),
xaxis = list(title = "Dose"),
yaxis = list(title = "Time"),
zaxis = list(title = "Exp")))
}
# p = plot_ly(x=~exp(x), y=~exp(y), z=~t(z),
# colors = c("#f5cb11", "#b31d83"),type="surface") %>%
# add_trace(data=pd, x=exp(pd$Dose), y=exp(pd$Day), z=pd$Counts, mode="markers", type="scatter3d",
# marker = list(opacity=0.7, symbol=105)) %>%
# layout(scene = list(
# aspectmode = "manual",
# aspectratio = list(x=1, y=1, z=1),
# xaxis = list(title = "Dose"),
# yaxis = list(title = "Time"),
# zaxis = list(title = "Counts")))
return(p)
}
angy89/TinderMIX documentation built on Nov. 26, 2020, 9:26 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Defines functions plot3d plot_contour_plot create_contour model_fitting_and_validation
Documented in create_contour plot3d
model_fitting_and_validation = function(Exp, pvalFitting= 0.05, modelSelection=1:3){
#TODO: if we decide to add different models, e.g. hill we can ask the user which method they want to use (e.g. with an extra parameter modelSelectionStrategy = c("anova","bic"))
# or we can still choose the optimal across the polynomials with the anova strategy and then compare this optimal one against the other by comparing their akaike criterion values
# fitting linear, poly2 and poly3 models
ModList = list()
null_model <- stats::lm(Exp ~ 1, data = Exp)
lin_model <- stats::lm(Exp ~ Dose + Time,data = Exp)
poly2_model <- stats::lm(Exp ~ I(Dose^2) + I(Time^2) + I(Dose * Time) + Dose + Time, data = Exp)
poly3_model <- stats::lm(Exp ~ I(Dose^3) + I(Dose^2 * Time) + I(Dose * Time^2) + I(Time^3) + I(Dose^2) + I(Time^2) + I(Dose * Time) + Dose + Time ,data = Exp)
ModList[["linear"]] = lin_model
ModList[["poly2"]] = poly2_model
ModList[["poly3"]] = poly3_model
#store fitted models in a list without names for do.call on anova
MML = list(lin_model, poly2_model, poly3_model)
# select only the ones in modelSelection parameter
ModList = ModList[modelSelection]
MML = MML[modelSelection]
# computing model statistics and p-value
SingleStats = c()
for(model in ModList){
f <- summary(model)$fstatistic
p <- stats::pf(f[1],f[2],f[3],lower.tail=F)
adj.r.square = summary(model)$adj.r.squared
RSS <- c(crossprod(model$residuals))
MSE <- RSS / length(model$residuals)
RMSE <- sqrt(MSE)
BICmod = stats::BIC(model)
AICmod = stats::AIC(model)
SingleStats = rbind(SingleStats, c(p, adj.r.square,RMSE, BICmod, AICmod))
}
colnames(SingleStats) = c("PValue","Adj.R.Square","RMSE","BIC","AIC")
rownames(SingleStats) = names(ModList)
# create a list including the null model and compute the anova between the null model, the linear, the poly2 and poly3 models
MMListN = c(list(null_model), MML)
anovaTest = do.call(anova, MMListN)
#take anova pvalues
anovaPVal = anovaTest$`Pr(>F)`
#assign to pvalues model names
names(anovaPVal) = c("Null",names(ModList))
#set to 1 the pvalue for the null model
anovaPVal[is.na(anovaPVal)] = 1
#identify the smallest pvalue
optIdx = which.min(anovaPVal)
# get the statistics for the model associated to the smallest pvalue
SS = SingleStats[(optIdx-1),]
# the model associated with the smallest pvalue is returned. We don't filter for sign pvalues here, since we will probably apply a correction after
optModel = ModList[[(optIdx-1)]]
SS["PValue"] = anovaPVal[optIdx] #N.B. the pvalue returned for the model is the one of the anova with respect to the previous model in the nested hierarchy
SS = c(SS, names(ModList)[(optIdx-1)])
names(SS)[length(SS)] = "OptMod"
toRet = list(optModel = optModel, stats = SS, modelsStats = SingleStats,anovaTest=anovaTest)
return(toRet)
# # if the model with smallest pvalue is significative with respect to his previous model in the anova and with respect to the null model, than that gene is selected as the otpimal one
# if(anovaPVal[optIdx] < pvalFitting & SS["PValue"]<pvalFitting){
# optModel = ModList[[(optIdx-1)]]
# SS["PValue"] = anovaPVal[optIdx] #N.B. the pvalue returned for the model is the one of the anova with respect to the previous model in the nested hierarchy
# SS = c(SS, names(ModList)[(optIdx-1)])
# names(SS)[length(SS)] = "OptMod"
# toRet = list(optModel = optModel, stats = SS)
# return(toRet)
# }else{
# #otherwise the gene is considered not fitted and the function return an empy list
# return(list(c(),SS*NA))
# }
}
#'
#' This function fits a 3D regression model for every gene in the dataset and creates an N x N contour plot
#'
#' @importFrom reshape melt
#' @importFrom utils txtProgressBar setTxtProgressBar
#' @importFrom stats predict lm pf
#' @param exp_data is the expression matrix with genes on the rows and samples on the columns
#' @param pheno_data is a dataframe with phenodata informations. Samples are on the rows. The columns should include the dose and time point information.
#' @param responsive_genes responsive_genes character vector with the genes statistically significant for the two-way anova
#' @param dose_index numeric value specifing the column of the phenodata table containing the doses
#' @param time_point_index numeric value specifing the column of the phenodata table containing the time points
#' @param gridSize numeric value specifing size of the z-grid
#' @param logScale boolean specifying if the fitting is performed by using the dose and time in log or linear scale
#' @param modelSelection is a vector of indices specifying which model to fit. 1:linear 2: poly2, 3: poly3
#' @return a list with list with estimated contour objects, 3D fitted objects, fitting statistics and feature values for time and dose
#' \item{GenesMap}{a matrix with the z-maps computed for each gene}
#' \item{RPGenes}{a list with the 3D fitted objects}
#' \item{Statis}{a matrix witht the fitting statistics: PValue,Adj.R.Square,RMSE}
#' \item{DFList}{a list with the data used for the fitting}
#' \item{ModList}{a list with the fitted models}
#' @export
#'
#modelType string showing the model to fit. Possible options: poly2 and loess
create_contour = function(exp_data, pheno_data, responsive_genes,dose_index, time_point_index, gridSize = 50,
pvalFitting.adj.method = "fdr",pvalFitting=0.05, logScale = FALSE,
modelSelection = c(1,2)){ #models is a vector of indices specifying which model to fit. 1:linear 2: poly2, 3: poly3
GenesMap = matrix(NA, ncol = length(responsive_genes), nrow = (gridSize*gridSize) )
colnames(GenesMap) = responsive_genes
RPGenes = list()
DFList = list()
ModList = list()
Stats = matrix(0, nrow = length(responsive_genes), ncol = 6)
rownames(Stats) = responsive_genes
colnames(Stats) = c("PValue","Adj.R.Square","RMSE","BIC","AIC", "OptMod")
index = 1
pb = txtProgressBar(min = 1, max = length(responsive_genes), style = 3)
for(g in responsive_genes){
Exp = reshape::melt(exp_data[g,])
Exp = cbind(rownames(Exp), Exp)
colnames(Exp)[1] = "variable"
Exp = cbind(Exp, pheno_data[as.character(Exp$variable),dose_index])
Exp = cbind(Exp, pheno_data[as.character(Exp$variable),time_point_index])
colnames(Exp) = c("Sample","Exp","Dose","Time")
Exp = as.data.frame(Exp)
Exp$Dose = as.numeric(as.vector(Exp$Dose))
Exp$Time = as.numeric(as.vector(Exp$Time))
if(logScale){
Exp$Dose = log(Exp$Dose)
Exp$Time = log(Exp$Time)
#fit three different models and store the results in a list
fitted_models = model_fitting_and_validation(Exp, modelSelection=modelSelection)
optModel = fitted_models$optModel
Stats[index,] = fitted_models$stats
x1 <-range(as.numeric(as.vector(exp(Exp$Dose))))
x1 <- seq(x1[1], x1[2], length.out=gridSize)
y1 <- range(as.numeric(as.vector(exp(Exp$Time))))
y1 <- seq(y1[1], y1[2], length.out=gridSize)
x <-range(as.numeric(as.vector(Exp$Dose)))
x <- seq(x[1], x[2], length.out=gridSize)
y <- range(as.numeric(as.vector(Exp$Time)))
y <- seq(y[1], y[2], length.out=gridSize)
z <- outer(x,y, function(Dose,Time) stats::predict(optModel, data.frame(Dose,Time)))
List3d=list(x,y,z)
#plot3d(toPlot = List3d,logScale = TRUE, DF = Exp)
}else{
fitted_models = model_fitting_and_validation(Exp, modelSelection=modelSelection)
optModel = fitted_models$optModel
Stats[index,] = fitted_models$stats
x <-range(as.numeric(as.vector(Exp$Dose)))
x <- seq(x[1], x[2], length.out=gridSize)
y <- range(as.numeric(as.vector(Exp$Time)))
y <- seq(y[1], y[2], length.out=gridSize)
#Z is a matrix dose X time Z[1,1] = first time and first dose
z <- outer(x,y, function(Dose,Time) stats::predict(optModel, data.frame(Dose,Time)))
List3d=list(x,y,z)
#plot3d(toPlot = List3d,logScale = FALSE, DF = Exp)
}
#GenesMap = cbind(GenesMap,as.vector(z)) # I want to compute the distance by the genes based on the mapping z
GenesMap[,g] = as.vector(z)
RPGenes[[g]] = List3d
DFList[[g]] = Exp
ModList[[g]] = optModel
# p = plot3d(toPlot = list(x,y,z),DF = Exp)
# p
setTxtProgressBar(pb,index)
index = index + 1
}
colnames(GenesMap) = responsive_genes#responsive_genes
close(pb)
SST = as.data.frame(Stats)
SST[,1] = as.numeric(as.vector(SST[,1]))
SST[,2] = as.numeric(as.vector(SST[,2]))
SST[,3] = as.numeric(as.vector(SST[,3]))
adj.pval = p.adjust(SST[,1],method = pvalFitting.adj.method)
SST = cbind(adj.pval,SST)
ggenes = rownames(SST)[SST[,1]<pvalFitting]
Stats = SST
return(list(GenesMap=GenesMap, RPGenes = RPGenes,Stats=Stats,DFList=DFList,ggenes=ggenes, ModList=ModList))
}
plot_contour_plot = function(immy, coord,geneName){
# geneName = "Hspbp1" #"Acmsd
# immy = contour_res$RPGenes[[geneName]][[3]]
#
# immy2 = clpr[[1]][[3]]
# coord = cbind(contour_res$RPGenes[[geneName]][[1]],contour_res$RPGenes[[geneName]][[2]])
res2 = compute_BMD_IC50(immy,coord, geneName,
activity_threshold = activity_threshold,
BMD_resonse_threhold = BMD_resonse_threhold,
mode = mode,
nTimeInt = nTimeInt,nDoseInt=nDoseInt,
timeLabels = timeLabels,
doseLabels = doseLabels, toPlot = TRUE, tosave = FALSE)
}
#'
#'This function plots the fitted 3d surface for the expression value of a gene
#'
#' @param toPlot is a list containing the predicted value for the x, y and z axis
#' @param DF is the data frame containing the information for the samples used in the fitting process
#' @return a plotly object
#'
#' @export
#'
#'
plot3d = function(toPlot = list(x,y,z), DF, logScale = FALSE){
if(logScale){
p = plot_ly(x=~toPlot[[1]], y=~toPlot[[2]], z=~t(toPlot[[3]]),
colors = c("#f5cb11", "#b31d83"),type="surface") %>%
add_trace(data=DF, x=exp(DF$Dose), y=exp(DF$Time), z=DF$Exp, mode="markers", type="scatter3d",
marker = list(opacity=0.7, symbol=105)) %>%
layout(scene = list(
aspectmode = "manual",
aspectratio = list(x=1, y=1, z=1),
xaxis = list(title = "Dose"),
yaxis = list(title = "Time"),
zaxis = list(title = "Exp")))
}else{
p = plot_ly(x=~toPlot[[1]], y=~toPlot[[2]], z=~t(toPlot[[3]]),
colors = c("#f5cb11", "#b31d83"),type="surface") %>%
add_trace(data=DF, x=DF$Dose, y=DF$Time, z=DF$Exp, mode="markers", type="scatter3d",
marker = list(opacity=0.7, symbol=105)) %>%
layout(scene = list(
aspectmode = "manual",
aspectratio = list(x=1, y=1, z=1),
xaxis = list(title = "Dose"),
yaxis = list(title = "Time"),
zaxis = list(title = "Exp")))
}
# p = plot_ly(x=~exp(x), y=~exp(y), z=~t(z),
# colors = c("#f5cb11", "#b31d83"),type="surface") %>%
# add_trace(data=pd, x=exp(pd$Dose), y=exp(pd$Day), z=pd$Counts, mode="markers", type="scatter3d",
# marker = list(opacity=0.7, symbol=105)) %>%
# layout(scene = list(
# aspectmode = "manual",
# aspectratio = list(x=1, y=1, z=1),
# xaxis = list(title = "Dose"),
# yaxis = list(title = "Time"),
# zaxis = list(title = "Counts")))
return(p)
}
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.