################################## #create output# ##################### pdf(paste(filenameprefix,added,".pdf",sep=""), paper = "a4") layout (matrix(1)) par(cex=0.6, cex.axis=1.2, las=1) Arrlength <- 0 oldpar = par() ############### control same number on each side: graph # data$scorerelative = (data$X8_test2 -data$meanbefore)/10000 data$scorenotrelative = (data$X8_test2 )/10000 plot(data$side_punished ~ data$genotype) # # # # # # data$colorwhennegative=2-(as.numeric(data$colorwhenpositive)-1) # data$colorwhennegative2 = levels(data$colorwhenpositive)[data$colorwhennegative] # data$colorwhenpositive2 = levels(data$colorwhenpositive)[data$colorwhenpositive] # # # data$colorpunished= data$colorwhenpositive2 # data$colorpunished= ifelse(data$side_punished =="positive-left",data$colorwhenpositive2,data$colorwhennegative2) # # data$colorpunished=as.factor(data$colorpunished) # # # plot(data$colorpunished ~ data$genotype) plot (data$genotype, data$meanbefore/10000, main = "preference before training", ylim = c(-1,1)) points (data$genotype, data$meanbefore/10000) abline(h=0, col =2) ############### #for each genotype ######################################################################################## for (j in c(1:length(levels(data$genotype)))){ thesubset = levels (data$genotype)[j] data_ord = subset (data_ordtot, data_ordtot$genotype == thesubset) datanumber= as.matrix (summary (data$genotype) [j]) par(xpd= TRUE, mar=c(5.1,4,4,8)) plot(data_ord$ind,data_ord$values, type ="l", axes = FALSE, main =paste(thesubset, "n=",datanumber ),xlim=c(0,10)) axis(2) axis(1, at=seq(1,7,by=1), labels=c("pretest","training","training","test","training","training","test"), las = 2) abline(h=0, col =2) s <- seq(length(data_ord$ind)-1)# one shorter than data segments (as.numeric(data_ord$ind[s]),data_ord$values[s],as.numeric(data_ord$ind[s+datanumber]),data_ord$values[s+datanumber], col= 1:datanumber) for (i in c(1:datanumber)){ k=length(data_ord$ind)-datanumber +i #print(k) text (9, data_ord$values[k], paste(ifelse(is.na(data_ord$remarks [k]),as.character(data_ord$remarksnotimportant [k]),data_ord$remarks [k]) , row.names(data_ord)[i]),col=i) } par(oldpar) violinplot(data_ord$values~ data_ord$ind, ylim = c(-2,2)) abline(h=0, col =2) title (thesubset) } ################# # genotypes together ######## layout(matrix(c(1,1,1,1), 1, 1, byrow = TRUE)) variables = list (as.factor (data_ordtot$genotype), data_ordtot$ind) LI_mean <- tapply(data_ordtot$values, variables, mean, na.rm= TRUE) LI_sem <- tapply(data_ordtot$values, variables, std.error) samplesize <-as.matrix(summary (data$genotype)) bar <- barplot(LI_mean, beside=T, ylim = c(-0.6,1), main=("nocolorlearning"), axisnames =F, legend.text=( paste(levels(data$genotype), "(",samplesize,")" ) ), args.legend = list(x = "bottomright") ) # store location of bars in bar, and # plot the barplot. ylim is used to # make room for the error bar later arrows(bar, LI_mean + LI_sem, bar, LI_mean - LI_sem, length = Arrlength, # width of the arrowhead angle = 90, # angle of the arrowhead code = 3 # arrowhead in both ends ) axis(1, at=seq(bar [2,1], bar [2,7],by= (length(levels(data_ordtot$genotype))+1)), tick =T, labels=c("pretest","training","training","test","training","training","test"), las = 2) abline(h=seq(0,1,.2),col=2) plot(data$X8_test2/10000~data$genotype, las=2, ylab="learning score", xlab="", main="score absolute during final test") abline(h=0, col=2) violinplot(data$X8_test2/10000~data$genotype, las=2, ylab="learning score", xlab="", main="score absolute during final test") abline(h=0, col=2) plot(data$scorerelative~data$genotype, las=2, ylab="learning score", xlab="", main="score final test - pretest") abline(h=0, col=2) violinplot(data$scorerelative~data$genotype, las=2, ylab="learning score", xlab="", main="score final test - pretest") abline(h=0, col=2) data$scorenotrel_asin = asin(sqrt((data$scorenotrelative+1)/2)) #data$color_asin = asin(sqrt((data$X5_test1/10000 +1)/2)) #change=ddata$scorenotrel -data$Xb_test3/10000 plot(data$scorenotrel_asin~data$genotype, main="score absolute Arcsin transformed during final test", ylab="",ylim=c (asin(sqrt((0))),asin(sqrt((1)))),las=2) abline(h=asin(sqrt((0.5))), col=2) abline(h=asin(sqrt((0)))) abline(h=asin(sqrt((1)))) #plot(data$scorerelative~data$genotype) # abline(h=0) aov = aov(data$scorenotrel_asin~data$genotype) kruskal.test(data$scorenotrelative~data$genotype) #aov2 = aov(data$color_asin~data$genotype) summary(aov) TukeyHSD(aov) #########Textplots ########################## L=length(levels(data$genotype)) layout (matrix(1:L,L,1)) #textplot (data, cmar=0, halign ="left") table=c("anova", summary(aov)) for(i in 1: length (levels(data$genotype))){ Genotypetested= levels(data$genotype)[i] EX=with(data, wilcox.test(data$scorerelative[data$genotype == Genotypetested],alternative = "greater", mu = 0)) #textplot(summary (subset (data,data$genotype == Genotypetested))) #textplot(paste (Genotypetested, "scorerelative wilcox.test, pvalue diff from 0:",EX$p.value,"by",EX$method,EX$alternative)) table= rbind(table,c("wilcox",paste (Genotypetested, "scorerelative wilcox.test, pvalue diff from 0:",EX$p.value,"by",EX$method,EX$alternative))) EX=with(data, wilcox.test(data$scorenotrelative[data$genotype == Genotypetested],alternative = "greater", mu = 0)) #textplot(summary (subset (data,data$genotype == Genotypetested))) #textplot(paste (Genotypetested, "scorerelative wilcox.test, pvalue diff from 0:",EX$p.value,"by",EX$method,EX$alternative)) table= rbind(table,c("wilcox",paste (Genotypetested, "scorenotrelative wilcox.test, pvalue diff from 0:",EX$p.value,"by",EX$method,EX$alternative))) } textplot(table) layout (matrix(1:1,1,1)) ######################if 3 genotypes only # # ############################## # # enlever preference inne########## # ########## # # # # Testvar =data$scorenotrelative # # if (length(levels(data$genotype))==2){ # stat= wilcox.test(Testvar ~data$genotype2 ) # STATLAB=paste("difference: wilcox.test, p=",stat$p.value,",", stat$alternative) # } else{STATLAB= "no statistique more than 3 groups"} # # Title= "score normal" # # LI_mean <- tapply(Testvar, as.factor (data$genotype2), mean, na.rm= TRUE) # LI_sem <- tapply(Testvar, as.factor (data$genotype2), std.error, na.rm= TRUE) # samplesize <-as.matrix(summary (data$genotype2)) # # bar <- barplot(LI_mean, beside=T, ylim = c(-0.5,1), main=(Title), axisnames =T, xlab=STATLAB,legend.text=( paste(levels(data$genotype2), "(",samplesize,")" ) ) ) # store location of bars in bar, and # # plot the barplot. ylim is used to # # make room for the error bar later # # arrows(bar, LI_mean + LI_sem, bar, LI_mean - LI_sem, # length = Arrlength, # width of the arrowhead # angle = 90, # angle of the arrowhead # code = 3 # arrowhead in both ends # ) # # Testvar =data$scorerelative # if (length(levels(data$genotype))==2){ # stat= wilcox.test(Testvar ~data$genotype2 ) # STATLAB=paste("difference: wilcox.test, p=",stat$p.value,",", stat$alternative) # } else{STATLAB= "no statistique more than 3 groups"} # # Title= "score relative to pretest score(substraction)" # # LI_mean <- tapply(Testvar, as.factor (data$genotype2), mean, na.rm= TRUE) # LI_sem <- tapply(Testvar, as.factor (data$genotype2), std.error) # samplesize <-as.matrix(summary (data$genotype2)) # # bar <- barplot(LI_mean, beside=T, ylim = c(-0.5,1), main=(Title), axisnames =T, xlab=STATLAB,legend.text=( paste(levels(data$genotype2), "(",samplesize,")" ) ) ) # store location of bars in bar, and # # plot the barplot. ylim is used to # # make room for the error bar later # # arrows(bar, LI_mean + LI_sem, bar, LI_mean - LI_sem, # length = Arrlength, # width of the arrowhead # angle = 90, # angle of the arrowhead # code = 3 # arrowhead in both ends # ) # # # # plot (data$genotype2, data$scorerelative, main = "self learning score while naive preference taken out") # points (data$genotype2, data$scorerelative) # abline(h=0, col =2) # # # # plot (data$genotype2, data$scorenotrelative, main = "self learning score while naive preference not taken out", las=2) # points (data$genotype2, data$scorenotrelative) # abline(h= 0, col=2) # # # # #table= rbind(table,as.matrix(summary(aov)) # nparcomp(data$scorenotrelative~data$genotype, data=data, type ="Tukey") # a =nparcomp(data$scorenotrelative~data$genotype, data=data, type ="Sequen") # plot(a$Analysis.of.relative.effects$p.value.adjusted~a$Analysis.of.relative.effects$Comparison, ylim=c(0,1), main ="significance of the effects,Nonparametric relative contrast effects ") # abline(h=0.05, col=2) # # # # data2= subset (data, data$genotype != "Gal4 controls") # data2=drop.levels(data2) # wilcox.test(data2$scorerelative~data2$genotype, data=data2) # #plot(data2$scorenotrelative~data2$genotype) # # data2= subset (data, data$genotype != "UAS controls") # data2=drop.levels(data2) # wilcox.test(data2$scorerelative~data2$genotype, data=data2) # #plot(data2$scorenotrelative~data2$genotype) dev.off() violinplot(data$X8_test2/10000~data$genotype, las=2, ylab="learning score", xlab="", main="score absolute during final test") abline(h=0, col=2)