Incorporado el equilibrio entre sexos.

invivos/app.R

@@ -88,6 +88,7 @@ server <- function(input, output) {
   dades<-reactiveValues()
   dades$taula<-NULL
   dades$groups<-NULL
+  dades$sex<-NULL
   dades$db<-NULL
   observe({
     if (!is.null(input$file_sizes)){ 
@@ -103,6 +104,7 @@ server <- function(input, output) {
       }
       dades$taula<-taula
       dades$groups<-read.xlsx(input$file_sizes$datapath, sheet = 2)[,1]
+      dades$sex<-read.xlsx(input$file_sizes$datapath, sheet = 3, sep.names = " ")
     }
   })
   output$firstPlot <- renderPlot({
@@ -148,56 +150,76 @@ server <- function(input, output) {
     low_cuttof<-input$lowcut
     print(up_cuttof)
     df<-df[df$Volume < up_cuttof & df$Volume >= low_cuttof,]
-    
+    df<-merge(df, dades$sex)
     
     # df["Mouse"]<-gsub("[a-zA-Z]", "", df$MouseID)
     print(df$Volume)
     s<-shapiro.test(df$Volume)[[2]]
     
     ngroup<-length(dades$groups)
-    ind.list<-list()
-    pval.list<-list()
-    lvn.list<-list()
-    test.list<-list()
-    for (data in 1:input$iterations){
-      interr=T
-      while(interr == T){
-        ind<-sample(rep(dades$groups, each=ceiling(length(unique(df$`ID animal`))/ngroup)), length(unique(df$`ID animal`)))
-        df_temp<-merge(df[,c("ID animal", "ID tumor","Volume")], data.frame("ID animal"=unique(df$`ID animal`), "group"=as.factor(ind),check.names=F))
-        if ((nrow(df_temp)/ngroup) %% 2 == 0){
-          interr<-any(table(df_temp$group) < floor(nrow(df_temp)/ngroup) | table(df_temp$group) > ceiling(nrow(df_temp)/ngroup))  
+    df_def<-list()
+    print(head(df))
+    for (sex.var in c("male","female")){
+      print(sex.var)
+      df_sex<-df %>% filter(`sex` == sex.var)
+      ind.list<-list()
+      pval.list<-list()
+      lvn.list<-list()
+      test.list<-list()
+      for (data in 1:input$iterations){
+        interr=T
+        while(interr == T){
+          ind<-sample(rep(dades$groups, each=ceiling(length(unique(df_sex$`ID animal`))/ngroup)), length(unique(df_sex$`ID animal`)))
+          df_sex<-merge(df_sex[,c("ID animal", "ID tumor","Volume")], data.frame("ID animal"=unique(df_sex$`ID animal`), "group"=as.factor(ind),check.names=F))
+          if ((nrow(df_sex)/ngroup) %% 2 == 0){
+            interr<-any(table(df_sex$group) < floor(nrow(df_sex)/ngroup) | table(df_sex$group) > ceiling(nrow(df_sex)/ngroup))  
+          }else{
+            # interr<-any(table(df_sex$group) < (floor(nrow(df_sex)/ngroup)-1) | table(df_sex$group) > (ceiling(nrow(df_sex)/ngroup)+1))
+            interr<-diff(range(table(ind))) > 1
+          }
+        }
+        ind.list[[data]]<-df_sex[,c("ID animal","ID tumor","group","Volume")]
+        lvn.list[data]<-leveneTest(Volume ~ group, data = df_sex[,3:4])[[2]][1]
+        if (s < 0.05){
+          k<-kruskal.test(df_sex$Volume,df_sex$group)
+          test.list[data]<-k[[1]][1]
+          pval.list[data]<-k[[3]][1]
         }else{
-          # interr<-any(table(df_temp$group) < (floor(nrow(df_temp)/ngroup)-1) | table(df_temp$group) > (ceiling(nrow(df_temp)/ngroup)+1))
-          interr<-diff(range(table(ind))) > 1
+          res.aov<-aov(Volume~group, data=df_sex)
+          pval.list[data]<-summary(res.aov)[[1]][[5]][1]
+          test.list[data]<-summary(res.aov)[[1]][[4]][1]
         }
       }
-      ind.list[[data]]<-df_temp[,c("ID animal", "ID tumor","group","Volume")]
-      lvn.list[data]<-leveneTest(Volume ~ group, data = df_temp[,3:4])[[2]][1]
-      if (s < 0.05){
-        k<-kruskal.test(df_temp$Volume,df_temp$group)
-        test.list[data]<-k[[1]][1]
-        pval.list[data]<-k[[3]][1]
-      }else{
-        res.aov<-aov(Volume~group, data=df_temp)
-        pval.list[data]<-summary(res.aov)[[1]][[5]][1]
-        test.list[data]<-summary(res.aov)[[1]][[4]][1]
-      }
+      index<-which(unlist(lvn.list) == min(unlist(lvn.list)[which(unlist(pval.list) %in% sort(unlist(pval.list), decreasing = T)[1:20])]))
+      print(df_sex)
+      df_def[[sex.var]]<-merge(df_sex %>% select(-group), ind.list[[index]])
     }
-    index<-which(unlist(lvn.list) == min(unlist(lvn.list)[which(unlist(pval.list) %in% sort(unlist(pval.list), decreasing = T)[1:20])]))
-    df_def<-merge(df, ind.list[[index]])
+    # df_def<-do.call(rbind, c(df_def, make.row.names=F))
+    # lapply(df_def, function(x) x %>% as_tibble %>% print(n=Inf))
+    df_def<-rbind(df_def[[1]], df_def[[2]], make.row.names=F)
     if ("Group" %in% colnames(df_def)){
       df_def<-df_def %>% select(-"Group") 
     }
-    df_def<-merge(dades$taula %>% select(-Group), df_def[,c("ID animal", "group")] %>% unique, all=T, by="ID animal") %>% select(c(`ID animal`, `ID tumor`, Volume, Cage, Major, Minor, group))
+    df_def<-merge(merge(dades$taula, dades$sex) %>% select(-Group), df_def[,c("ID animal", "group")] %>% unique, all=T, by="ID animal") %>% select(c(`ID animal`, `sex`,`ID tumor`, Volume, Cage, Major, Minor, group))
     df_def[!paste0(df_def$`ID animal`, df_def$`ID tumor`) %in% paste0(df$`ID animal`, df$`ID tumor`),"group"]<-NA
     dades$db<-df_def
 
-    
+    ggarrange(
     ggplot(df_def, aes(group, Volume))+
       geom_boxplot(outlier.alpha = F)+
-      geom_jitter(width=0.25)+
-      geom_point(stat="summary", color="blue", size=3)
+      geom_jitter(width=0.25, aes(color=sex))+
+      geom_point(stat="summary", color="blue", size=3)+
+      theme(axis.text.x = element_text(angle=90, hjust=1, vjust=0.5)),
       # lims(y=c(0,max(df_def$Volume)+10))
+    ggarrange(
+      ggplot(df_def, aes(sex, Volume))+
+        geom_boxplot(outlier.alpha = F)+
+        geom_quasirandom(width=0.3),
+      ggplot(df_def, aes(group, fill=sex))+
+        geom_bar(stat="count", color="black", position="dodge")+
+        guides(fill="none")+
+        theme(axis.text.x = element_text(angle=90, hjust=1, vjust=0.5)), ncol = 1, heights = c(0.35, 0.65)),
+    nrow = 1, aligh="h", widths = c(0.65, 0.35))
   })
   output$distPlot <- renderPlot({
     observeEvent(dades$taula, {})
@@ -295,7 +317,7 @@ server <- function(input, output) {
   analysis$taula_vol<-NULL
   observe({
     if (!is.null(input$file_analy)){ 
-      analysis$taula<-read.xlsx(input$file_analy$datapath, sheet = 1, check.names = F, sep.names = " ")
+      analysis$taula<-read.xlsx(input$file_analy$datapath, sheet = 1, check.names = F, sep.names = " ") %>% select(-sex)
     }
   })
   output$cutoffUI<-renderUI({