-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy path09_halibut_RDA_envdata
225 lines (168 loc) · 12 KB
/
09_halibut_RDA_envdata
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
#RDA by environment ####
#Recode as RAW for RDA
library(data.table)
library(dplyr)
library(ggplot2)
library(vegan)
library(ggman)
library(qvalue)
system("~/Desktop/Software/plink_mac_20200219/plink --file Halibut_finalgenome_mindp15_24chrom_ming90_HWE_noWF_allENV --recodeA --out Halibut_genome2_d15gt90HWE_withENV --chr-set 24")
ENV_Metadata <- read.delim("Metadata_Halibut_ENV_noNAnoWF.txt", stringsAsFactors = F)
GENO_RAW <- fread("Halibut_genome2_d15gt90HWE_withENV.raw",stringsAsFactors = F, data.table = F)
colnames(GENO_RAW)[1] <- "ID"
GENO_METADATA_SORT <- inner_join(ENV_Metadata, GENO_RAW)
METADATA_PHENO_ENV_SORT <- GENO_METADATA_SORT[1:33]
GENO_SORT <- GENO_METADATA_SORT[39:length(GENO_METADATA_SORT)]
#impute
GENO_SORT_IMP <- apply(GENO_SORT, 2, function(x) replace(x, is.na(x), as.numeric(names(which.max(table(x))))))
#check env variable correlation - remove if > 0.7
cor.test(METADATA_PHENO_ENV_SORT$TEMP_PC, METADATA_PHENO_ENV_SORT$O2_PC)
cor.test(METADATA_PHENO_ENV_SORT$TEMP_PC, METADATA_PHENO_ENV_SORT$SALT_PC)
cor.test(METADATA_PHENO_ENV_SORT$O2_PC, METADATA_PHENO_ENV_SORT$SALT_PC)
#nope.
#Biplot environmental variables
TEMPS <- data.frame((cbind(METADATA_PHENO_ENV_SORT$BO2_tempmean_bdmean_sd, METADATA_PHENO_ENV_SORT$BO2_tempmin_bdmean_sd, METADATA_PHENO_ENV_SORT$BO2_tempmax_bdmean_sd)))
colnames(TEMPS) <- c("temp_mean", "temp_min", "temp_max")
TEMPPC <- prcomp(TEMPS)
library(ggbiplot)
ggbiplot(pcobj = TEMPPC, groups = METADATA_PHENO_ENV_SORT$Region) + theme_classic()
mtcars.pca <- prcomp(mtcars)
O2S <- data.frame((cbind(METADATA_PHENO_ENV_SORT$BO2_dissoxmean_bdmean_sd, METADATA_PHENO_ENV_SORT$BO2_dissoxmin_bdmean_sd, METADATA_PHENO_ENV_SORT$BO2_dissoxmax_bdmean_sd)))
colnames(O2S) <- c("O2_mean", "O2_min", "O2_max")
O2PC <- prcomp(O2S)
ggbiplot(pcobj = O2PC, groups = METADATA_PHENO_ENV_SORT$Region) + theme_classic()
SALTS<- data.frame((cbind(METADATA_PHENO_ENV_SORT$BO2_salinitymean_bdmean_sd, METADATA_PHENO_ENV_SORT$BO2_salinitymin_bdmean_sd, METADATA_PHENO_ENV_SORT$BO2_salinitymax_bdmean_sd)))
colnames(SALTS) <- c("Salt_mean", "Salt_min", "Salt_max")
SALTPC <- prcomp(SALTS)
ggbiplot(pcobj = SALTPC, groups = METADATA_PHENO_ENV_SORT$Region) + theme_classic()
#Environmental RDA
ALL_ENVPC_RDA <- rda(GENO_SORT_IMP ~ TEMP_PC + O2_PC + SALT_PC, data = METADATA_PHENO_ENV_SORT)
#condition on geography
ALL_ENVPC_RDA_cond <- rda(GENO_SORT_IMP ~ TEMP_PC + O2_PC + SALT_PC + Condition(Lat + Lon), data = METADATA_PHENO_ENV_SORT)
#RDA with just geographic info
ALL_ENVPC_RDA_dists <- rda(GENO_SORT_IMP ~ Lat + Lon, data = METADATA_PHENO_ENV_SORT)
#adjusted R2, Significance testing, screeplots
RsquareAdj(ALL_ENVPC_RDA)
RsquareAdj(ALL_ENVPC_RDA_cond)
RsquareAdj(ALL_ENVPC_RDA_dists)
anova.cca(ALL_ENVPC_RDA, parallel= 8,by = "terms")
anova.cca(ALL_ENVPC_RDA_cond, parallel= 8,by = "terms")
anova.cca(ALL_ENVPC_RDA_dists, parallel= 8,by = "terms")
screeplot(ALL_ENVPC_RDA)
screeplot(ALL_ENVPC_RDA_cond)
screeplot(ALL_ENVPC_RDA_dists)
#Get proportion variance per axis
summary(eigenvals(ALL_ENVPC_RDA, model = "constrained"))
summary(eigenvals(ALL_ENVPC_RDA_cond, model = "constrained"))
summary(eigenvals(ALL_ENVPC_RDA_dists, model = "constrained"))
#plot RDA biplot
bg <- c("dodgerblue", "red", "purple", "cyan", "orange", "goldenrod", "orchid")
bg2 <- colorRampPalette(c("dodgerblue", "red", "purple", "cyan", "orange", "goldenrod", "orchid"))(n=35)
eco <- as.factor(METADATA_PHENO_ENV_SORT$Region)
plot(ALL_ENVPC_RDA, type="n", scaling=3)
points(ALL_ENVPC_RDA, display="species", pch=20, cex=0.7, col="gray32", scaling=3) # the SNPs
points(ALL_ENVPC_RDA, display="sites", pch=21, cex=1.3, col="gray32", scaling=3, bg=bg[eco]) # the fish
text(ALL_ENVPC_RDA, scaling=3, display="bp", col="#0868ac", cex=1) # the predictors
legend("bottom", legend=levels(eco), bty="n", col="gray32", pch=21, cex=0.5, pt.bg=bg)
cor.test(METADATA_PHENO_ENV_SORT$TEMP_PC, METADATA_PHENO_ENV_SORT$O2_PC)
#plot axis 1 and 3
bg <- c("dodgerblue", "red", "purple", "cyan", "orange", "goldenrod", "orchid")
bg2 <- colorRampPalette(c("dodgerblue", "red", "purple", "cyan", "orange", "goldenrod", "orchid"))(n=35)
eco <- as.factor(METADATA_PHENO_ENV_SORT$Region)
plot(ALL_ENVPC_RDA, type="n", scaling=3, choices = c(1,3))
points(ALL_ENVPC_RDA, display="species", pch=20, cex=0.7, col="gray32", scaling=3, choices = c(1,3)) # the SNPs
points(ALL_ENVPC_RDA, display="sites", pch=21, cex=1.3, col="gray32", scaling=3, bg=bg[eco], choices = c(1,3)) # the fish
text(ALL_ENVPC_RDA, scaling=3, display="bp", col="#0868ac", cex=1, choices = c(1,3)) # the predictors
legend("bottom", legend=levels(eco), bty="n", col="gray32", pch=21, cex=0.5, pt.bg=bg)
library(robust)
library(qvalue)
#rdadapt
rdadapt<-function(rda,K)
{
loadings<-rda$CCA$v[,1:as.numeric(K)]
resscale <- apply(loadings, 2, scale)
resmaha <- covRob(resscale, distance = TRUE, na.action= na.omit, estim="pairwiseGK")$dist
lambda <- median(resmaha)/qchisq(0.5,df=K)
reschi2test <- pchisq(resmaha/lambda,K,lower.tail=FALSE)
qval <- qvalue(reschi2test)
q.values_rdadapt<-qval$qvalues
return(data.frame(p.values=reschi2test, q.values=q.values_rdadapt))
}
RDADPT_K3 <- rdadapt(ALL_ENVPC_RDA, K = 3)
#Get SNP scores
ENV_RDA_SNPSCORES <- data.frame(ALL_ENVPC_RDA$CCA$v[,1:3], stringsAsFactors = F)
SNPnames <- as.character(rownames(ENV_RDA_SNPSCORES))
library(rsed)
ENV_RDA_SNPSCORES$SNP <- as.character(sed_substitute(SNPnames, "_.*", ""))
Chrom_map <- read.delim("Halibut_genome2_d15gt90HWE_withENV_12.map", stringsAsFactors = F, header = F)
colnames(Chrom_map) <- c("Chrom", "SNP", "CM", "BP")
ENV_RDA_SNPSCORES_MAPPED <- inner_join(Chrom_map, ENV_RDA_SNPSCORES)
ENV_RDA_SNPSCORES_MAPPED$RDA1_abs <- abs(as.numeric(as.character(ENV_RDA_SNPSCORES$RDA1)))
ENV_RDA_SNPSCORES_MAPPED$RDA2_abs <- abs(as.numeric(as.character(ENV_RDA_SNPSCORES$RDA2)))
ENV_RDA_SNPSCORES_MAPPED$RDA3_abs <- abs(as.numeric(as.character(ENV_RDA_SNPSCORES$RDA3)))
ENV_RDA_SNPSCORES_MAPPED <- data.frame(cbind(ENV_RDA_SNPSCORES_MAPPED, RDADPT_K3), stringsAsFactors = F)
class(ENV_RDA_SNPSCORES_MAPPED$q.values)
RDA_ADAPT_Q01 <- ENV_RDA_SNPSCORES_MAPPED[which(ENV_RDA_SNPSCORES_MAPPED$q.values < 0.01),]
RDA_ADAPT_Q05 <- ENV_RDA_SNPSCORES_MAPPED[which(ENV_RDA_SNPSCORES_MAPPED$q.values < 0.05),]
write.table(RDA_ADAPT_Q05, "RDA_ADAPT_Q05.txt", sep = "\t", col.names = F, row.names = F, quote = F)
write.table(RDA_ADAPT_Q05$SNP, "RDA_ADAPT_Q05_SNPs.txt", sep = "\t", col.names = F, row.names = F, quote = F)
write.table(RDA_ADAPT_Q01, "RDA_ADAPT_Q01.txt", sep = "\t", col.names = F, row.names = F, quote = F)
write.table(RDA_ADAPT_Q01$SNP, "RDA_ADAPT_Q01_SNPs.txt", sep = "\t", col.names = F, row.names = F, quote = F)
#Plot RDA scores
ggman(ENV_RDA_SNPSCORES_MAPPED, chrom = "Chrom", pvalue = "q.values", snp = "SNP", bp="BP", pointSize = 1, title = "Halibut", xlabel = "Chromosome", logTransform = T, ylabel = "RDA1", sigLine = -log10(0.05) ) + theme_classic()
#Get outliers for each axis 99.9%
RDA1_ENV_OL <- ENV_RDA_SNPSCORES_MAPPED[which(ENV_RDA_SNPSCORES_MAPPED$RDA1_abs > quantile(x = ENV_RDA_SNPSCORES_MAPPED$RDA1_abs, 0.999 )),]
RDA2_ENV_OL <- ENV_RDA_SNPSCORES_MAPPED[which(ENV_RDA_SNPSCORES_MAPPED$RDA2_abs > quantile(x = ENV_RDA_SNPSCORES_MAPPED$RDA2_abs, 0.999 )),]
RDA3_ENV_OL <- ENV_RDA_SNPSCORES_MAPPED[which(ENV_RDA_SNPSCORES_MAPPED$RDA3_abs > quantile(x = ENV_RDA_SNPSCORES_MAPPED$RDA3_abs, 0.999 )),]
RDA_999OL <- unique(c(RDA1_ENV_OL$SNP, RDA2_ENV_OL$SNP, RDA3_ENV_OL$SNP))
RDA_999OL_RDADAPT <- ENV_RDA_SNPSCORES_MAPPED[ENV_RDA_SNPSCORES_MAPPED$SNP %in% RDA_999OL,]
RDA_999OL_RDADAPT <- RDA_999OL_RDADAPT[which(RDA_999OL_RDADAPT$q.values < 0.05),]
write.table(RDA_999OL_RDADAPT, "RDA_999OL_RDADAPT.txt", sep = "\t", col.names = F, row.names = F, quote = F)
write.table(RDA_999OL_RDADAPT$SNP, "RDA_999OL_RDADAPT_SNPs.txt", sep = "\t", col.names = F, row.names = F, quote = F)
#Conditioned by geography
ALL_ENVPC_RDA_cond <- rda(GENO_SORT_IMP ~ TEMP_PC + O2_PC + SALT_PC + Condition(Lat + Lon), data = METADATA_PHENO_ENV_SORT)
#plot
bg <- c("dodgerblue", "red", "purple", "cyan", "orange", "goldenrod", "orchid")
bg2 <- colorRampPalette(c("dodgerblue", "red", "purple", "cyan", "orange", "goldenrod", "orchid"))(n=35)
eco <- as.factor(METADATA_PHENO_ENV_SORT$Region)
plot(ALL_ENVPC_RDA_cond, type="n", scaling=3)
points(ALL_ENVPC_RDA_cond, display="species", pch=20, cex=0.7, col="gray32", scaling=3) # the SNPs
points(ALL_ENVPC_RDA_cond, display="sites", pch=21, cex=1.3, col="gray32", scaling=3, bg=bg[eco]) # the fish
text(ALL_ENVPC_RDA_cond, scaling=3, display="bp", col="#0868ac", cex=1) # the predictors
legend("bottom", legend=levels(eco), bty="n", col="gray32", pch=21, cex=0.5, pt.bg=bg)
#plot axis 1 and 3
bg <- c("dodgerblue", "red", "purple", "cyan", "orange", "goldenrod", "orchid")
bg2 <- colorRampPalette(c("dodgerblue", "red", "purple", "cyan", "orange", "goldenrod", "orchid"))(n=35)
eco <- as.factor(METADATA_PHENO_ENV_SORT$Region)
plot(ALL_ENVPC_RDA_cond, type="n", scaling=3, choices = c(1,3))
points(ALL_ENVPC_RDA_cond, display="species", pch=20, cex=0.7, col="gray32", scaling=3, choices = c(1,3)) # the SNPs
points(ALL_ENVPC_RDA_cond, display="sites", pch=21, cex=1.3, col="gray32", scaling=3, bg=bg[eco], choices = c(1,3)) # the fish
text(ALL_ENVPC_RDA_cond, scaling=3, display="bp", col="#0868ac", cex=1, choices = c(1,3)) # the predictors
legend("bottom", legend=levels(eco), bty="n", col="gray32", pch=21, cex=0.5, pt.bg=bg)
RDADPT_K3_cond <- rdadapt(ALL_ENVPC_RDA_cond, K = 3)
#Get SNP scores
ENV_RDA_SNPSCORES_cond <- data.frame(ALL_ENVPC_RDA_cond$CCA$v[,1:3], stringsAsFactors = F)
SNPnames <- as.character(rownames(ENV_RDA_SNPSCORES_cond))
library(rsed)
ENV_RDA_SNPSCORES_cond$SNP <- as.character(sed_substitute(SNPnames, "_.*", ""))
Chrom_map <- read.delim("Halibut_genome2_d15gt90HWE_withENV_12.map", stringsAsFactors = F, header = F)
colnames(Chrom_map) <- c("Chrom", "SNP", "CM", "BP")
ENV_RDA_SNPSCORES_MAPPED_cond <- inner_join(Chrom_map, ENV_RDA_SNPSCORES_cond)
ENV_RDA_SNPSCORES_MAPPED_cond$RDA1_abs <- abs(as.numeric(as.character(ENV_RDA_SNPSCORES_cond$RDA1)))
ENV_RDA_SNPSCORES_MAPPED_cond$RDA2_abs <- abs(as.numeric(as.character(ENV_RDA_SNPSCORES_cond$RDA2)))
ENV_RDA_SNPSCORES_MAPPED_cond$RDA3_abs <- abs(as.numeric(as.character(ENV_RDA_SNPSCORES_cond$RDA3)))
ENV_RDA_SNPSCORES_MAPPED_cond <- data.frame(cbind(ENV_RDA_SNPSCORES_MAPPED_cond, RDADPT_K3_cond), stringsAsFactors = F)
class(ENV_RDA_SNPSCORES_MAPPED_cond$q.values)
RDA_ADAPT_Q01_cond <- ENV_RDA_SNPSCORES_MAPPED_cond[which(ENV_RDA_SNPSCORES_MAPPED_cond$q.values < 0.01),]
RDA_ADAPT_Q05_cond <- ENV_RDA_SNPSCORES_MAPPED_cond[which(ENV_RDA_SNPSCORES_MAPPED_cond$q.values < 0.05),]
#Plot RDA scores
ggman(ENV_RDA_SNPSCORES_MAPPED_cond, chrom = "Chrom", pvalue = "q.values", snp = "SNP", bp="BP", pointSize = 1, title = "Halibut", xlabel = "Chromosome", logTransform = T, ylabel = "RDA1", sigLine = -log10(0.05) ) + theme_classic()
ggman(ENV_RDA_SNPSCORES_MAPPED, chrom = "Chrom", pvalue = "q.values", snp = "SNP", bp="BP", pointSize = 1, title = "Halibut", xlabel = "Chromosome", logTransform = T, ylabel = "RDA1", sigLine = -log10(0.05) ) + theme_classic()
#Get outliers for each axis 99.9%
RDA1_ENV_OL_cond <- ENV_RDA_SNPSCORES_MAPPED_cond[which(ENV_RDA_SNPSCORES_MAPPED_cond$RDA1_abs > quantile(x = ENV_RDA_SNPSCORES_MAPPED_cond$RDA1_abs, 0.999 )),]
RDA2_ENV_OL_cond <- ENV_RDA_SNPSCORES_MAPPED_cond[which(ENV_RDA_SNPSCORES_MAPPED_cond$RDA2_abs > quantile(x = ENV_RDA_SNPSCORES_MAPPED_cond$RDA2_abs, 0.999 )),]
RDA3_ENV_OL_cond <- ENV_RDA_SNPSCORES_MAPPED_cond[which(ENV_RDA_SNPSCORES_MAPPED_cond$RDA3_abs > quantile(x = ENV_RDA_SNPSCORES_MAPPED_cond$RDA3_abs, 0.999 )),]
RDA_999OL_cond <- unique(c(RDA1_ENV_OL_cond$SNP, RDA2_ENV_OL_cond$SNP, RDA3_ENV_OL_cond$SNP))
RDA_999OL_RDADAPT_cond <- ENV_RDA_SNPSCORES_MAPPED_cond[ENV_RDA_SNPSCORES_MAPPED_cond$SNP %in% RDA_999OL_cond,]
RDA_999OL_RDADAPT_cond <- RDA_999OL_RDADAPT_cond[which(RDA_999OL_RDADAPT_cond$q.values < 0.05),]
RDA_OL_both <- RDA_999OL_RDADAPT_cond$SNP[RDA_999OL_RDADAPT_cond$SNP %in% RDA_999OL_RDADAPT$SNP]