library(openxlsx) library(ggplot2) library(coxme) library(glmmTMB) library(multcomp) library(DHARMa) library(multcompView) #library(ggpubr) library(patchwork) library(ggbeeswarm) library(survminer) #install.packages("viridis") # Install library("viridis") # Load library(RColorBrewer) library(dplyr) setwd("~/Documents/PhD Tesis db/") maternal_data <- read.xlsx("Pop_survey.xlsx",sheet = 2) #Creating the ID for each queen colony maternal_data$Exp_nest <- paste(maternal_data$Code, maternal_data$W, "_", maternal_data$Replicate, sep = "") # The 16 maternal queens are mat16 <- c("A30_10", "A30_17", "A30_2", "A30_29", "A30_7", "A30_9", "C10_1", "C10_13", "C10_30", "C10_9", "E20_16", "E20_17", "E20_2", "E20_27", "F10_10", "F10_6") maternal_data2 <- maternal_data[maternal_data$Exp_nest %in% mat16, ] maternal_data3 <- data.frame (Exp_nest = maternal_data2$Exp_nest, W = maternal_data2$W, Eggs_total = maternal_data2$Eggs, Queen_Ptotal = maternal_data2$Pupae_Q, Worker_Ptotal = maternal_data2$Pupae_W, Males_Ptotal = maternal_data2$Pupae_M, Treatment = as.numeric(maternal_data2$W)) maternal_summarized <- maternal_data3 %>% group_by(Exp_nest) %>% summarise(across(everything(),~ sum(.x, na.rm = T))) # Maternal Lifespan ### data_na <- subset(maternal_data2, Queen_alive==1)# get ride of lines of queens after been sacrified or died # cols <- c("Code", "Replicate") # data_na$queen.id <- do.call(paste, c(data_na[cols], sep="_")) # data_na$queen.id <- as.factor(data_na$queen.id) # f0_names_115q <- c("A_10", "A_17" ,"A_29", "A_7" , "C_1", "C_13" ,"A_2" , # "A_9" , "C_9" , "E_27" ,"F_10" ,"C_30", "E_16", "E_17", "E_2", "F_6" , "A_11" , "A_14" , "A_15" , "A_16" , "A_18" , "A_19" , "A_2" , "A_21" , "A_22" , "A_23" , "A_24" , "A_25" , "A_27" , "A_27b" , "A_28" , "A_29" , "A_3" , "A_4" , "A_5" , "A_6" , "A_6b" , "A_7" , "A_9" , "B_1" , "B_10" , "B_11" , "B_14" , "B_15" , "B_16" , "B_17" , "B_2" , "B_20" , "B_23" , "B_25" , "B_27" , "B_28" , "B_32b" , "B_34" , "B_36" , "B_38" , "B_7" , "B_9" , "C_1" , "C_11" , "C_12" , "C_15" , "C_17" , "C_18" , "C_19" , "C_19b" , "C_2" , "C_20" , "C_21" , "C_22" , "C_25" , "C_26" , "C_28" , "C_3" , "C_4" , "C_5" , "C_6" , "C_7" , "C_8" , "D_1" , "D_11" , "D_14" , "D_16" , "D_17" , "D_18" , "D_2" , "D_3" , "D_7" , "D_8" , "D_9" , "E_1" , "E_10" , "E_13" , "E_14", "E_15" , "E_19" , "E_2" , "E_26" , "E_28" , "E_29" , "E_3" , "E_30" , "E_4" , "E_6" , "E_8" , "F_10" , "F_14" , "F_15" , "F_20" , "F_3" , "F_4" , "F_5" , "F_7" , "F_8" , "F_9" ) # # colonies <- data_na[(data_na$queen.id %in% f0_names_115q), ] max_week <- aggregate(data_na$Week, list(data_na$Exp_nest), max) #but inprefer to have not only the week data but the whole row colnames(max_week) <- c("Exp_nest", "Week") maternal_summarized <- merge(maternal_summarized, max_week, by= "Exp_nest") #write.xlsx(maternal_summarized, "Selection_lines_records_241011.xlsx") setwd("~/Documents/PhD_Tesis/04_selected_lines/") f1_data <- read.xlsx("Selection_lines_records_241011.xlsx",sheet = 1) f0_data <- read.xlsx("Selection_lines_records_241011.xlsx",sheet = 2) batch_data <- read.xlsx("Selection_lines_records_241011.xlsx",sheet = 3) f1_0_data <- merge(f0_data, f1_data, by = "Exp_nest") f1_0_data$casteR <- f1_0_data$QueenP_sel/(f1_0_data$QueenP_sel + f1_0_data$Worker_sel) f1_0_data$casteRf0 <- f1_0_data$Queen_Ptotal /(f1_0_data$Worker_Ptotal + f1_0_data$Queen_Ptotal) summary(f1_0_data) f1_0_data$Recorded_dead <- as.factor(f1_0_data$Recorded_dead) f1_0_data$Treatment <- as.factor(f1_0_data$Treatment) f1_0_data$Productivity <- factor(f1_0_data$Productivity, levels = c("Low", "High")) ##### Figure 1 ##### setwd("~/Dropbox/2020_Review_Social_aging_Jaimes_etal/Data") correlation <- read.xlsx("Egg_estimate_productivity.xlsx",sheet = 1) pvalplot <- ggplot()+ geom_smooth(data = correlation, aes(x=Week, y=Pearson_r, color = "#5050F3"), span = 0.8)+ geom_point(data = correlation, aes(x=Week, y=Pearson_r, color ="#5050F3"))+ geom_smooth(data = correlation, aes(x= Week, y = Pvalue, color= "black"), span = 0.8, se = FALSE)+ geom_point(data = correlation, aes(x=Week, y=Pvalue), color = "black")+ scale_y_continuous(sec.axis = sec_axis(~., name = "P-value"))+ xlab("Weeks censored")+ ylab("Pearson correlation")+ scale_color_manual(values = c("#5050F3", "black"), labels = c( "Correlation","P-value"))+ theme_minimal()+ theme(legend.title = element_blank(), legend.key = element_rect(colour = "white"), legend.position = "bottom", legend.text=element_text(size=12), axis.title = element_text(size = 12), axis.text = element_text(size = 12)) f1_0_data_mat <- f1_0_data[!duplicated(f1_0_data$Exp_nest),] colnames(f1_0_data_mat) mat_eggs_product <- lm(Eggs_total~ Productivity, data=f1_0_data_mat) summary(mat_eggs_product) # Call: # lm(formula = Eggs_total ~ Productivity, data = f1_0_data_mat) # # Coefficients: # (Intercept) ProductivityHigh # 414.9 344.7 # # > mat_eggs_product <- lm(Eggs_total~ Productivity, data=f1_0_data_mat) # > summary(mat_eggs_product) # # Call: # lm(formula = Eggs_total ~ Productivity, data = f1_0_data_mat) # # Residuals: # Min 1Q Median 3Q Max # -212.57 -84.31 -42.73 76.43 295.11 # # Coefficients: # Estimate Std. Error t value Pr(>|t|) # (Intercept) 414.89 54.49 7.613 2.43e-06 *** # ProductivityHigh 344.68 82.39 4.184 0.000919 *** egg15_plot <- ggplot(data = f1_0_data_mat, aes(x = Productivity , y = Eggs_15) )+ geom_violin()+ geom_boxplot(width=.2)+ geom_quasirandom(aes(col = Productivity))+ scale_color_manual(values = c( "#5050F3", "#FF6666"), labels = c("High", "Low" ))+ # labs(title = "(a)") + theme_classic(base_size = 12)+ xlab("Productivity in maternal colonies")+ ylab("Egg production")+ theme(legend.position = "none", axis.text.x = element_text(size = 12), axis.text.y = element_text(size = 12)) f0_data <- subset(f1_0_data, Maternal != 1) #16 f0 and 99 col f0_data$Week <- as.numeric(f0_data$Week) f0_data$Death <- as.factor(f0_data$Death) Fig1C <- ggplot(data = batch_data, aes(x = Week , y = Eggs_sel) )+ geom_point(aes(col = Productivity))+ scale_color_manual(values = c("#FF6666", "#5050F3", "black"), labels = c("High", "Low" ))+ geom_smooth(method = lm, colour = "black")+ # labs(title = "(b)") + xlab("Lifespan (weeks)")+ ylab("Total eggs")+ theme_classic()+ theme(legend.position = "bottom", legend.title = element_text(size = 12), legend.text = element_text(size = 12), axis.text.x = element_text(size = 12), axis.text.y = element_text(size = 12), axis.title = element_text(size = 12)) (pvalplot + egg15_plot ) / Fig1C + plot_annotation(tag_levels = 'A') ###### Test Correlation #### f1_0_numbers <- data.frame (Exp_nest = f1_0_data$Exp_nest, Week = f1_0_data$Week, Eggs_sel = f1_0_data$Eggs_sel,QueenP_sel = f1_0_data$QueenP_sel, Worker_sel= f1_0_data$Worker_sel, casteR = f1_0_data$casteR, Eggs_total = f1_0_data$Eggs_total, Queen_Ptotal = f1_0_data$Queen_Ptotal, Worker_Ptotal = f1_0_data$Worker_Ptotal, Lifespan_weeks = f1_0_data$Lifespan, Treatment = as.numeric(f1_0_data$Treatment),Mat_castR = f1_0_data$casteRf0) lapply(f1_0_numbers[,-c(1)], shapiro.test) # $Worker_Ptotal # # Shapiro-Wilk normality test # # data: X[[i]] # W = 0.84405, p-value = 3.904e-06 # $Eggs_total # # Shapiro-Wilk normality test # # data: X[[i]] # W = 0.89453, p-value = 0.0001414 # # # $Queen_Ptotal # # Shapiro-Wilk normality test # # data: X[[i]] # W = 0.91535, p-value = 0.0007927 fig2_a <- ggplot(data = f1_0_numbers, aes(x = Eggs_total , y = Eggs_sel) )+ geom_smooth (method = lm, colour = "black" )+ geom_point()+ # ylim(0,350)+ # scale_color_manual(values = c( "orange","darkorchid1"), labels = c("High", "Low" ))+ xlab("Maternal eggs")+ ylab("Daughter eggs")+ theme_classic()+ theme(legend.position = "bottom", axis.title = element_text(size = 15), axis.text.x = element_text(size = 12), axis.text.y = element_text(size = 12)) fig2_b <- ggplot(data = f1_0_numbers, aes(x = Worker_Ptotal , y = Worker_sel) )+ geom_smooth(method = lm, colour = "darkorange" , fill ="tan2")+ geom_point(color = "darkorange")+ # ylim(0,250)+ # scale_color_manual(values = c( "orange","darkorchid1"), labels = c("High", "Low" ))+ xlab("Maternal workers")+ ylab("Daughter workers")+ theme_classic()+ theme(legend.position = "bottom", axis.title = element_text(size = 15), axis.text.x = element_text(size = 12), axis.text.y = element_text(size = 12)) fig2_c <- ggplot(data = f1_0_numbers, aes(x = Queen_Ptotal , y = QueenP_sel) )+ geom_smooth(method = lm, colour = "deeppink3" , fill ="deeppink")+ geom_point( colour = "deeppink3" )+ # scale_color_manual(values = c( "orange","darkorchid1"), labels = c("High", "Low" ))+ xlab("Maternal queen pupae")+ ylab("Daughter queen pupae")+ theme_classic()+ theme(legend.position = "bottom", axis.title = element_text(size = 15), axis.text.x = element_text(size = 12), axis.text.y = element_text(size = 12)) fig2_d <- ggplot(data = f1_0_numbers, aes(x = Lifespan_weeks , y = Week) )+ geom_smooth(method = lm, colour = "#4444EC" , fill ="#9595F6")+ geom_point( colour ="#4444EC" )+ # scale_color_manual(values = c( "orange","darkorchid1"), labels = c("High", "Low" ))+ xlab("Maternal lifespan")+ ylab("Daughter lifespan")+ theme_classic()+ theme(legend.position = "bottom", axis.title = element_text(size = 15), axis.text.x = element_text(size = 12), axis.text.y = element_text(size = 12)) fig2 <- fig2_a + fig2_b + fig2_c + fig2_d + plot_annotation(tag_levels = 'A') #### Eggs #### model_eggs = glmmTMB(Eggs_sel ~ Eggs_total + (Treatment) + (1|Exp_nest) , zi = ~1, family = nbinom2, data = f1_0_numbers) res <- simulateResiduals(model_eggs, plot = T) summary(model_eggs) # Family: nbinom2 ( log ) # Formula: Eggs_sel ~ Eggs_total + (Treatment) + (1 | Exp_nest) # Zero inflation: ~1 # Data: f1_0_numbers # # AIC BIC logLik deviance df.resid # 604.6 616.7 -296.3 592.6 50 # # Random effects: # # Conditional model: # Groups Name Variance Std.Dev. # Exp_nest (Intercept) 4.122e-09 6.42e-05 # Number of obs: 56, groups: Exp_nest, 16 # # Dispersion parameter for nbinom2 family (): 1.48 # # Conditional model: # Estimate Std. Error z value Pr(>|z|) # (Intercept) 4.1759238 0.3565185 11.713 <2e-16 *** # Eggs_total 0.0002150 0.0004281 0.502 0.616 # Treatment 0.0156686 0.1378498 0.114 0.910 # --- # Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 # # Zero-inflation model: # Estimate Std. Error z value Pr(>|z|) # (Intercept) -4.179 1.207 -3.462 0.000537 *** # --- # Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 #### Workers ### model_workers = glmmTMB(Worker_sel ~ Worker_Ptotal + (Treatment) + (1|Exp_nest) , family = nbinom2, data = f1_0_numbers) res <- simulateResiduals(model_workers, plot = T) summary(model_workers) # Family: nbinom2 ( log ) # Formula: # Worker_sel ~ Worker_Ptotal + (Treatment) + (1 | Exp_nest) # Data: f1_0_numbers # # AIC BIC logLik deviance df.resid # 582.3 592.4 -286.2 572.3 51 # # Random effects: # # Conditional model: # Groups Name Variance Std.Dev. # Exp_nest (Intercept) 0.1362 0.369 # Number of obs: 56, groups: Exp_nest, 16 # # Dispersion parameter for nbinom2 family (): 1.32 # # Conditional model: # Estimate Std. Error z value Pr(>|z|) # (Intercept) 3.8185615 0.4954620 7.707 1.29e-14 *** # Worker_Ptotal 0.0002132 0.0003969 0.537 0.591 # Treatment 0.0362845 0.1877243 0.193 0.847 #### Queen pupae ### model_qp= glmmTMB(QueenP_sel ~ Queen_Ptotal + (Treatment) + (1|Exp_nest) , family = nbinom2, data = f1_0_numbers) res <- simulateResiduals(model_qp, plot = T) summary(model_qp) # Family: nbinom2 ( log ) # Formula: QueenP_sel ~ Queen_Ptotal + (Treatment) + (1 | Exp_nest) # Data: f1_0_numbers # # AIC BIC logLik deviance df.resid # 550.2 560.4 -270.1 540.2 51 # # Random effects: # # Conditional model: # Groups Name Variance Std.Dev. # Exp_nest (Intercept) 0.3743 0.6118 # Number of obs: 56, groups: Exp_nest, 16 # # Dispersion parameter for nbinom2 family (): 0.708 # # Conditional model: # Estimate Std. Error z value Pr(>|z|) # (Intercept) 3.451803 0.645247 5.35 8.82e-08 *** # Queen_Ptotal 0.001416 0.002722 0.52 0.603 # Treatment 0.030184 0.301338 0.10 0.920 #### Caste Ratio ### model_cr= glmmTMB(casteR ~ Mat_castR + (Treatment) + (1|Exp_nest) , data = f1_0_numbers) res <- simulateResiduals(model_cr, plot = T) summary(model_cr) # Family: gaussian ( identity ) # Formula: casteR ~ Mat_castR + (Treatment) + (1 | Exp_nest) # Data: f1_0_numbers # # AIC BIC logLik deviance df.resid # -16.2 -6.1 13.1 -26.2 51 # # Random effects: # # Conditional model: # Groups Name Variance Std.Dev. # Exp_nest (Intercept) 3.520e-12 1.876e-06 # Residual 3.665e-02 1.915e-01 # Number of obs: 56, groups: Exp_nest, 16 # # Dispersion estimate for gaussian family (sigma^2): 0.0367 # # Conditional model: # Estimate Std. Error z value Pr(>|z|) # (Intercept) 0.39156 0.07200 5.439 5.37e-08 *** # Mat_castR -0.08829 0.27045 -0.326 0.744 # Treatment -0.00722 0.03290 -0.219 0.826 ###### Maternal Lifespan and Lansing effect #### shapiro.test(f0_data$Week) # does the daugther lifespan exhibit normality? # # Shapiro-Wilk normality test # # data: f0_data$Week # W = 0.91894, p-value = 0.001087 # Survival F1 setwd("~/Documents/PhD_Tesis/04_selected_lines/") f1_0_data <- read.xlsx("Selection_lines_records_complete.xlsx",sheet = 3) f1_0_data$Week <- as.numeric(f1_0_data$Week) #The proportion of total reproduction accomplished until the moment of laying the F1 egg was _> percentage of reprotuction left. Produce a colum with the proportion of cummulative egss produced at each time point colonies2 <- data_na %>% group_by(Exp_nest) %>% mutate(Cum = cumsum(Eggs)/sum(Eggs)) %>% right_join(colonies) %>% select(names(colonies), everything()) colonies3 <- as.data.frame(cbind(queen.id =as.character(colonies2$Exp_nest), Cum=as.numeric(colonies2$Cum), Week=colonies2$Week)) f1_0_data_eggprop <-merge(f1_0_data, colonies3, by.x=c("Exp_nest", "Age_f0_egg"), by.y=c("queen.id","Week")) f1_0_data_eggprop$Cum <- as.numeric(f1_0_data_eggprop$Cum) # residuals only work for coxph that doesnt handle random effects coxph_model <- coxph(Surv(Week, Death) ~ Lifespan_weeks+ Cum + Productivity, data = f1_0_data_eggprop) resid_mart <- residuals(coxph_model, type = "martingale") plot(resid_mart, ylab = "Martingale Residuals", main = "Martingale Residuals Plot") abline(h = c(-2, 2), col = "red", lty = 2) # Highlight potential outliers sort(resid_mart) # Orders residuals from smallest to largest outliers <- order(abs(resid_mart), decreasing = TRUE)[1] # Get indices of extreme values f1_0_data_eggprop[outliers, ] # Show the corresponding rows in the dataset survival_prod <- coxme(Surv(Week, Death) ~ Lifespan_weeks+ Cum + Productivity + (1| Exp_nest) + (1|Recorded_dead) , data = f1_0_data_eggprop[-c(18),]) summary(survival_prod) # Mixed effects coxme model # Formula: Surv(Week, Death) ~ Lifespan_weeks + Cum + Productivity + (1 | Exp_nest) + (1 | Recorded_dead) # Data: f1_0_data_eggprop[-c(18), ] # # events, n = 54, 55 # # Random effects: # group variable sd variance # 1 Exp_nest Intercept 0.016880455 2.849498e-04 # 2 Recorded_dead Intercept 0.004214828 1.776477e-05 # Chisq df p AIC BIC # Integrated loglik 1.42 5.00 0.9224 -8.58 -18.53 # Penalized loglik 1.44 3.01 0.6972 -4.58 -10.57 # # Fixed effects: # coef exp(coef) se(coef) z p # Lifespan_weeks -0.005748 0.994268 0.023671 -0.24 0.808 # Cum 1.000021 2.718340 1.027988 0.97 0.331 # ProductivityLow 0.087027 1.090926 0.307323 0.28 0.777 # Maternal age at oviposition and daughter lifespan were not correlated survival_prod <- coxme(Surv(Week, Death) ~ Lifespan_weeks+ Age_f0_egg + Productivity + (1| Exp_nest) + (1|Recorded_dead) , data = f1_0_data_eggprop[-c(18),]) summary(survival_prod) # Mixed effects coxme model # Formula: Surv(Week, Death) ~ Lifespan_weeks + Age_f0_egg + Productivity + (1 | Exp_nest) + (1 | Recorded_dead) # Data: f1_0_data_eggprop[-c(18), ] # # events, n = 54, 55 # # Random effects: # group variable sd variance # 1 Exp_nest Intercept 0.011349702 1.288157e-04 # 2 Recorded_dead Intercept 0.003488909 1.217249e-05 # Chisq df p AIC BIC # Integrated loglik 2.28 5 0.8099 -7.72 -17.67 # Penalized loglik 2.29 3 0.5160 -3.72 -9.70 # # Fixed effects: # coef exp(coef) se(coef) z p # Lifespan_weeks -0.029478 0.970952 0.024269 -1.21 0.225 # Age_f0_egg 0.055958 1.057553 0.040953 1.37 0.172 # ProductivityLow 0.008159 1.008192 0.309060 0.03 0.979 ##### Lansing sensu lato ##### mod_lansing2 <- glmmTMB(log(Eggs_sel+1) ~ Cum + (1|Exp_nest), data = f1_0_data_eggprop) res <- simulateResiduals(mod_lansing2 , plot = T) summary(mod_lansing2) # Family: gaussian ( identity ) # Formula: log(Eggs_sel + 1) ~ Cum + (1 | Exp_nest) # Data: f1_0_data_eggprop # # AIC BIC logLik deviance df.resid # 173.1 181.2 -82.5 165.1 52 # # Random effects: # # Conditional model: # Groups Name Variance Std.Dev. # Exp_nest (Intercept) 0.121 0.3479 # Residual 1.013 1.0062 # Number of obs: 56, groups: Exp_nest, 16 # # Dispersion estimate for gaussian family (sigma^2): 1.01 # # Conditional model: # Estimate Std. Error z value Pr(>|z|) # (Intercept) 4.0494 0.7676 5.275 1.33e-07 *** # Cum -0.1759 0.9593 -0.183 0.854 mod_lansing3 <- glmmTMB(log(Worker_sel+1) ~ Cum + (1|Exp_nest), data = f1_0_data_eggprop) res <- simulateResiduals(mod_lansing3 , plot = T) summary(mod_lansing3) # Family: gaussian ( identity ) # Formula: log(Worker_sel + 1) ~ Cum + (1 | Exp_nest) # Data: f1_0_data_eggprop # # AIC BIC logLik deviance df.resid # 176.3 184.4 -84.2 168.3 52 # # Random effects: # # Conditional model: # Groups Name Variance Std.Dev. # Exp_nest (Intercept) 0.1527 0.3908 # Residual 1.0558 1.0275 # Number of obs: 56, groups: Exp_nest, 16 # # Dispersion estimate for gaussian family (sigma^2): 1.06 # # Conditional model: # Estimate Std. Error z value Pr(>|z|) # (Intercept) 4.1081 0.8139 5.047 4.48e-07 *** # Cum -0.5858 1.0144 -0.578 0.564 ggplot (data = f1_0_data_eggprop, aes(x=Age_f0, y=Cum))+ geom_point() mod_lansing4 <- glmmTMB(log(QueenP_sel+1) ~ Cum + (1|Exp_nest), data = f1_0_data_eggprop) res <- simulateResiduals(mod_lansing4 , plot = T) summary(mod_lansing4) # Family: gaussian ( identity ) # Formula: log(QueenP_sel + 1) ~ Cum + (1 | Exp_nest) # Data: f1_0_data_eggprop # # AIC BIC logLik deviance df.resid # 205.4 213.5 -98.7 197.4 52 # # Random effects: # # Conditional model: # Groups Name Variance Std.Dev. # Exp_nest (Intercept) 0.2109 0.4592 # Residual 1.8077 1.3445 # Number of obs: 56, groups: Exp_nest, 16 # # Dispersion estimate for gaussian family (sigma^2): 1.81 # # Conditional model: # Estimate Std. Error z value Pr(>|z|) # (Intercept) 3.829 1.038 3.688 0.000226 *** # Cum -1.016 1.296 -0.784 0.432820 #### Visualization Lansing per category - senescent #### f1_0_data$time_to_die <- f1_0_data$Lifespan - f1_0_data$Age_f0_egg f1_0_data_nosac <- subset(f1_0_data, Recorded_dead == 1) ggplot(f1_0_data, aes(x=Age_f0_egg)) + geom_histogram(aes(y = ..density..),color="black", fill="white", bins = 14)+ geom_density(color = "#000000", fill = "#69b3a2", alpha = 0.6)+ geom_vline(aes(xintercept=mean(Age_f0_egg)), color="black", linetype="dashed", size=1)+ theme_minimal()+ xlab("Maternal age") summary(f1_0_data$Age_f0_egg) aggregate(f1_0_data_eggprop$Cum, by = list(Age_f0cat = f1_0_data_eggprop$Age_f0cat), fivenum) #. Age_f0cat x.1 x.2 x.3 x.4 x.5 # 1 before 0.4789916 0.5741578 0.6476965 0.7314050 0.8995816 # 2 past 0.5913313 0.8149557 0.8970406 0.9680556 0.9972900 library(survival) f1_0_data_eggprop$Recorded_dead <- as.numeric(f1_0_data_eggprop$Recorded_dead) model_reprod_output <- glmmTMB(Cum ~ Age_f0cat + (1|Recorded_dead),data=f1_0_data_eggprop) res <- simulateResiduals(model_reprod_output, plot =T) summary(model_reprod_output) # Family: gaussian ( identity ) # Formula: Cum ~ Age_f0cat + (1 | Recorded_dead) # Data: f1_0_data_eggprop # # AIC BIC logLik deviance df.resid # -76.8 -68.7 42.4 -84.8 52 # # Random effects: # # Conditional model: # Groups Name Variance Std.Dev. # Recorded_dead (Intercept) 0.000315 0.01775 # Residual 0.012631 0.11239 # Number of obs: 56, groups: Recorded_dead, 2 # # Dispersion estimate for gaussian family (sigma^2): 0.0126 # # Conditional model: # Estimate Std. Error z value Pr(>|z|) # (Intercept) 0.65607 0.02615 25.09 < 2e-16 *** # Age_f0catpast 0.21391 0.03132 6.83 8.48e-12 *** fig3_1 <- ggplot(f1_0_data_eggprop, aes(x = Age_f0cat, y= Cum))+ geom_boxplot()+ geom_point()+ theme_minimal()+ ylab("Proportion of reproductive output")+ xlab("Maternal age")+ theme(axis.title = element_text(size = 15), axis.text.x = element_text(size = 12), axis.text.y = element_text(size = 12)) survplot_f3 <- ggsurvplot( fit = survfit(Surv(Week, Recorded_dead) ~Age_f0cat , data = f1_0_data_eggprop), xlab = "Weeks", ylab = "Overall survival probability", palette = c("#5050F3", "#FF6666"), conf.int = TRUE, legend.title = "Maternal age", font.legend = 12, font.y =12, font.tickslab = 12, font.title = 12, legend.labs = c("Not senescent", "Senescent")) fig3c <- ggplot(f1_0_data_eggprop, aes(x = Age_f0cat, y= QueenP_sel, fill =Age_f0cat))+ geom_boxplot()+ geom_point()+ scale_fill_manual(values= c("#9595F6", "#FF9494"))+ theme_classic()+ ylab("Daughter queen pupae")+ xlab("Maternal age")+ theme(legend.position = "bottom", axis.title = element_text(size = 15), axis.text.x = element_text(size = 12), axis.text.y = element_text(size = 12)) fig3_1 + survplot_f3$plot +fig3c +plot_annotation(tag_levels = 'A') ######## Testing maternal lineages #### # For this any colony with less than 3 replicates is removed setwd("~/Documents/PhD_Tesis/04_selected_lines/") f1_0_data <- read.xlsx("Selection_lines_records_complete.xlsx",sheet = 2) #f1_0_data$sEggs_total <- as.numeric(f1_0_data$sEggs_total) #f1_0_data$sLifespan_weeks <- as.numeric(f1_0_data$sLifespan_weeks) f1_0_data$casteR <- f1_0_data$QueenP_sel/(f1_0_data$QueenP_sel + f1_0_data$Worker_sel) f1_0_data$casteRf0 <- f1_0_data$Worker_Ptotal /(f1_0_data$Worker_Ptotal + f1_0_data$Queen_Ptotal) f1_0_data$Maternal <- as.factor(f1_0_data$Maternal ) summary(f1_0_data) #f1_0_data$Comments_f0 <- as.factor(f1_0_data$Comments_f0 ) f1_0_data$Comments <- as.factor(f1_0_data$Comments) f1_0_data$Recorded_dead <- as.factor(f1_0_data$Recorded_dead) f1_0_data <- f1_0_data[f1_0_data$Exp_nest %in% c("C10_30", "A30_7", "A30_10", "A30_29","C10_13", "C10_1", "E20_2", "E20_17", "F10_6","A30_17", "E20_16"),] f1_0_data$Exp_nest <- as.factor(f1_0_data$Exp_nest) f1_0_data$Exp_nest <- factor(f1_0_data$Exp_nest, levels = c("C10_30", "A30_7", "A30_10", "A30_29","C10_13", "C10_1", "E20_2", "E20_17", "F10_6","A30_17", "E20_16")) #by qp f1_0_data$Exp_nest <- factor(f1_0_data$Exp_nest, levels = c("C10_13", "A30_29", "A30_10", "C10_1","E20_17", "A30_17", "C10_30", "A30_7","E20_2", "F10_6", "E20_16")) #by qp? f1_0_data$Productivity <- factor(f1_0_data$Productivity, levels = c("Low", "High")) f1_0_data_nomat <- f1_0_data[f1_0_data$Maternal == 0,] f1_0_data_nomat$Productivity <- factor(f1_0_data_nomat$Productivity, levels = c("Low", "High")) Lines_eggs_model = glmmTMB(Eggs_sel ~ Exp_nest + (1 | Recorded_dead), data = f1_0_data_nomat , family = "nbinom2", na.action = "na.fail") res <- simulateResiduals(Lines_eggs_model , plot = T) car::Anova(Lines_eggs_model) # Analysis of Deviance Table (Type II Wald chisquare tests) # # Response: Eggs_sel # Chisq Df Pr(>Chisq) # Exp_nest 22.504 10 0.01273 * # --- summary(Lines_eggs_model) # Family: nbinom2 ( log ) # Formula: Eggs_sel ~ Exp_nest + (1 | Recorded_dead) # Data: f1_0_data_nomat # # AIC BIC logLik deviance df.resid # 523.3 547.6 -248.6 497.3 35 # # Random effects: # # Conditional model: # Groups Name Variance Std.Dev. # Recorded_dead (Intercept) 9.545e-10 3.09e-05 # Number of obs: 48, groups: Recorded_dead, 2 # # Dispersion parameter for nbinom2 family (): 1.56 # # Conditional model: # Estimate Std. Error z value Pr(>|z|) # (Intercept) 3.7917 0.4708 8.054 8.02e-16 *** # Exp_nestA30_29 0.1253 0.6223 0.201 0.8405 # Exp_nestA30_10 0.5416 0.5939 0.912 0.3618 # Exp_nestC10_1 0.7830 0.5934 1.319 0.1871 # Exp_nestE20_17 0.5390 0.5939 0.908 0.3641 # Exp_nestA30_17 0.8293 0.5934 1.398 0.1622 # Exp_nestC10_30 -1.1891 0.6331 -1.878 0.0603 . # Exp_nestA30_7 0.5624 0.5939 0.947 0.3436 # Exp_nestE20_2 0.3514 0.5756 0.610 0.5416 # Exp_nestF10_6 0.8397 0.6626 1.267 0.2050 # Exp_nestE20_16 1.2560 0.6617 1.898 0.0577 . summary(glht(Lines_eggs_model,linfct=mcp( Exp_nest = "Tukey"))) # Simultaneous Tests for General Linear Hypotheses # # Multiple Comparisons of Means: Tukey Contrasts # # # Fit: glmmTMB(formula = Eggs_sel ~ Exp_nest + (1 | Recorded_dead), # data = f1_0_data_nomat, family = "nbinom2", na.action = "na.fail", # ziformula = ~0, dispformula = ~1) # # Linear Hypotheses: # Estimate Std. Error z value Pr(>|z|) # A30_29 - C10_13 == 0 0.125270 0.622266 0.201 1.0000 # A30_10 - C10_13 == 0 0.541620 0.593924 0.912 0.9980 # C10_1 - C10_13 == 0 0.782970 0.593450 1.319 0.9650 # E20_17 - C10_13 == 0 0.538993 0.593930 0.908 0.9981 # A30_17 - C10_13 == 0 0.829303 0.593371 1.398 0.9485 # C10_30 - C10_13 == 0 -1.189051 0.633055 -1.878 0.7280 # A30_7 - C10_13 == 0 0.562401 0.593878 0.947 0.9973 # E20_2 - C10_13 == 0 0.351394 0.575633 0.610 0.9999 # F10_6 - C10_13 == 0 0.839747 0.662584 1.267 0.9736 # E20_16 - C10_13 == 0 1.255976 0.661749 1.898 0.7155 # A30_10 - A30_29 == 0 0.416350 0.544674 0.764 0.9996 # C10_1 - A30_29 == 0 0.657700 0.544157 1.209 0.9813 # E20_17 - A30_29 == 0 0.413722 0.544680 0.760 0.9996 # A30_17 - A30_29 == 0 0.704032 0.544071 1.294 0.9694 # C10_30 - A30_29 == 0 -1.314322 0.587096 -2.239 0.4734 # A30_7 - A30_29 == 0 0.437131 0.544624 0.803 0.9993 # E20_2 - A30_29 == 0 0.226124 0.524669 0.431 1.0000 # F10_6 - A30_29 == 0 0.714476 0.618823 1.155 0.9867 # E20_16 - A30_29 == 0 1.130706 0.617929 1.830 0.7593 # C10_1 - A30_10 == 0 0.241350 0.511505 0.472 1.0000 # E20_17 - A30_10 == 0 -0.002628 0.512062 -0.005 1.0000 # A30_17 - A30_10 == 0 0.287682 0.511414 0.563 1.0000 # C10_30 - A30_10 == 0 -1.730672 0.556968 -3.107 0.0679 . # A30_7 - A30_10 == 0 0.020781 0.512002 0.041 1.0000 # E20_2 - A30_10 == 0 -0.190226 0.490722 -0.388 1.0000 # F10_6 - A30_10 == 0 0.298126 0.590316 0.505 1.0000 # E20_16 - A30_10 == 0 0.714356 0.589379 1.212 0.9809 # E20_17 - C10_1 == 0 -0.243978 0.511512 -0.477 1.0000 # A30_17 - C10_1 == 0 0.046332 0.510863 0.091 1.0000 # C10_30 - C10_1 == 0 -1.972022 0.556462 -3.544 0.0169 * # A30_7 - C10_1 == 0 -0.220569 0.511452 -0.431 1.0000 # E20_2 - C10_1 == 0 -0.431576 0.490149 -0.881 0.9985 # F10_6 - C10_1 == 0 0.056776 0.589839 0.096 1.0000 # E20_16 - C10_1 == 0 0.473006 0.588901 0.803 0.9993 # A30_17 - E20_17 == 0 0.290310 0.511420 0.568 1.0000 # C10_30 - E20_17 == 0 -1.728044 0.556974 -3.103 0.0696 . # A30_7 - E20_17 == 0 0.023408 0.512009 0.046 1.0000 # E20_2 - E20_17 == 0 -0.187598 0.490729 -0.382 1.0000 # F10_6 - E20_17 == 0 0.300754 0.590321 0.509 1.0000 # E20_16 - E20_17 == 0 0.716984 0.589384 1.216 0.9804 # C10_30 - A30_17 == 0 -2.018354 0.556378 -3.628 0.0125 * # A30_7 - A30_17 == 0 -0.266902 0.511361 -0.522 1.0000 # E20_2 - A30_17 == 0 -0.477909 0.490053 -0.975 0.9965 # F10_6 - A30_17 == 0 0.010444 0.589760 0.018 1.0000 # E20_16 - A30_17 == 0 0.426674 0.588822 0.725 0.9997 # A30_7 - C10_30 == 0 1.751452 0.556919 3.145 0.0605 . # E20_2 - C10_30 == 0 1.540445 0.537421 2.866 0.1316 # F10_6 - C10_30 == 0 2.028798 0.629671 3.222 0.0487 * # E20_16 - C10_30 == 0 2.445028 0.628792 3.888 <0.01 ** # E20_2 - A30_7 == 0 -0.211007 0.490667 -0.430 1.0000 # F10_6 - A30_7 == 0 0.277346 0.590270 0.470 1.0000 # E20_16 - A30_7 == 0 0.693575 0.589333 1.177 0.9847 # F10_6 - E20_2 == 0 0.488352 0.571910 0.854 0.9989 # E20_16 - E20_2 == 0 0.904582 0.570943 1.584 0.8874 # E20_16 - F10_6 == 0 0.416230 0.658513 0.632 0.9999 Lines_worker_model = glmmTMB(Worker_sel ~ Exp_nest + (1 | Recorded_dead), data = f1_0_data_nomat , family = "nbinom2", na.action = "na.fail") res <- simulateResiduals(Lines_worker_model , plot = T) car::Anova(Lines_worker_model) summary(Lines_worker_model) Lines_qp_model = glmmTMB(QueenP_sel ~ Exp_nest + (1 | Recorded_dead), data = f1_0_data_nomat , family = "nbinom2", na.action = "na.fail") res <- simulateResiduals(Lines_qp_model , plot = T) f1_0_data_nomat$Exp_nest car::Anova(Lines_qp_model) # Analysis of Deviance Table (Type II Wald chisquare tests) # # Response: QueenP_sel # Chisq Df Pr(>Chisq) # Exp_nest 34.732 10 0.0001387 *** summary(Lines_qp_model) # Family: nbinom2 ( log ) # Formula: QueenP_sel ~ Exp_nest + (1 | Recorded_dead) # Data: f1_0_data_nomat # # AIC BIC logLik deviance df.resid # 474.6 498.9 -224.3 448.6 35 # # Random effects: # # Conditional model: # Groups Name Variance Std.Dev. # Recorded_dead (Intercept) 1.234e-09 3.512e-05 # Number of obs: 48, groups: Recorded_dead, 2 # # Dispersion parameter for nbinom2 family (): 0.867 # # Conditional model: # Estimate Std. Error z value Pr(>|z|) # (Intercept) 3.1918 0.6309 5.059 4.21e-07 *** # Exp_nestA30_29 -0.8893 0.8433 -1.054 0.2917 # Exp_nestA30_10 -0.7939 0.8042 -0.987 0.3235 # Exp_nestC10_1 0.7594 0.7953 0.955 0.3396 # Exp_nestE20_17 0.9764 0.7948 1.228 0.2193 # Exp_nestA30_17 1.1149 0.7945 1.403 0.1605 # Exp_nestC10_30 -1.0816 0.8465 -1.278 0.2013 # Exp_nestA30_7 1.6894 0.7938 2.128 0.0333 * # Exp_nestE20_2 0.6332 0.7706 0.822 0.4113 # Exp_nestF10_6 0.8512 0.8878 0.959 0.3377 # Exp_nestE20_16 1.7402 0.8858 1.964 0.0495 * summary(glht(Lines_qp_model,linfct=mcp( Exp_nest = "Tukey"))) # Simultaneous Tests for General Linear Hypotheses # Multiple Comparisons of Means: Tukey Contrasts # # # Fit: glmmTMB(formula = QueenP_sel ~ Exp_nest + (1 | Recorded_dead), # data = f1_0_data_nomat, family = "nbinom2", na.action = "na.fail", # ziformula = ~0, dispformula = ~1) # # Linear Hypotheses: # Estimate Std. Error z value Pr(>|z|) # A30_29 - C10_13 == 0 -0.88926 0.84335 -1.054 0.9934 # A30_10 - C10_13 == 0 -0.79395 0.80422 -0.987 0.9962 # C10_1 - C10_13 == 0 0.75940 0.79526 0.955 0.9971 # E20_17 - C10_13 == 0 0.97637 0.79478 1.228 0.9789 # A30_17 - C10_13 == 0 1.11492 0.79453 1.403 0.9471 # C10_30 - C10_13 == 0 -1.08163 0.84648 -1.278 0.9721 # A30_7 - C10_13 == 0 1.68944 0.79379 2.128 0.5528 # E20_2 - C10_13 == 0 0.63317 0.77058 0.822 0.9992 # F10_6 - C10_13 == 0 0.85121 0.88778 0.959 0.9970 # E20_16 - C10_13 == 0 1.74023 0.88584 1.964 0.6703 # A30_10 - A30_29 == 0 0.09531 0.74966 0.127 1.0000 # C10_1 - A30_29 == 0 1.64866 0.74004 2.228 0.4817 # E20_17 - A30_29 == 0 1.86563 0.73953 2.523 0.2867 # A30_17 - A30_29 == 0 2.00418 0.73926 2.711 0.1924 # C10_30 - A30_29 == 0 -0.19237 0.79483 -0.242 1.0000 # A30_7 - A30_29 == 0 2.57870 0.73846 3.492 0.0197 * # E20_2 - A30_29 == 0 1.52243 0.71346 2.134 0.5487 # F10_6 - A30_29 == 0 1.74047 0.83868 2.075 0.5914 # E20_16 - A30_29 == 0 2.62949 0.83662 3.143 0.0614 . # C10_1 - A30_10 == 0 1.55335 0.69512 2.235 0.4763 # E20_17 - A30_10 == 0 1.77032 0.69458 2.549 0.2726 # A30_17 - A30_10 == 0 1.90887 0.69429 2.749 0.1753 # C10_30 - A30_10 == 0 -0.28768 0.75319 -0.382 1.0000 # A30_7 - A30_10 == 0 2.48339 0.69345 3.581 0.0147 * # E20_2 - A30_10 == 0 1.42712 0.66675 2.140 0.5443 # F10_6 - A30_10 == 0 1.64516 0.79932 2.058 0.6038 # E20_16 - A30_10 == 0 2.53418 0.79717 3.179 0.0560 . # E20_17 - C10_1 == 0 0.21697 0.68418 0.317 1.0000 # A30_17 - C10_1 == 0 0.35552 0.68389 0.520 1.0000 # C10_30 - C10_1 == 0 -1.84103 0.74361 -2.476 0.3151 # A30_7 - C10_1 == 0 0.93004 0.68303 1.362 0.9567 # E20_2 - C10_1 == 0 -0.12623 0.65591 -0.192 1.0000 # F10_6 - C10_1 == 0 0.09181 0.79031 0.116 1.0000 # E20_16 - C10_1 == 0 0.98083 0.78812 1.245 0.9769 # A30_17 - E20_17 == 0 0.13855 0.68334 0.203 1.0000 # C10_30 - E20_17 == 0 -2.05800 0.74311 -2.769 0.1676 # A30_7 - E20_17 == 0 0.71307 0.68248 1.045 0.9939 # E20_2 - E20_17 == 0 -0.34320 0.65534 -0.524 1.0000 # F10_6 - E20_17 == 0 -0.12516 0.78983 -0.158 1.0000 # E20_16 - E20_17 == 0 0.76386 0.78765 0.970 0.9967 # C10_30 - A30_17 == 0 -2.19655 0.74284 -2.957 0.1038 # A30_7 - A30_17 == 0 0.57452 0.68219 0.842 0.9990 # E20_2 - A30_17 == 0 -0.48175 0.65503 -0.735 0.9997 # F10_6 - A30_17 == 0 -0.26371 0.78958 -0.334 1.0000 # E20_16 - A30_17 == 0 0.62531 0.78739 0.794 0.9994 # A30_7 - C10_30 == 0 2.77107 0.74205 3.734 <0.01 ** # E20_2 - C10_30 == 0 1.71480 0.71716 2.391 0.3687 # F10_6 - C10_30 == 0 1.93284 0.84183 2.296 0.4324 # E20_16 - C10_30 == 0 2.82186 0.83979 3.360 0.0312 * # E20_2 - A30_7 == 0 -1.05627 0.65413 -1.615 0.8747 # F10_6 - A30_7 == 0 -0.83823 0.78883 -1.063 0.9930 # E20_16 - A30_7 == 0 0.05079 0.78664 0.065 1.0000 # F10_6 - E20_2 == 0 0.21804 0.76547 0.285 1.0000 # E20_16 - E20_2 == 0 1.10706 0.76322 1.451 0.9342 # E20_16 - F10_6 == 0 0.88902 0.88140 1.009 0.9954 mean_C10_30 <- predict(Lines_qp_model, newdata = f1_0_data_nomat[f1_0_data_nomat$Exp_nest == "C10_30", ], type = "response") mean_E20_16 <- predict(Lines_qp_model, newdata = f1_0_data_nomat[f1_0_data_nomat$Exp_nest == "E20_16", ], type = "response") f1_0_data_nomat$logQueenP_sel <- log(f1_0_data_nomat$QueenP_sel) aggregate(f1_0_data_nomat$logQueenP_sel,by = list(f1_0_data_nomat$Exp_nest), median) # Group.1 x # 1 C10_13 3.4657359 # 2 A30_29 2.2499048 # 3 A30_10 2.3025851 # 4 C10_1 3.7376696 # 5 E20_17 3.8918203 # 6 A30_17 4.3040651 # 7 C10_30 0.8958797 # 8 A30_7 1.9459101 # 9 E20_2 3.7903499 # 10 F10_6 3.9889840 # 11 E20_16 5.0998664 tukey_qp <- glht(Lines_qp_model,linfct=mcp( Exp_nest = "Tukey")) cld(tukey_qp) # C10_30 A30_7 A30_10 A30_29 C10_13 C10_1 E20_2 E20_17 F10_6 A30_17 E20_16 # "a" "c" "ab" "ab" "ac" "ac" "ac" "ac" "ac" "ac" "bc" f1_0_data_nomat$qp_letters <- c(rep("ab",5),rep("ac",5), rep("ab", 4), rep("c", 5), rep("ac", 5), rep("ac", 3), rep("a", 4), rep("bc", 3), rep("ac",5), rep("ac", 6), rep("ac",3)) f1_0_data_nomat$Exp_nest <- factor(f1_0_data_nomat$Exp_nest, levels = c("C10_30", "A30_7", "A30_10", "A30_29","C10_13", "C10_1", "E20_2", "E20_17", "F10_6","A30_17", "E20_16")) #by qp library(viridis) library(ggpubr) library(RColorBrewer) Fig2_Aqp_sel <- ggplot(data = f1_0_data_nomat, aes(x =Exp_nest , y = log(QueenP_sel), fill = Exp_nest))+ geom_boxplot(outlier.shape = NA)+ scale_fill_brewer(palette = "PRGn")+ #scale_fill_viridis(discrete = TRUE, option = "D")+ geom_quasirandom()+ # ylim(0,350)+ xlab("F0 Maternal colony")+ ylab("F1 log queen pupae")+ theme_minimal()+ theme(legend.position = "none", axis.title = element_text(size = 15), axis.text.x = element_text(size = 12), axis.text.y = element_text(size = 12))+ geom_text (aes(y = 7,label= qp_letters)) # annotate("text", x=1, y = 250, label = "5")+ # annotate("text", x=2, y = 250, label = "5")+ # annotate("text", x=3, y = 250, label = "3")+ # annotate("text", x=4, y = 250, label = "5")+ # annotate("text", x=5, y = 250, label = "3")+ # annotate("text", x=6, y = 250, label = "5")+ # annotate("text", x=7, y = 250, label = "4")+ # annotate("text", x=8, y = 250, label = "4")+ # annotate("text", x=9, y = 250, label = "3")+ # annotate("text", x=10, y = 250, label = "6")+ # annotate("text", x=11, y = 250, label = "3") library(emmeans) cld_eggs = as.data.frame( cld(emmeans(correctedModelGAM, ~ Exp_nest ), Letters = letters ) ) f1_0_data <- merge(f1_0_data, cld_eggs, by = "Exp_nest") FigS2_egg_sel <- ggplot(data = f1_0_data_nomat, aes(x =Exp_nest , y = log(Eggs_sel), fill = Exp_nest))+ geom_boxplot(outlier.shape = NA)+ #scale_fill_viridis_d()+ scale_fill_brewer(palette = "PRGn")+ geom_quasirandom()+ # labs(title="(a)")+ xlab("F0 Maternal colony")+ ylab("F1 log eggs")+ # ylim(0,350)+ geom_text(aes(y = 6.5, label= eggs_letters))+ theme_minimal()+ theme(legend.position = "none", axis.title = element_text(size=15), axis.text.x = element_text(size = 12), axis.text.y = element_text(size = 12)) Fig2_Aqp_sel / FigS2_egg_sel + plot_annotation(tag_levels = 'A') # annotate("text", x=1, y = 250, label = "5")+ # annotate("text", x=2, y = 250, label = "5")+ # annotate("text", x=3, y = 250, label = "3")+ # annotate("text", x=4, y = 250, label = "5")+ # annotate("text", x=5, y = 250, label = "3")+ # annotate("text", x=6, y = 250, label = "5")+ # annotate("text", x=7, y = 250, label = "4")+ # annotate("text", x=8, y = 250, label = "4")+ # annotate("text", x=9, y = 250, label = "3")+ # annotate("text", x=10, y = 250, label = "6")+ # annotate("text", x=11, y = 250, label = "3") lines_casteratio_model = glmmTMB(casteR ~ Exp_nest + (1|Recorded_dead) , data = f1_0_data ) res <- simulateResiduals(lines_casteratio_model, plot = T) summary(lines_casteratio_model) # Family: gaussian ( identity ) # Formula: casteR ~ Exp_nest + (1 | Recorded_dead) # Data: f1_0_data # # AIC BIC logLik deviance df.resid # 2.1 26.4 12.0 -23.9 35 # # Random effects: # # Conditional model: # Groups Name Variance Std.Dev. # Recorded_dead (Intercept) 1.439e-12 0.0000012 # Residual 3.555e-02 0.1885530 # Number of obs: 48, groups: Recorded_dead, 2 # # Dispersion estimate for gaussian family (sigma^2): 0.0356 # # Conditional model: # Estimate Std. Error z value Pr(>|z|) # (Intercept) 0.296125 0.108861 2.720 0.00652 ** # Exp_nestA30_29 -0.057241 0.144010 -0.398 0.69101 # Exp_nestA30_10 0.004197 0.137700 0.030 0.97568 # Exp_nestC10_1 0.134218 0.137700 0.975 0.32970 # Exp_nestE20_17 0.087007 0.137700 0.632 0.52748 # Exp_nestA30_17 0.065419 0.137700 0.475 0.63473 # Exp_nestC10_30 -0.034421 0.144010 -0.239 0.81109 # Exp_nestA30_7 0.142733 0.137700 1.037 0.29995 # Exp_nestE20_2 0.084306 0.133327 0.632 0.52717 # Exp_nestF10_6 0.115357 0.153953 0.749 0.45368 # Exp_nestE20_16 0.161002 0.153953 1.046 0.29566 car::Anova(lines_casteratio_model) # Analysis of Deviance Table (Type II Wald chisquare tests) # # Response: casteR # Chisq Df Pr(>Chisq) # Exp_nest 6.316 10 0.7881 Fig2B_cr_sel <- ggplot(data = f1_0_data_nomat, aes(x =Exp_nest , y = casteR, fill = Exp_nest))+ geom_boxplot( outlier.shape = NA, colour = "grey20")+ scale_fill_viridis( discrete = TRUE, option = "D")+ geom_quasirandom()+ labs(title="(b)")+ xlab("F0 Maternal colony")+ ylab("F1 caste ratio")+ ylim(0,1)+ theme_pubr()+ theme(legend.position = "none", axis.text.x = element_text(size = 12), axis.text.y = element_text(size = 12)) Fig2_Aqp_sel / Fig2B_cr_sel survival_lines <- coxme(Surv(Week, Death) ~ (Exp_nest) + (1|Recorded_dead) , data = f1_0_data_nomat) survival_lines2 <- coxme(Surv(Week, Death) ~ (1|Recorded_dead) , data = f1_0_data_nomat) anova(survival_lines2, survival_lines) # Analysis of Deviance Table # Cox model: response is Surv(Week, Death) # Model 1: ~(1 | Recorded_dead) # Model 2: ~(Exp_nest) + (1 | Recorded_dead) # loglik Chisq Df P(>|Chi|) # 1 -137.73 # 2 -133.24 8.9891 10 0.5331 car::Anova(survival_lines) # Analysis of Deviance Table (Type II tests) # # Response: Surv(Week, Death) # Df Chisq Pr(>Chisq) # Exp_nest 10 9.4903 0.4863 summary(survival_lines) lif_sel <- ggplot(data = f1_0_data_nomat, aes(x =Exp_nest , y = Week, fill = Exp_nest))+ geom_boxplot( outlier.shape = NA)+ scale_fill_viridis_d()+ geom_quasirandom()+ # labs(title="(b)")+ xlab("F0 Maternal colony")+ ylab("F1 lifespan (weeks)")+ theme_pubr()+ # ylim(13,75)+ theme(legend.position = "none", axis.text.x = element_text(size = 12), axis.text.y = element_text(size = 12)) egg_sel / qp_sel/ cr_sel /lif_sel -> F1_plots qp_sel/ lif_sel -> F1_2plot egg_sel / cr_sel -> supFig ######## Testing maternal lineages #### # For this any colony with less than 3 replicates is removed setwd("~/Documents/PhD_Tesis/04_selected_lines/") f1_0_data <- read.xlsx("Selection_lines_records_complete.xlsx",sheet = 2) #f1_0_data$sEggs_total <- as.numeric(f1_0_data$sEggs_total) #f1_0_data$sLifespan_weeks <- as.numeric(f1_0_data$sLifespan_weeks) f1_0_data$casteR <- f1_0_data$QueenP_sel/(f1_0_data$QueenP_sel + f1_0_data$Worker_sel) f1_0_data$casteRf0 <- f1_0_data$Worker_Ptotal /(f1_0_data$Worker_Ptotal + f1_0_data$Queen_Ptotal) f1_0_data$Maternal <- as.factor(f1_0_data$Maternal ) summary(f1_0_data) #f1_0_data$Comments_f0 <- as.factor(f1_0_data$Comments_f0 ) f1_0_data$Comments <- as.factor(f1_0_data$Comments) f1_0_data$Recorded_dead <- as.factor(f1_0_data$Recorded_dead) f1_0_data <- f1_0_data[f1_0_data$Exp_nest %in% c("C10_30", "A30_7", "A30_10", "A30_29","C10_13", "C10_1", "E20_2", "E20_17", "F10_6","A30_17", "E20_16"),] f1_0_data$Exp_nest <- as.factor(f1_0_data$Exp_nest) f1_0_data$Exp_nest <- factor(f1_0_data$Exp_nest, levels = c("C10_30", "A30_7", "A30_10", "A30_29","C10_13", "C10_1", "E20_2", "E20_17", "F10_6","A30_17", "E20_16")) #by qp f1_0_data$Exp_nest <- factor(f1_0_data$Exp_nest, levels = c("C10_13", "A30_29", "A30_10", "C10_1","E20_17", "A30_17", "C10_30", "A30_7","E20_2", "F10_6", "E20_16")) #by qp? f1_0_data$Productivity <- factor(f1_0_data$Productivity, levels = c("Low", "High")) f1_0_data_nomat$Productivity <- factor(f1_0_data_nomat$Productivity, levels = c("Low", "High")) correctedModelGAM = glmmTMB(QueenP_sel ~ Exp_nest + (1 | Recorded_dead), data = f1_0_data_nomat , family = "nbinom2", na.action = "na.fail") res <- simulateResiduals(correctedModelGAM, plot = T) summary(correctedModelGAM) glmmTMB:::Anova.glmmTMB(QueenP_sel ~ Exp_nest + (1 | Recorded_dead), data = f1_0_data_nomat , family = "nbinom2", na.action = "na.fail") car::Anova(correctedModelGAM) summary(glht(correctedModelGAM,linfct=mcp( Exp_nest = "Tukey"))) tukey_qp <- glht(correctedModelGAM,linfct=mcp( Exp_nest = "Tukey")) # Simultaneous Tests for General Linear Hypotheses # # Multiple Comparisons of Means: Tukey Contrasts # # # Fit: glmmTMB(formula = QueenP_sel ~ Exp_nest + (1 | Recorded_dead) + # (1 | Treatment), data = f1_0_data, family = "nbinom2", na.action = "na.fail", # ziformula = ~0, dispformula = ~1) # # Linear Hypotheses: # Estimate Std. Error z value Pr(>|z|) # A30_17 - A30_10 == 0 1.90887 0.69429 2.749 0.1760 # A30_29 - A30_10 == 0 -0.09531 0.74966 -0.127 1.0000 # A30_7 - A30_10 == 0 2.48339 0.69344 3.581 0.0146 * # C10_1 - A30_10 == 0 1.55335 0.69512 2.235 0.4756 # C10_13 - A30_10 == 0 0.79395 0.80422 0.987 0.9961 # C10_30 - A30_10 == 0 -0.28768 0.75319 -0.382 1.0000 # E20_16 - A30_10 == 0 2.53418 0.79717 3.179 0.0555 . # E20_17 - A30_10 == 0 1.77032 0.69458 2.549 0.2727 # E20_2 - A30_10 == 0 1.42712 0.66675 2.140 0.5443 # F10_6 - A30_10 == 0 1.64516 0.79932 2.058 0.6031 # A30_29 - A30_17 == 0 -2.00418 0.73926 -2.711 0.1912 # A30_7 - A30_17 == 0 0.57452 0.68219 0.842 0.9990 # C10_1 - A30_17 == 0 -0.35552 0.68389 -0.520 1.0000 # C10_13 - A30_17 == 0 -1.11492 0.79453 -1.403 0.9470 # C10_30 - A30_17 == 0 -2.19655 0.74284 -2.957 0.1039 # E20_16 - A30_17 == 0 0.62531 0.78739 0.794 0.9994 # E20_17 - A30_17 == 0 -0.13855 0.68334 -0.203 1.0000 # E20_2 - A30_17 == 0 -0.48175 0.65503 -0.735 0.9997 # F10_6 - A30_17 == 0 -0.26371 0.78958 -0.334 1.0000 # A30_7 - A30_29 == 0 2.57870 0.73846 3.492 0.0202 * # C10_1 - A30_29 == 0 1.64866 0.74004 2.228 0.4803 # C10_13 - A30_29 == 0 0.88926 0.84335 1.054 0.9934 # C10_30 - A30_29 == 0 -0.19237 0.79483 -0.242 1.0000 # E20_16 - A30_29 == 0 2.62949 0.83662 3.143 0.0616 . # E20_17 - A30_29 == 0 1.86563 0.73953 2.523 0.2876 # E20_2 - A30_29 == 0 1.52243 0.71346 2.134 0.5491 # F10_6 - A30_29 == 0 1.74047 0.83868 2.075 0.5909 # C10_1 - A30_7 == 0 -0.93004 0.68303 -1.362 0.9565 # C10_13 - A30_7 == 0 -1.68944 0.79379 -2.128 0.5521 # C10_30 - A30_7 == 0 -2.77107 0.74205 -3.734 <0.01 ** # E20_16 - A30_7 == 0 0.05079 0.78664 0.065 1.0000 # E20_17 - A30_7 == 0 -0.71307 0.68248 -1.045 0.9939 # E20_2 - A30_7 == 0 -1.05627 0.65413 -1.615 0.8744 # F10_6 - A30_7 == 0 -0.83823 0.78883 -1.063 0.9930 # C10_13 - C10_1 == 0 -0.75940 0.79526 -0.955 0.9971 # C10_30 - C10_1 == 0 -1.84103 0.74361 -2.476 0.3151 # E20_16 - C10_1 == 0 0.98083 0.78812 1.245 0.9769 # E20_17 - C10_1 == 0 0.21697 0.68418 0.317 1.0000 # E20_2 - C10_1 == 0 -0.12623 0.65591 -0.192 1.0000 # F10_6 - C10_1 == 0 0.09181 0.79031 0.116 1.0000 # C10_30 - C10_13 == 0 -1.08163 0.84648 -1.278 0.9721 # E20_16 - C10_13 == 0 1.74022 0.88584 1.964 0.6698 # E20_17 - C10_13 == 0 0.97637 0.79479 1.228 0.9789 # E20_2 - C10_13 == 0 0.63316 0.77058 0.822 0.9992 # F10_6 - C10_13 == 0 0.85120 0.88778 0.959 0.9970 # E20_16 - C10_30 == 0 2.82186 0.83979 3.360 0.0313 * # E20_17 - C10_30 == 0 2.05800 0.74311 2.769 0.1675 # E20_2 - C10_30 == 0 1.71480 0.71716 2.391 0.3683 # F10_6 - C10_30 == 0 1.93284 0.84183 2.296 0.4319 # E20_17 - E20_16 == 0 -0.76386 0.78765 -0.970 0.9967 # E20_2 - E20_16 == 0 -1.10706 0.76322 -1.451 0.9342 # F10_6 - E20_16 == 0 -0.88902 0.88140 -1.009 0.9954 # E20_2 - E20_17 == 0 -0.34320 0.65534 -0.524 1.0000 # F10_6 - E20_17 == 0 -0.12516 0.78983 -0.158 1.0000 # F10_6 - E20_2 == 0 0.21804 0.76547 0.285 1.0000 f1_0_data_nomat$logQueenP_sel <- log(f1_0_data_nomat$QueenP_sel) aggregate(f1_0_data_nomat$logQueenP_sel,by = list(f1_0_data_nomat$Exp_nest), median) # Group.1 x # 1 C10_13 3.4657359 # 2 A30_29 2.2499048 # 3 A30_10 2.3025851 # 4 C10_1 3.7376696 # 5 E20_17 3.8918203 # 6 A30_17 4.3040651 # 7 C10_30 0.8958797 # 8 A30_7 1.9459101 # 9 E20_2 3.7903499 # 10 F10_6 3.9889840 # 11 E20_16 5.0998664 cld(tukey_qp) # C10_30 A30_7 A30_10 A30_29 C10_13 C10_1 E20_2 E20_17 F10_6 A30_17 E20_16 # "a" "c" "ab" "ab" "ac" "ac" "ac" "ac" "ac" "ac" "bc" f1_0_data_nomat$qp_letters <- c(rep("ab",5),rep("ac",5), rep("ab", 4), rep("c", 5), rep("ac", 5), rep("ac", 3), rep("a", 4), rep("bc", 3), rep("ac",5), rep("ac", 6), rep("ac",3)) f1_0_data_nomat$Exp_nest <- factor(f1_0_data_nomat$Exp_nest, levels = c("C10_30", "A30_7", "A30_10", "A30_29","C10_13", "C10_1", "E20_2", "E20_17", "F10_6","A30_17", "E20_16")) #by qp library(viridis) library(ggpubr) library(RColorBrewer) Fig2_Aqp_sel <- ggplot(data = f1_0_data_nomat, aes(x =Exp_nest , y = log(QueenP_sel), fill = Productivity))+ geom_boxplot(outlier.shape = NA)+ scale_fill_manual(values = c("#9595F6", "#FF9494"), labels = c("High", "Low" ))+ #scale_fill_viridis(discrete = TRUE, option = "D")+ geom_quasirandom()+ # ylim(0,350)+ xlab("Maternal colony")+ ylab("Daughter log(queen pupae)")+ theme_minimal()+ theme(legend.position = "none", axis.title = element_text(size = 15), axis.text.x = element_text(size = 12), axis.text.y = element_text(size = 12))+ geom_text (aes(y = 7,label= qp_letters)) # annotate("text", x=1, y = 250, label = "5")+ # annotate("text", x=2, y = 250, label = "5")+ # annotate("text", x=3, y = 250, label = "3")+ # annotate("text", x=4, y = 250, label = "5")+ # annotate("text", x=5, y = 250, label = "3")+ # annotate("text", x=6, y = 250, label = "5")+ # annotate("text", x=7, y = 250, label = "4")+ # annotate("text", x=8, y = 250, label = "4")+ # annotate("text", x=9, y = 250, label = "3")+ # annotate("text", x=10, y = 250, label = "6")+ # annotate("text", x=11, y = 250, label = "3") library(emmeans) cld_eggs = as.data.frame( cld(emmeans(correctedModelGAM, ~ Exp_nest ), Letters = letters ) ) f1_0_data <- merge(f1_0_data, cld_eggs, by = "Exp_nest") FigS2_egg_sel <- ggplot(data = f1_0_data_nomat, aes(x =Exp_nest , y = log(Eggs_sel), fill = Productivity))+ geom_boxplot(outlier.shape = NA)+ #scale_fill_viridis_d()+ scale_fill_manual(values = c("#9595F6", "#FF9494"), labels = c("High", "Low" ))+ geom_quasirandom()+ # labs(title="(a)")+ xlab("Maternal colony")+ ylab("Daughter log(eggs)")+ # ylim(0,350)+ geom_text(aes(y = 6.5, label= eggs_letters))+ theme_minimal()+ theme(legend.position = "none", axis.title = element_text(size=15), axis.text.x = element_text(size = 12), axis.text.y = element_text(size = 12)) Fig2_Aqp_sel / FigS2_egg_sel + plot_annotation(tag_levels = 'A')