-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathFigure9.R
266 lines (197 loc) · 12.7 KB
/
Figure9.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
rm(list=ls())
gc()
library(MCMCglmm)
library(matrixStats)
load("output_files/RData/G_matrices_high_salt.RData")
load('output_files/RData/Gqw_High_Salt.RData')
load('output_files/RData/Gqw_Low_Salt.RData')
angle_theta <- function(x, y) {
dot.prod <- x %*% y
norm.x <- norm(x, type = "2")
norm.y <- norm(y, type = "2")
theta <- 180/pi * as.numeric(acos(dot.prod/(norm.x * norm.y)))
as.numeric(theta)
}
## Opposite sign to match the direction of evolution
load('output_files/RData/SSCP_eigenvectors.RData')
dmax = EV_div[,1]
gmax = eigen(VCV_mat_NaCl[[1]]$G1_mat/2)$vectors[,1]
## Load the coefficients from the Manova
coef_manova <- read.table('output_files/txt/Manova.coefficients.txt',sep="\t")
coef_manova_GA_evolution <- coef_manova[2:4,]
coef_manova_GA_evolution_NGM <- coef_manova[2:4,] +
coef_manova[110:112,] # Add the interaction
avg_NaCl = colMeans(coef_manova_GA_evolution)
avg_NGM = colMeans(coef_manova_GA_evolution_NGM)
# New sampled mean
all_ratios <- NULL; all_ratios_drift <- NULL
all_ratios_NGM <- NULL; all_ratios_drift_NGM <- NULL
for(i in 1:1000){
delta_Z_sampled <- matrix(VCV_with_w_NaCl$VCV_Mat[i,1:64]/2,8,8)[1:7,8]
delta_Z_sampled_NGM <- matrix(VCV_with_w_NGM$VCV_Mat[i,1:64]/2,8,8)[1:7,8]
all_ratios <- rbind(all_ratios,c(
as.numeric(coef_manova_GA_evolution[1,])/delta_Z_sampled,
as.numeric(coef_manova_GA_evolution[2,])/delta_Z_sampled,
as.numeric(coef_manova_GA_evolution[3,])/delta_Z_sampled,
avg_NaCl/delta_Z_sampled))
all_ratios_NGM <- rbind(all_ratios_NGM,c(
as.numeric(coef_manova_GA_evolution_NGM[1,])/delta_Z_sampled_NGM,
as.numeric(coef_manova_GA_evolution_NGM[2,])/delta_Z_sampled_NGM,
as.numeric(coef_manova_GA_evolution_NGM[3,])/delta_Z_sampled_NGM,
avg_NGM/delta_Z_sampled_NGM))
}
avg_gen_change <- mean(colMedians(all_ratios[,8:28])) # 3.5 generations on average
angle_pred_dmax = NULL
angle_pred_gmax = NULL
angle_pred_GA1 = NULL;angle_pred_GA2 = NULL;angle_pred_GA4 = NULL
angle_pred_GA1_NGM = NULL;angle_pred_GA2_NGM = NULL;angle_pred_GA4_NGM = NULL
angle_avg_NaCl = NULL;angle_avg_NGM = NULL
for(i in 1:nrow(VCV_with_w_NaCl$VCV_Mat)){
temp_deltaZ = matrix(VCV_with_w_NaCl$VCV_Mat[i,1:64]/2,8,8)[1:7,8]
temp_deltaZ_NGM = matrix(VCV_with_w_NGM$VCV_Mat[i,1:64]/2,8,8)[1:7,8]
angle_pred_dmax= c(angle_pred_dmax,angle_theta(dmax,temp_deltaZ))
angle_pred_gmax= c(angle_pred_gmax,angle_theta(gmax,temp_deltaZ))
angle_pred_GA1= c(angle_pred_GA1,angle_theta(as.numeric(coef_manova_GA_evolution[1,]),temp_deltaZ))
angle_pred_GA2= c(angle_pred_GA2,angle_theta(as.numeric(coef_manova_GA_evolution[2,]),temp_deltaZ))
angle_pred_GA4= c(angle_pred_GA4,angle_theta(as.numeric(coef_manova_GA_evolution[3,]),temp_deltaZ))
angle_pred_GA1_NGM= c(angle_pred_GA1_NGM,angle_theta(as.numeric(coef_manova_GA_evolution_NGM[1,]),temp_deltaZ_NGM))
angle_pred_GA2_NGM= c(angle_pred_GA2_NGM,angle_theta(as.numeric(coef_manova_GA_evolution_NGM[2,]),temp_deltaZ_NGM))
angle_pred_GA4_NGM= c(angle_pred_GA4_NGM,angle_theta(as.numeric(coef_manova_GA_evolution_NGM[3,]),temp_deltaZ_NGM))
angle_avg_NaCl= c(angle_avg_NaCl,angle_theta(avg_NaCl,temp_deltaZ))
angle_avg_NGM= c(angle_avg_NGM,angle_theta(avg_NGM,temp_deltaZ))
}
angle_Rdm=NULL
for(i in 1:4500){
angle_Rdm=c(angle_Rdm,angle_theta(runif(7,min=(-1),max=1),runif(7,min=(-1),max=1)))
}
all_angles=cbind(angle_pred_GA1,angle_pred_GA2,angle_pred_GA4,angle_pred_GA1_NGM,angle_pred_GA2_NGM,angle_pred_GA4_NGM,angle_avg_NaCl,angle_avg_NGM,angle_Rdm)
## Vect that contains the beta P
vect_betasP_withBL_NaCl <- NULL
vect_betasP_withBL_NaCl_GA1 <- NULL
vect_betasP_withBL_NaCl_GA2 <- NULL
vect_betasP_withBL_NaCl_GA3 <- NULL
vect_betas_withBL_NaCl_no_cov_sampling <- NULL
vect_betas_withBL_NaCl_no_cov_sampling_only5 <- NULL
for(k in 1:nrow(VCV_with_w_NaCl$VCV_Mat)){
temp_Gzw_NaCl <- matrix(VCV_with_w_NaCl$VCV_Mat[k,1:64]/2,8,8)
vect_betas_withBL_NaCl_no_cov_sampling <- rbind(vect_betas_withBL_NaCl_no_cov_sampling ,t(solve(temp_Gzw_NaCl[1:7,1:7]/2)%*%VCV_with_w_NaCl$G1_mat[1:7,8]/2))
vect_betas_withBL_NaCl_no_cov_sampling_only5 <- rbind(vect_betas_withBL_NaCl_no_cov_sampling_only5 ,
t(solve(temp_Gzw_NaCl[c(1,2,4,6,7),c(1,2,4,6,7)]/2)%*%VCV_with_w_NaCl$G1_mat[c(1,2,4,6,7),8]/2))
vect_betasP_withBL_NaCl <- rbind(vect_betasP_withBL_NaCl ,t(solve(temp_Gzw_NaCl[1:7,1:7]/2)%*%avg_NaCl))
vect_betasP_withBL_NaCl_GA1 <- rbind(vect_betasP_withBL_NaCl_GA1 ,t(solve(temp_Gzw_NaCl[1:7,1:7]/2)%*%as.numeric(coef_manova_GA_evolution[1,])))
vect_betasP_withBL_NaCl_GA2 <- rbind(vect_betasP_withBL_NaCl_GA2 ,t(solve(temp_Gzw_NaCl[1:7,1:7]/2)%*%as.numeric(coef_manova_GA_evolution[2,])))
vect_betasP_withBL_NaCl_GA3 <- rbind(vect_betasP_withBL_NaCl_GA3 ,t(solve(temp_Gzw_NaCl[1:7,1:7]/2)%*%as.numeric(coef_manova_GA_evolution[3,])))
}
v_col = c("cadetblue1", "cornflowerblue", "slateblue2")
output_table=NULL
pdf(file='plots/Figure9.pdf',h=8,w=8)
layout(matrix(c(1,1,2,3),2,2),h=c(.8,1))
par(mar=c(5,4,4,2))
plot(solve(VCV_with_w_NaCl$G1_mat[1:7,1:7])%*%(VCV_with_w_NaCl$G1_mat[1:7,8]),ylim=c(-6,4),type="n",ylab="Selection gradients",xlab="",xlim=c(0.5,8),las=1,bty="n",xaxt="n")
axis(side=1,at=1:7,c("SF","SB","FS","FB","BS","BF","Size"))
legend(1,-4.5,c(expression(beta),expression(beta[g])),lwd=2,col=c("blue","gray"),cex=1.4)
#####
temp_95 <- apply(vect_betasP_withBL_NaCl_GA1,2,function(x){HPDinterval(mcmc(x))})/avg_gen_change
temp_80 <- apply(vect_betasP_withBL_NaCl_GA1,2,function(x){HPDinterval(mcmc(x),prob=.8)})/avg_gen_change
arrows(c(1:7)+.2,temp_95[1,],c(1:7)+.2, temp_95[2,],code=3,length=.025,angle=90,col='black')
arrows(c(1:7)+.2,temp_80[1,],c(1:7)+.2, temp_80[2,],code=0,length=.0,angle=90,col=v_col[1],lwd=2)
output_table=cbind(output_table,t(temp_80),t(temp_95))
temp_95 <- apply(vect_betasP_withBL_NaCl_GA2,2,function(x){HPDinterval(mcmc(x))})/avg_gen_change
temp_80 <- apply(vect_betasP_withBL_NaCl_GA2,2,function(x){HPDinterval(mcmc(x),prob=.8)})/avg_gen_change
arrows(c(1:7)+.3,temp_95[1,],c(1:7)+.3, temp_95[2,],code=3,length=.025,angle=90,col='black')
arrows(c(1:7)+.3,temp_80[1,],c(1:7)+.3, temp_80[2,],code=0,length=.0,angle=90,col=v_col[2],lwd=2)
output_table=cbind(output_table,t(temp_80),t(temp_95))
temp_95 <- apply(vect_betasP_withBL_NaCl_GA3,2,function(x){HPDinterval(mcmc(x))})/avg_gen_change
temp_80 <- apply(vect_betasP_withBL_NaCl_GA3,2,function(x){HPDinterval(mcmc(x),prob=.8)})/avg_gen_change
arrows(c(1:7)+.4,temp_95[1,],c(1:7)+.4, temp_95[2,],code=3,length=.025,angle=90,col='black')
arrows(c(1:7)+.4,temp_80[1,],c(1:7)+.4, temp_80[2,],code=0,length=.0,angle=90,col=v_col[3],lwd=2)
output_table=cbind(output_table,t(temp_80),t(temp_95))
# No sampling of the trait covariances with fitness
temp_95 <- apply(vect_betas_withBL_NaCl_no_cov_sampling,2,function(x){HPDinterval(mcmc(x))})
temp_80 <- apply(vect_betas_withBL_NaCl_no_cov_sampling,2,function(x){HPDinterval(mcmc(x),prob=.8)})
arrows(c(1:7)-.2,temp_95[1,],c(1:7)-.2, temp_95[2,],code=3,length=.05,angle=90,col='black')
arrows(c(1:7)-.2,temp_80[1,],c(1:7)-.2, temp_80[2,],code=0,length=.0,angle=90,col=c(rep("grey",6),"grey"),lwd=2)
output_table=cbind(output_table,t(temp_80),t(temp_95))
temp_Post <- apply(vect_betas_withBL_NaCl_no_cov_sampling,2,function(x){posterior.mode(mcmc(x))})
points(1:7-.2, temp_Post,pch=16,col='black')
output_table=cbind(output_table,temp_Post)
abline(h=0,lty=2)
mtext(side=3,"A",at=0,cex=1.2)
par(mar=c(5,4,1,2))
plot(rep(45,4),1:4,xlim=c(0,180),ylim=c(0.5,6.5),bty="n",las=1,pch=21,bg="grey",type="n",
xlab="Angle (°)",yaxt="n",ylab="",xaxt="n")
axis(1,at=c(0,45,90,135,180))
axis(2,at=c(1,2,3,5.5),c("GA150","GA250","GA450","Random"),las=1)
abline(v=90,lty=2)
temp_arrows <- apply(all_angles,2,function(x){ HPDinterval(as.mcmc(x))})
temp_arrows_80 <- apply(all_angles,2,function(x){ HPDinterval(as.mcmc(x),prob=0.83)})
arrows(temp_arrows[1,c(1:6,9)],c(1:3-.1,1:3+.1,5.5),temp_arrows[2,c(1:6,9)],c(1:3-.1,1:3+.1,5.5),angle=90,length=.05,code=3)
arrows(temp_arrows_80[1,c(1:6,9)],c(1:3-.1,1:3+.1,5.5), temp_arrows_80[2,c(1:6,9)],c(1:3-.1,1:3+.1,5.5),angle=90,length=.05,code=0,lwd=2,col=c(rep(v_col,2),"gray"))
points(apply(all_angles[,1:6],2,function(x){posterior.mode(as.mcmc(x))}),c(1:3-.1,1:3+.1),pch=rep(c(21,8),each=3),bg="grey",cex=1.3)
legend(95,3,c("Low Salt","High Salt"),pch=c(8,21))
mtext(side=3,"B",at=0,cex=1.2,padj=1)
par(mar=c(5,4,1,2))
plot(apply(all_ratios,2,function(x){posterior.mode(as.mcmc(x))}),rep(1:7,4),xlim=c(-2,9),ylim=c(0.5,8),bty="n",las=1,pch=21,bg="grey",type="n",
xlab="Ratio Observed/Predicted divergence",yaxt="n",ylab="")
axis(2,at=c(1:7)+.3,c("SF","SB","FS","FB","BS","BF","Size"))
lines(c(0,0),c(.8,7.5),lty=2)
temp_arrows <- apply(all_ratios,2,function(x){ HPDinterval(as.mcmc(x))})
temp_arrows_80 <- apply(all_ratios,2,function(x){ HPDinterval(as.mcmc(x),prob=0.8)})
arrows(temp_arrows[1,8:28],rep(1:7)+rep(c(0,.1,.2),each=7),temp_arrows[2,8:28],rep(1:7)+rep(c(0,.1,.2),each=7),angle=90,length=.05,code=3,lwd=rep(c(1,1,1,1.5),each=7))
arrows(temp_arrows_80[1,8:28],rep(1:7)+rep(c(0,.1,.2),each=7), temp_arrows_80[2,8:28],rep(1:7)+rep(c(0,.1,.2),each=7),angle=90,length=.05,code=0,lwd=rep(c(2,2,2,3),each=7),col=rep(c(v_col),each=7))
points(apply(all_ratios[,8:28],2,function(x){median(x)}),rep(1:7,3)+rep(c(0,.1,.2),each=7),pch=21,bg="grey")
mtext(side=3,"C",at=-3,cex=1.2)
dev.off()
# Output tables
colnames(output_table)=c("GA150 - 83% CI (lower)","GA150 - 83% CI (upper)",
"GA150 - 95% CI (lower)","GA150 - 95% CI (upper)",
"GA250 - 83% CI (lower)","GA250 - 83% CI (upper)",
"GA250 - 95% CI (lower)","GA250 - 95% CI (upper)",
"GA450 - 83% CI (lower)","GA450 - 83% CI (upper)",
"GA450 - 95% CI (lower)","GA450 - 95% CI (upper)",
"Beta G - 83% CI (lower)","Beta G - 83% CI (upper)",
"Beta G - 95% CI (lower)","Beta G - 95% CI (upper)",
"Beta G - posterior mode")
write.table(output_table,file='output_files/txt/Selection_gradients.txt',row.names=FALSE,sep='\t',quote=FALSE)
pdf(file='plots/Figure9_figure_supplement1.pdf',w=5)
v_col = c("cadetblue1", "cornflowerblue", "slateblue2")
plot(solve(VCV_with_w_NaCl$G1_mat[1:7,1:7])%*%(VCV_with_w_NaCl$G1_mat[1:7,8]),ylim=c(-2,2),type="n",ylab="Selection gradients",xlab="",xlim=c(0.5,8),las=1,bty="n",xaxt="n")
axis(side=1,at=1:7,c("SF","SB","FS","FB","BS","BF","Size"))
temp_95 <- apply(vect_betas_withBL_NaCl_no_cov_sampling,2,function(x){HPDinterval(mcmc(x))})
temp_80 <- apply(vect_betas_withBL_NaCl_no_cov_sampling,2,function(x){HPDinterval(mcmc(x),prob=.8)})
arrows(c(1:7)-.2,temp_95[1,],c(1:7)-.2, temp_95[2,],code=3,length=.05,angle=90,col='black')
arrows(c(1:7)-.2,temp_80[1,],c(1:7)-.2, temp_80[2,],code=0,length=.0,angle=90,col=c(rep("grey",6),"grey"),lwd=2)
temp_Post <- apply(vect_betas_withBL_NaCl_no_cov_sampling,2,function(x){posterior.mode(mcmc(x))})
points(1:7-.2, temp_Post,pch=16,col='black')
temp_95 <- apply(vect_betas_withBL_NaCl_no_cov_sampling_only5,2,function(x){HPDinterval(mcmc(x))})
temp_80 <- apply(vect_betas_withBL_NaCl_no_cov_sampling_only5,2,function(x){HPDinterval(mcmc(x),prob=.8)})
arrows(c(1,2,4,6,7)+.2,temp_95[1,],c(1,2,4,6,7)+.2, temp_95[2,],code=3,length=.05,angle=90,col='black')
arrows(c(1,2,4,6,7)+.2,temp_80[1,],c(1,2,4,6,7)+.2, temp_80[2,],code=0,length=.0,angle=90,col=c(rep("firebrick3",6),"firebrick3"),lwd=2)
temp_Post <- apply(vect_betas_withBL_NaCl_no_cov_sampling_only5,2,function(x){posterior.mode(mcmc(x))})
points(c(1,2,4,6,7)+.2, temp_Post,pch=16,col='firebrick3')
abline(h=0,lty=2)
dev.off()
vect_betas_NaCl <- NULL
for(k in 1:nrow(VCV_with_w_NaCl$VCV_Mat)){
temp_Gzw_NaCl <- matrix(VCV_with_w_NaCl$VCV_Mat[k,1:64]/2,8,8)
vect_betas_NaCl <- rbind(vect_betas_NaCl ,t(solve(temp_Gzw_NaCl[1:7,1:7]/2)%*%(temp_Gzw_NaCl[1:7,8]/2)))
}
pdf(file='plots/Figure9_figure_supplement2.pdf',w=5)
plot(solve(VCV_with_w_NaCl$G1_mat[1:7,1:7])%*%(VCV_with_w_NaCl$G1_mat[1:7,8]),ylim=c(-1.5,1.5),type="n",ylab=expression(paste("Selection gradient ",beta[g])),xlab="",xlim=c(0.5,7),xaxt="n",las=1,bty="n")
axis(side=1,c("SF","SB","FS","FB","BS","BF","Size"),at=1:7)
temp_95 <- apply(vect_betas_NaCl,2,function(x){HPDinterval(mcmc(x))})
temp_80 <- apply(vect_betas_NaCl,2,function(x){HPDinterval(mcmc(x),prob=.8)})
arrows(c(1:7),temp_95[1,],c(1:7), temp_95[2,],code=3,length=.05,angle=90,col='black')
arrows(c(1:7),temp_80[1,],c(1:7), temp_80[2,],code=0,length=.0,angle=90,col="gray",lwd=2)
temp_Post <- apply(vect_betas_NaCl,2,function(x){posterior.mode(mcmc(x))})
points(1:7, temp_Post,pch=16,col='black')
abline(h=0,lty=2)
dev.off()
## Export the data in a table :
output_df = data.frame(cbind(temp_Post,t(temp_80),t(temp_95)))
rownames(output_df)=NULL
output_df$trait=c(c("SF","SB","FS","FB","BS","BF","Size"))
output_df=output_df[,c(6,1:5)]
names(output_df)[2:6]=c("Post.mode","lower_83","upper_83","lower_95","upper_95")
output_df[,2:6]=round(output_df[,2:6],digits=3)
write.table(output_df,file=("output_files/txt/Selection_gradients_bg.txt"),sep='\t',quote=FALSE,row.names=FALSE)