R/vis_covariates_continuous.R
In EESI/themetagenomics: Exploring Thematic Structure and Predicted Functionality of 16s rRNA Amplicon Data
Defines functions
vis.continuous
Documented in
vis.continuous
#' @rdname vis
#'
#' @param taxa_reg_n Number of most relevant taxa within topic to regress. Defaults to 8.
#'
#' @export
vis.continuous <- function(object,lambda_step=.1,taxa_reg_n=8,...){
topics <- object$topics
topic_effects <- object$topic_effects
otu_table <- object$topics$otu_table
tax_table <- object$topics$tax_table
metadata <- object$modelframe
covariates <- colnames(metadata)
cov_f <- sapply(metadata,class) == 'factor'
cov_cont <- covariates[!cov_f]
if (length(cov_cont) == 0) stop('For continuous, formula provided must have contained a continuous, numeric, or integer covariate.')
cov_fact <- covariates[cov_f]
names(cov_cont) <- tolower(unlist(cov_cont))
mod_fact <- names(topic_effects[[cov_cont[1]]]$fitted_switch)
pretty_names <- pretty_taxa_names(tax_table)
otu_table <- otu_table + 1
fit <- topics$fit
vocab <- fit$vocab
K <- fit$settings$dim$K
doc_lengths <- sapply(topics$docs,function(x) sum(x[2,]))
theta <- fit$theta
rownames(theta) <- names(topics$docs)
colnames(theta) <- paste0('T',1:K)
if (min(theta) == 0){
min_theta <- min(theta[theta > 0])
theta <- theta + min_theta
theta <- theta/rowSums(theta)
}
beta <- exp(fit$beta$logbeta[[1]])
colnames(beta) <- vocab
rownames(beta) <- paste0('T',1:K)
if (min(beta) == 0){
min_beta <- min(beta[beta > 0])
beta <- beta + min_beta
beta <- beta/rowSums(beta)
}
topic_freq <- colSums(theta*doc_lengths)
topic_marg <- topic_freq/sum(topic_freq)
term_topic_freq <- beta*topic_freq
# compute term frequencies as column sums of term.topic.frequency
# we actually won't use the user-supplied term.frequency vector.
# the term frequencies won't match the user-supplied frequencies exactly
# this is a work-around to solve the bug described in Issue #32 on github:
# https://github.com/cpsievert/LDAvis/issues/32
term_freq <- colSums(term_topic_freq) # topic_freq <- colSums(theta*doc_lengths)
# marginal distribution over terms (width of blue bars)
term_marg <- term_freq/sum(term_freq)
beta <- t(beta)
# compute the distinctiveness and saliency of the terms:
# this determines the R terms that are displayed when no topic is selected
topic_g_term <- beta/rowSums(beta) # (W x K)
kernel <- topic_g_term*log(t(t(topic_g_term)/topic_marg))
distinctiveness <- rowSums(kernel)
saliency <- term_marg*distinctiveness
# Order the terms for the "default" view by decreasing saliency:
default_terms <- vocab[order(saliency,decreasing=TRUE)][1:taxa_reg_n]
counts <- as.integer(term_freq[match(default_terms,vocab)])
taxa_n_rev <- rev(seq_len(taxa_reg_n))
default <- data.frame(Term=default_terms,
logprob=taxa_n_rev,
loglift=taxa_n_rev,
Freq=counts,
Total=counts,
Category='Default',
stringsAsFactors=FALSE)
default$Term <- factor(default$Term,levels=rev(default$Term),ordered=TRUE)
topic_seq <- rep(seq_len(K),each=taxa_reg_n)
category <- paste0('T',topic_seq)
lift <- beta/term_marg
# Collect taxa_reg_n most relevant terms for each topic/lambda combination
# Note that relevance is re-computed in the browser, so we only need
# to send each possible term/topic combination to the browser
find_relevance <- function(i){
relevance <- i*log(beta) + (1 - i)*log(lift)
idx <- apply(relevance,2,function(x) order(x,decreasing=TRUE)[seq_len(taxa_reg_n)])
indices <- cbind(c(idx),topic_seq)
data.frame(Term=vocab[idx],
Category=category,
logprob=round(log(beta[indices]),4),
loglift=round(log(lift[indices]),4),
stringsAsFactors=FALSE)
}
lambda_seq <- seq(0,1,by=lambda_step) # 0.01
tinfo <- lapply(as.list(lambda_seq),find_relevance)
tinfo <- unique(do.call('rbind', tinfo))
tinfo$Total <- term_freq[match(tinfo$Term,vocab)]
rownames(term_topic_freq) <- paste0('T',seq_len(K))
colnames(term_topic_freq) <- vocab
tinfo$Freq <- term_topic_freq[as.matrix(tinfo[c('Category','Term')])]
tinfo <- rbind(default[,-7],tinfo)
new_order <- default
shinyApp(
ui <- fluidPage(
tags$head(tags$style(type='text/css','.side{font-size: 10px;} .side{color: gray;} .side{font-weight: bold;}')),
tags$head(tags$style(type='text/css','.below{font-size: 10px;} .below{color: gray;} .below{font-weight: bold;}')),
tags$head(tags$style(type='text/css','.capt{font-size: 9px;} .capt{color: gray;} .capt{font-weight: bold;} .capt{margin-top: -10px;}')),
titlePanel('Topic-Covariate Effects'),
fixedRow(
column(2,selectInput('choose_cov', label='Covariate',
choices=cov_cont,selected=cov_cont[[1]]),
fixedRow(column(1,''),
column(11,tags$div(paste0('Choosing a covariate determines which weight estimates will shown',
' The order of the topics will be adjusted accordingly. By clicking',
' an estimate, all figures below will rerender.'),class='side')))),
column(10,plotlyOutput('est',height='200px'))
),
fixedRow(
column(2,''),
column(8,htmlOutput('text')),
column(2,'')
),
plotlyOutput('theta'),
fixedRow(
column(3,
conditionalPanel(condition='output.show_mods',
selectInput('choose_mod',label='Group',
choices=mod_fact,selected=mod_fact[[1]]))),
column(6,
conditionalPanel(condition='output.show',
sliderInput('lambda',label=strong('Lambda'),min=0,max=1,value=1,step=lambda_step))),
column(3,
conditionalPanel(condition='output.show',
checkboxGroupInput('z',
label='',
c('Ignore zeros'=1,'Normalize'=2,'Split'=3),
selected=c(1,2))))
),
fixedRow(
column(3,conditionalPanel(condition='output.show_mods',
tags$div(paste0('Choice of factor to split scatter plots.'),class='capt'))),
column(6,conditionalPanel(condition='output.show',
tags$div(paste0('Relative weighting that influences taxa shown in the scatterplots',
' If equal to 1, p(taxa|topic)l if 0, p(taxa|topic)/p(taxa).'),class='capt'))),
column(3,conditionalPanel(condition='output.show',
tags$div(paste0('Adjust scatter plots by ignoring zeros for smoothing, normalizing to relative abundances,',
' or splitting the plots as a function of the selected factor (if present).'),class='capt')))
),
plotlyOutput('ss'),
plotlyOutput('full')
),
server = function(input,output,session){
current_mods <- reactiveValues(mods=mod_fact)
observeEvent(input$choose_cov,{
covariate <- input$choose_cov
mods <- names(topic_effects[[covariate]]$fitted_switch)
current_mods$mods <- mods
updateNumericInput(session,'k_in',value=mods)
})
EST <- reactive({
suppressWarnings({
covariate <- input$choose_cov
df0 <- data.frame(p=c(0,max(theta)),
covariate=range(metadata[,covariate]),
stringsAsFactors=TRUE)
est_mat <- topic_effects[[covariate]]$est
df <- data.frame(topic=paste0('T',1:K),
est=est_mat[,1],
lower=est_mat[,2],
upper=est_mat[,3],
sig=ifelse(1:K %in% topic_effects[[covariate]]$sig,'1','0'),
stringsAsFactors=TRUE)[order(topic_effects[[covariate]]$rank),]
df$sig <- factor(as.character(sign(df$est) * as.numeric(as.character.factor(df$sig))),levels=c('0','1','-1'),ordered=TRUE)
df$topic <- factor(df$topic,levels=df$topic,ordered=TRUE)
p_est <- ggplot(df,aes_(~topic,y=~est,ymin=~lower,ymax=~upper,color=~sig)) +
geom_hline(yintercept=0,linetype=3)
p_est <- p_est +
geom_pointrange(size=2) +
theme_minimal() +
labs(x='',y='Estimate') +
scale_color_manual(values=c('gray','indianred3','dodgerblue3'),drop=FALSE) +
scale_fill_brewer(type='qual',palette=6,direction=-1) +
theme(legend.position='none',
axis.text.x=element_text(angle=-90,hjust=0,vjust=.5))
list(p_est=p_est,k_levels=levels(df$topic),df0=df0,covariate=covariate)
})
})
output$est <- renderPlotly({
ggplotly(EST()$p_est,source='reg_est')
})
TAB <- reactive({
if (any(input$z %in% '2')) x <- otu_table/rowSums(otu_table) else x <- otu_table
df_temp <- EST()$df0
df_temp$abundance <- c(min(x),max(x))
list(table=x,df=df_temp)
})
P <- reactive({
suppressWarnings({
s <- event_data('plotly_click',source='reg_est')
validate(need(!is.null(s),''))
current_k <- EST()$k_levels[s[['x']]]
l <- input$lambda
tinfo_k <- tinfo[tinfo$Category == current_k,] #subset(tinfo,Category == current_k)
rel_k <- l*tinfo_k$logprob + (1-l)*tinfo_k$loglift
new_order <- tinfo_k[order(rel_k,decreasing=TRUE)[1:taxa_reg_n],]
new_order$Term <- as.character.factor(new_order$Term)
otu_subset <- t(TAB()$table[,new_order$Term])
df_temp <- data.frame(abundance=matrix(otu_subset,ncol=1),
otu=rownames(otu_subset),
sample=rep(colnames(otu_subset),each=nrow(otu_subset)),
stringsAsFactors=FALSE)
df_temp$covariate <- metadata[df_temp$sample,EST()$covariate]
df_temp$switch <- metadata[df_temp$sample,input$choose_mod]
df_temp$p <- theta[df_temp$sample,current_k]
df_temp$taxon <- paste0(pretty_names[df_temp$otu],' (',df_temp$otu,')')
df_temp$taxon <- factor(df_temp$taxon,levels=unique(df_temp$taxon),ordered=TRUE)
p_full <- ggplot(data=TAB()$df,aes_(~covariate,~abundance,color=~p)) +
geom_blank() +
geom_point(data=df_temp,aes_(~covariate,~abundance,color=~p),alpha=.5,size=1.2) +
scale_y_log10(labels=scales::comma) +
viridis::scale_color_viridis('Theta Prob.') +
theme_classic() +
theme(aspect.ratio=.5) +
labs(x=EST()$covariate,y='Abundance')
p_ss <- p_full +
facet_wrap(~taxon,ncol=4) +
theme(aspect.ratio=.5,strip.text.x=element_text(size=7)) +
xlab('')
list(df=df_temp,min=min(df_temp$abundance),p_ss=p_ss,p_full=p_full)
})
})
observeEvent(event_data('plotly_click',source='reg_est'),
{
output$text <- renderUI({
HTML(sprintf("The scatterplot below shows the sample over topic distribution as a function of %s.
If additional factors were present in the model formula,
then their posterior predictive estimates will be shown as a function of the given
factor being set to 1 and 0 with all other covariates held fixed (i.e., averaged across samples).
These factors can be cycled through via the selection box. Each point represents a sample's probability of
containing the selected topic.
The next set of scatterplots show the top %s highest probability taxa in the selected topic -- that is,
p(taxa|k). Each point is the taxa's abundance in the raw data for a given sample, shaded based on the probability of that
sample occuring in the chosen topic. By adjusting lambda, the top %s taxa in terms of p(taxa|k)/p(taxa) will be shown.
The large scatter plot in the bottom row shows all %s taxa combined.",
EST()$covariate,
taxa_reg_n,
taxa_reg_n,
taxa_reg_n))
})
}
)
output$ss <- renderPlotly({
suppressWarnings({
if (any(input$z %in% '1')){
df2 <- P()$df[P()$df$abundance > round(P()$min,5),]
p_ss <- P()$p_ss +
stat_smooth(data=df2,aes_(~covariate,~abundance),color='black',size=1.1,method='loess',se=FALSE) +
stat_smooth(data=df2,aes_(~covariate,~abundance),color='darkred',size=1.1,method='lm',linetype=2,se=FALSE)
}else{
p_ss <- P()$p_ss +
stat_smooth(data=P()$df,aes_(~covariate,~abundance),color='black',size=1.1,method='loess',se=FALSE) +
stat_smooth(data=P()$df,aes_(~covariate,~abundance),color='darkred',size=1.1,method='lm',linetype=2,se=FALSE)
}
if (any(input$z %in% '3') & input$choose_mod != '')
p_ss <- p_ss + facet_wrap(switch~taxon,ncol=6) + theme(strip.text.x=element_text(size=7,margin=margin(t=15,b=15)))
ggplotly(p_ss)
})
})
output$theta <- renderPlotly({
suppressWarnings({
s <- event_data('plotly_click',source='reg_est')
validate(need(!is.null(s),'Please choose a topic by clicking a point.'))
k <- EST()$k_levels[s[['x']]]
if (input$choose_mod != ''){
topic_fitted <- topic_effects[[EST()$covariate]]$fitted[[input$choose_mod]][[k]]
topic_switch <- topic_effects[[EST()$covariate]]$fitted_switch[[input$choose_mod]][[k]]
df_th <- data.frame(Value=theta[rownames(metadata),k],Covariate=metadata[,EST()$covariate],stringsAsFactors=TRUE)
df_fit <- data.frame(topic_fitted,stringsAsFactors=TRUE)
df_switch <- data.frame(topic_switch,stringsAsFactors=TRUE)
colnames(df_fit) <- c('Value','Lower','Upper','Covariate')
colnames(df_switch) <- c('Value','Lower','Upper','Covariate')
R <- range(c(df_th$Covariate,df_fit$covariate,df_switch$covariate))
}else{
topic_fitted <- topic_effects[[EST()$covariate]]$fitted[[1]][[k]]
df_th <- data.frame(Value=theta[rownames(metadata),k],Covariate=metadata[,EST()$covariate],stringsAsFactors=TRUE)
df_fit <- data.frame(topic_fitted,stringsAsFactors=TRUE)
colnames(df_fit) <- c('Value','Lower','Upper','Covariate')
R <- range(c(df_th$Covariate,df_fit$covariate))
}
p1 <- plot_ly(df_th,x=df_th$Covariate)
p1 <- add_trace(p1,
x=~df_th$Covariate,y=~df_th$Value,
type='scatter',mode='markers',
marker=list(symbol='circle',opacity=.5,sizemode='diameter',size=10,color=I('gray')),
showlegend=FALSE)
p1 <- add_lines(p1,
x=~df_fit$Covariate,y=~df_fit$Lower,
type='scatter',mode='lines',
line=list(color='transparent'),
showlegend=FALSE,
name=input$choose_mod)
p1 <- add_lines(p1,
x=~df_fit$Covariate,y=~df_fit$Upper,
type='scatter',mode='lines',
fill='tonexty',
fillcolor='rgba(205,85,85,0.2)',
line=list(color='transparent'),
showlegend=FALSE,
name=input$choose_mod)
p1 <- add_lines(p1,
x=~df_fit$Covariate,y=~df_fit$Value,
type='scatter',mode='lines',
line = list(width=5,color = 'rgba(205,85,85,0.9)'),
showlegend=TRUE,
name=input$choose_mod)
if (input$choose_mod != ''){
p1 <- add_lines(p1,
x=~df_switch$Covariate,y=~df_switch$Lower,
type='scatter',mode='lines',
line=list(color='transparent'),
showlegend=FALSE,
name='Reference/Control')
p1 <- add_lines(p1,
x=~df_switch$Covariate,y=~df_switch$Upper,
type='scatter',mode='lines',
fill='tonexty',
fillcolor='rgba(24,116,205,.2)',
line=list(color='transparent'),
showlegend=FALSE,
name='Reference/Control')
p1 <- add_lines(p1,
x=~df_switch$Covariate,y=~df_switch$Value,
type='scatter',mode='lines',
line=list(width=5,color='rgba(24,116,205,.9)'),
showlegend=TRUE,
name='Reference/Control')
p1 <- layout(p1,title=k,
legend=list(orientation='h'),
paper_bgcolor='rgb(255,255,255)',
xaxis=list(title=sprintf('%s',EST()$covariate),
gridcolor='rgb(255,255,255)',
showgrid=TRUE,
showline=FALSE,
showticklabels=TRUE,
tickcolor='rgb(127,127,127)',
ticks='outside',
zeroline=FALSE),
yaxis=list(title=sprintf('p(topic %s|sample)',k),
gridcolor='rgb(255,255,255)',
showgrid=TRUE,
showline=FALSE,
showticklabels=TRUE,
tickcolor='rgb(127,127,127)',
ticks='outside',
zeroline=FALSE))
}else{
p1 <- layout(p1,title=k,
showlegend=FALSE,
paper_bgcolor='rgb(255,255,255)',
xaxis=list(title=sprintf('%s',EST()$covariate),
gridcolor='rgb(255,255,255)',
showgrid=TRUE,
showline=FALSE,
showticklabels=TRUE,
tickcolor='rgb(127,127,127)',
ticks='outside',
zeroline=FALSE),
yaxis=list(title=sprintf('p(topic %s|sample)',k),
gridcolor='rgb(255,255,255)',
showgrid=TRUE,
showline=FALSE,
showticklabels=TRUE,
tickcolor='rgb(127,127,127)',
ticks='outside',
zeroline=FALSE))
}
p1
})
})
output$show_mods <- reactive({
s <- event_data('plotly_click',source='reg_est')
!is.null(s) & !is.null(mod_fact)
})
outputOptions(output,'show_mods',suspendWhenHidden=FALSE)
output$show <- reactive({
s <- event_data('plotly_click',source='reg_est')
!is.null(s)
})
outputOptions(output,'show',suspendWhenHidden=FALSE)
output$full <- renderPlotly({
suppressWarnings({
if (any(input$z %in% '1')){
df2 <- P()$df[P()$df$abundance > round(P()$min,5),]
p_full <- P()$p_full +
stat_smooth(data=df2,aes_(~covariate,~abundance),color='black',size=1.1,method='loess',se=FALSE) +
stat_smooth(data=df2,aes_(~covariate,~abundance),color='darkred',size=1.1,method='lm',linetype=2,se=FALSE)
}else{
p_full <- P()$p_full +
stat_smooth(data=P()$df,aes_(~covariate,~abundance),color='black',size=1.1,method='loess',se=FALSE) +
stat_smooth(data=P()$df,aes_(~covariate,~abundance),color='darkred',size=1.1,method='lm',linetype=2,se=FALSE)
}
if (any(input$z %in% '3') & input$choose_mod != '')
p_full <- p_full + facet_wrap(~switch,ncol=1)
ggplotly(p_full)
})
})
}
)
}
EESI/themetagenomics documentation built on May 10, 2020, 1:40 a.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Defines functions vis.continuous
Documented in vis.continuous
#' @rdname vis
#'
#' @param taxa_reg_n Number of most relevant taxa within topic to regress. Defaults to 8.
#'
#' @export
vis.continuous <- function(object,lambda_step=.1,taxa_reg_n=8,...){
topics <- object$topics
topic_effects <- object$topic_effects
otu_table <- object$topics$otu_table
tax_table <- object$topics$tax_table
metadata <- object$modelframe
covariates <- colnames(metadata)
cov_f <- sapply(metadata,class) == 'factor'
cov_cont <- covariates[!cov_f]
if (length(cov_cont) == 0) stop('For continuous, formula provided must have contained a continuous, numeric, or integer covariate.')
cov_fact <- covariates[cov_f]
names(cov_cont) <- tolower(unlist(cov_cont))
mod_fact <- names(topic_effects[[cov_cont[1]]]$fitted_switch)
pretty_names <- pretty_taxa_names(tax_table)
otu_table <- otu_table + 1
fit <- topics$fit
vocab <- fit$vocab
K <- fit$settings$dim$K
doc_lengths <- sapply(topics$docs,function(x) sum(x[2,]))
theta <- fit$theta
rownames(theta) <- names(topics$docs)
colnames(theta) <- paste0('T',1:K)
if (min(theta) == 0){
min_theta <- min(theta[theta > 0])
theta <- theta + min_theta
theta <- theta/rowSums(theta)
}
beta <- exp(fit$beta$logbeta[[1]])
colnames(beta) <- vocab
rownames(beta) <- paste0('T',1:K)
if (min(beta) == 0){
min_beta <- min(beta[beta > 0])
beta <- beta + min_beta
beta <- beta/rowSums(beta)
}
topic_freq <- colSums(theta*doc_lengths)
topic_marg <- topic_freq/sum(topic_freq)
term_topic_freq <- beta*topic_freq
# compute term frequencies as column sums of term.topic.frequency
# we actually won't use the user-supplied term.frequency vector.
# the term frequencies won't match the user-supplied frequencies exactly
# this is a work-around to solve the bug described in Issue #32 on github:
# https://github.com/cpsievert/LDAvis/issues/32
term_freq <- colSums(term_topic_freq) # topic_freq <- colSums(theta*doc_lengths)
# marginal distribution over terms (width of blue bars)
term_marg <- term_freq/sum(term_freq)
beta <- t(beta)
# compute the distinctiveness and saliency of the terms:
# this determines the R terms that are displayed when no topic is selected
topic_g_term <- beta/rowSums(beta) # (W x K)
kernel <- topic_g_term*log(t(t(topic_g_term)/topic_marg))
distinctiveness <- rowSums(kernel)
saliency <- term_marg*distinctiveness
# Order the terms for the "default" view by decreasing saliency:
default_terms <- vocab[order(saliency,decreasing=TRUE)][1:taxa_reg_n]
counts <- as.integer(term_freq[match(default_terms,vocab)])
taxa_n_rev <- rev(seq_len(taxa_reg_n))
default <- data.frame(Term=default_terms,
logprob=taxa_n_rev,
loglift=taxa_n_rev,
Freq=counts,
Total=counts,
Category='Default',
stringsAsFactors=FALSE)
default$Term <- factor(default$Term,levels=rev(default$Term),ordered=TRUE)
topic_seq <- rep(seq_len(K),each=taxa_reg_n)
category <- paste0('T',topic_seq)
lift <- beta/term_marg
# Collect taxa_reg_n most relevant terms for each topic/lambda combination
# Note that relevance is re-computed in the browser, so we only need
# to send each possible term/topic combination to the browser
find_relevance <- function(i){
relevance <- i*log(beta) + (1 - i)*log(lift)
idx <- apply(relevance,2,function(x) order(x,decreasing=TRUE)[seq_len(taxa_reg_n)])
indices <- cbind(c(idx),topic_seq)
data.frame(Term=vocab[idx],
Category=category,
logprob=round(log(beta[indices]),4),
loglift=round(log(lift[indices]),4),
stringsAsFactors=FALSE)
}
lambda_seq <- seq(0,1,by=lambda_step) # 0.01
tinfo <- lapply(as.list(lambda_seq),find_relevance)
tinfo <- unique(do.call('rbind', tinfo))
tinfo$Total <- term_freq[match(tinfo$Term,vocab)]
rownames(term_topic_freq) <- paste0('T',seq_len(K))
colnames(term_topic_freq) <- vocab
tinfo$Freq <- term_topic_freq[as.matrix(tinfo[c('Category','Term')])]
tinfo <- rbind(default[,-7],tinfo)
new_order <- default
shinyApp(
ui <- fluidPage(
tags$head(tags$style(type='text/css','.side{font-size: 10px;} .side{color: gray;} .side{font-weight: bold;}')),
tags$head(tags$style(type='text/css','.below{font-size: 10px;} .below{color: gray;} .below{font-weight: bold;}')),
tags$head(tags$style(type='text/css','.capt{font-size: 9px;} .capt{color: gray;} .capt{font-weight: bold;} .capt{margin-top: -10px;}')),
titlePanel('Topic-Covariate Effects'),
fixedRow(
column(2,selectInput('choose_cov', label='Covariate',
choices=cov_cont,selected=cov_cont[[1]]),
fixedRow(column(1,''),
column(11,tags$div(paste0('Choosing a covariate determines which weight estimates will shown',
' The order of the topics will be adjusted accordingly. By clicking',
' an estimate, all figures below will rerender.'),class='side')))),
column(10,plotlyOutput('est',height='200px'))
),
fixedRow(
column(2,''),
column(8,htmlOutput('text')),
column(2,'')
),
plotlyOutput('theta'),
fixedRow(
column(3,
conditionalPanel(condition='output.show_mods',
selectInput('choose_mod',label='Group',
choices=mod_fact,selected=mod_fact[[1]]))),
column(6,
conditionalPanel(condition='output.show',
sliderInput('lambda',label=strong('Lambda'),min=0,max=1,value=1,step=lambda_step))),
column(3,
conditionalPanel(condition='output.show',
checkboxGroupInput('z',
label='',
c('Ignore zeros'=1,'Normalize'=2,'Split'=3),
selected=c(1,2))))
),
fixedRow(
column(3,conditionalPanel(condition='output.show_mods',
tags$div(paste0('Choice of factor to split scatter plots.'),class='capt'))),
column(6,conditionalPanel(condition='output.show',
tags$div(paste0('Relative weighting that influences taxa shown in the scatterplots',
' If equal to 1, p(taxa|topic)l if 0, p(taxa|topic)/p(taxa).'),class='capt'))),
column(3,conditionalPanel(condition='output.show',
tags$div(paste0('Adjust scatter plots by ignoring zeros for smoothing, normalizing to relative abundances,',
' or splitting the plots as a function of the selected factor (if present).'),class='capt')))
),
plotlyOutput('ss'),
plotlyOutput('full')
),
server = function(input,output,session){
current_mods <- reactiveValues(mods=mod_fact)
observeEvent(input$choose_cov,{
covariate <- input$choose_cov
mods <- names(topic_effects[[covariate]]$fitted_switch)
current_mods$mods <- mods
updateNumericInput(session,'k_in',value=mods)
})
EST <- reactive({
suppressWarnings({
covariate <- input$choose_cov
df0 <- data.frame(p=c(0,max(theta)),
covariate=range(metadata[,covariate]),
stringsAsFactors=TRUE)
est_mat <- topic_effects[[covariate]]$est
df <- data.frame(topic=paste0('T',1:K),
est=est_mat[,1],
lower=est_mat[,2],
upper=est_mat[,3],
sig=ifelse(1:K %in% topic_effects[[covariate]]$sig,'1','0'),
stringsAsFactors=TRUE)[order(topic_effects[[covariate]]$rank),]
df$sig <- factor(as.character(sign(df$est) * as.numeric(as.character.factor(df$sig))),levels=c('0','1','-1'),ordered=TRUE)
df$topic <- factor(df$topic,levels=df$topic,ordered=TRUE)
p_est <- ggplot(df,aes_(~topic,y=~est,ymin=~lower,ymax=~upper,color=~sig)) +
geom_hline(yintercept=0,linetype=3)
p_est <- p_est +
geom_pointrange(size=2) +
theme_minimal() +
labs(x='',y='Estimate') +
scale_color_manual(values=c('gray','indianred3','dodgerblue3'),drop=FALSE) +
scale_fill_brewer(type='qual',palette=6,direction=-1) +
theme(legend.position='none',
axis.text.x=element_text(angle=-90,hjust=0,vjust=.5))
list(p_est=p_est,k_levels=levels(df$topic),df0=df0,covariate=covariate)
})
})
output$est <- renderPlotly({
ggplotly(EST()$p_est,source='reg_est')
})
TAB <- reactive({
if (any(input$z %in% '2')) x <- otu_table/rowSums(otu_table) else x <- otu_table
df_temp <- EST()$df0
df_temp$abundance <- c(min(x),max(x))
list(table=x,df=df_temp)
})
P <- reactive({
suppressWarnings({
s <- event_data('plotly_click',source='reg_est')
validate(need(!is.null(s),''))
current_k <- EST()$k_levels[s[['x']]]
l <- input$lambda
tinfo_k <- tinfo[tinfo$Category == current_k,] #subset(tinfo,Category == current_k)
rel_k <- l*tinfo_k$logprob + (1-l)*tinfo_k$loglift
new_order <- tinfo_k[order(rel_k,decreasing=TRUE)[1:taxa_reg_n],]
new_order$Term <- as.character.factor(new_order$Term)
otu_subset <- t(TAB()$table[,new_order$Term])
df_temp <- data.frame(abundance=matrix(otu_subset,ncol=1),
otu=rownames(otu_subset),
sample=rep(colnames(otu_subset),each=nrow(otu_subset)),
stringsAsFactors=FALSE)
df_temp$covariate <- metadata[df_temp$sample,EST()$covariate]
df_temp$switch <- metadata[df_temp$sample,input$choose_mod]
df_temp$p <- theta[df_temp$sample,current_k]
df_temp$taxon <- paste0(pretty_names[df_temp$otu],' (',df_temp$otu,')')
df_temp$taxon <- factor(df_temp$taxon,levels=unique(df_temp$taxon),ordered=TRUE)
p_full <- ggplot(data=TAB()$df,aes_(~covariate,~abundance,color=~p)) +
geom_blank() +
geom_point(data=df_temp,aes_(~covariate,~abundance,color=~p),alpha=.5,size=1.2) +
scale_y_log10(labels=scales::comma) +
viridis::scale_color_viridis('Theta Prob.') +
theme_classic() +
theme(aspect.ratio=.5) +
labs(x=EST()$covariate,y='Abundance')
p_ss <- p_full +
facet_wrap(~taxon,ncol=4) +
theme(aspect.ratio=.5,strip.text.x=element_text(size=7)) +
xlab('')
list(df=df_temp,min=min(df_temp$abundance),p_ss=p_ss,p_full=p_full)
})
})
observeEvent(event_data('plotly_click',source='reg_est'),
{
output$text <- renderUI({
HTML(sprintf("The scatterplot below shows the sample over topic distribution as a function of %s.
If additional factors were present in the model formula,
then their posterior predictive estimates will be shown as a function of the given
factor being set to 1 and 0 with all other covariates held fixed (i.e., averaged across samples).
These factors can be cycled through via the selection box. Each point represents a sample's probability of
containing the selected topic.
The next set of scatterplots show the top %s highest probability taxa in the selected topic -- that is,
p(taxa|k). Each point is the taxa's abundance in the raw data for a given sample, shaded based on the probability of that
sample occuring in the chosen topic. By adjusting lambda, the top %s taxa in terms of p(taxa|k)/p(taxa) will be shown.
The large scatter plot in the bottom row shows all %s taxa combined.",
EST()$covariate,
taxa_reg_n,
taxa_reg_n,
taxa_reg_n))
})
}
)
output$ss <- renderPlotly({
suppressWarnings({
if (any(input$z %in% '1')){
df2 <- P()$df[P()$df$abundance > round(P()$min,5),]
p_ss <- P()$p_ss +
stat_smooth(data=df2,aes_(~covariate,~abundance),color='black',size=1.1,method='loess',se=FALSE) +
stat_smooth(data=df2,aes_(~covariate,~abundance),color='darkred',size=1.1,method='lm',linetype=2,se=FALSE)
}else{
p_ss <- P()$p_ss +
stat_smooth(data=P()$df,aes_(~covariate,~abundance),color='black',size=1.1,method='loess',se=FALSE) +
stat_smooth(data=P()$df,aes_(~covariate,~abundance),color='darkred',size=1.1,method='lm',linetype=2,se=FALSE)
}
if (any(input$z %in% '3') & input$choose_mod != '')
p_ss <- p_ss + facet_wrap(switch~taxon,ncol=6) + theme(strip.text.x=element_text(size=7,margin=margin(t=15,b=15)))
ggplotly(p_ss)
})
})
output$theta <- renderPlotly({
suppressWarnings({
s <- event_data('plotly_click',source='reg_est')
validate(need(!is.null(s),'Please choose a topic by clicking a point.'))
k <- EST()$k_levels[s[['x']]]
if (input$choose_mod != ''){
topic_fitted <- topic_effects[[EST()$covariate]]$fitted[[input$choose_mod]][[k]]
topic_switch <- topic_effects[[EST()$covariate]]$fitted_switch[[input$choose_mod]][[k]]
df_th <- data.frame(Value=theta[rownames(metadata),k],Covariate=metadata[,EST()$covariate],stringsAsFactors=TRUE)
df_fit <- data.frame(topic_fitted,stringsAsFactors=TRUE)
df_switch <- data.frame(topic_switch,stringsAsFactors=TRUE)
colnames(df_fit) <- c('Value','Lower','Upper','Covariate')
colnames(df_switch) <- c('Value','Lower','Upper','Covariate')
R <- range(c(df_th$Covariate,df_fit$covariate,df_switch$covariate))
}else{
topic_fitted <- topic_effects[[EST()$covariate]]$fitted[[1]][[k]]
df_th <- data.frame(Value=theta[rownames(metadata),k],Covariate=metadata[,EST()$covariate],stringsAsFactors=TRUE)
df_fit <- data.frame(topic_fitted,stringsAsFactors=TRUE)
colnames(df_fit) <- c('Value','Lower','Upper','Covariate')
R <- range(c(df_th$Covariate,df_fit$covariate))
}
p1 <- plot_ly(df_th,x=df_th$Covariate)
p1 <- add_trace(p1,
x=~df_th$Covariate,y=~df_th$Value,
type='scatter',mode='markers',
marker=list(symbol='circle',opacity=.5,sizemode='diameter',size=10,color=I('gray')),
showlegend=FALSE)
p1 <- add_lines(p1,
x=~df_fit$Covariate,y=~df_fit$Lower,
type='scatter',mode='lines',
line=list(color='transparent'),
showlegend=FALSE,
name=input$choose_mod)
p1 <- add_lines(p1,
x=~df_fit$Covariate,y=~df_fit$Upper,
type='scatter',mode='lines',
fill='tonexty',
fillcolor='rgba(205,85,85,0.2)',
line=list(color='transparent'),
showlegend=FALSE,
name=input$choose_mod)
p1 <- add_lines(p1,
x=~df_fit$Covariate,y=~df_fit$Value,
type='scatter',mode='lines',
line = list(width=5,color = 'rgba(205,85,85,0.9)'),
showlegend=TRUE,
name=input$choose_mod)
if (input$choose_mod != ''){
p1 <- add_lines(p1,
x=~df_switch$Covariate,y=~df_switch$Lower,
type='scatter',mode='lines',
line=list(color='transparent'),
showlegend=FALSE,
name='Reference/Control')
p1 <- add_lines(p1,
x=~df_switch$Covariate,y=~df_switch$Upper,
type='scatter',mode='lines',
fill='tonexty',
fillcolor='rgba(24,116,205,.2)',
line=list(color='transparent'),
showlegend=FALSE,
name='Reference/Control')
p1 <- add_lines(p1,
x=~df_switch$Covariate,y=~df_switch$Value,
type='scatter',mode='lines',
line=list(width=5,color='rgba(24,116,205,.9)'),
showlegend=TRUE,
name='Reference/Control')
p1 <- layout(p1,title=k,
legend=list(orientation='h'),
paper_bgcolor='rgb(255,255,255)',
xaxis=list(title=sprintf('%s',EST()$covariate),
gridcolor='rgb(255,255,255)',
showgrid=TRUE,
showline=FALSE,
showticklabels=TRUE,
tickcolor='rgb(127,127,127)',
ticks='outside',
zeroline=FALSE),
yaxis=list(title=sprintf('p(topic %s|sample)',k),
gridcolor='rgb(255,255,255)',
showgrid=TRUE,
showline=FALSE,
showticklabels=TRUE,
tickcolor='rgb(127,127,127)',
ticks='outside',
zeroline=FALSE))
}else{
p1 <- layout(p1,title=k,
showlegend=FALSE,
paper_bgcolor='rgb(255,255,255)',
xaxis=list(title=sprintf('%s',EST()$covariate),
gridcolor='rgb(255,255,255)',
showgrid=TRUE,
showline=FALSE,
showticklabels=TRUE,
tickcolor='rgb(127,127,127)',
ticks='outside',
zeroline=FALSE),
yaxis=list(title=sprintf('p(topic %s|sample)',k),
gridcolor='rgb(255,255,255)',
showgrid=TRUE,
showline=FALSE,
showticklabels=TRUE,
tickcolor='rgb(127,127,127)',
ticks='outside',
zeroline=FALSE))
}
p1
})
})
output$show_mods <- reactive({
s <- event_data('plotly_click',source='reg_est')
!is.null(s) & !is.null(mod_fact)
})
outputOptions(output,'show_mods',suspendWhenHidden=FALSE)
output$show <- reactive({
s <- event_data('plotly_click',source='reg_est')
!is.null(s)
})
outputOptions(output,'show',suspendWhenHidden=FALSE)
output$full <- renderPlotly({
suppressWarnings({
if (any(input$z %in% '1')){
df2 <- P()$df[P()$df$abundance > round(P()$min,5),]
p_full <- P()$p_full +
stat_smooth(data=df2,aes_(~covariate,~abundance),color='black',size=1.1,method='loess',se=FALSE) +
stat_smooth(data=df2,aes_(~covariate,~abundance),color='darkred',size=1.1,method='lm',linetype=2,se=FALSE)
}else{
p_full <- P()$p_full +
stat_smooth(data=P()$df,aes_(~covariate,~abundance),color='black',size=1.1,method='loess',se=FALSE) +
stat_smooth(data=P()$df,aes_(~covariate,~abundance),color='darkred',size=1.1,method='lm',linetype=2,se=FALSE)
}
if (any(input$z %in% '3') & input$choose_mod != '')
p_full <- p_full + facet_wrap(~switch,ncol=1)
ggplotly(p_full)
})
})
}
)
}
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.