-
Notifications
You must be signed in to change notification settings - Fork 3
/
Cohesion_Analysis.R
425 lines (325 loc) · 15.3 KB
/
Cohesion_Analysis.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
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
### Processing the cohesion data
# RED 2017; [email protected]
# Loading in packages
library(reshape2)
library(ggplot2)
library(vegan)
library(ecodist)
library(Hmisc)
####################################
### Loading and sorting the data ###
####################################
# Setting work directory and loading data
setwd("~/Documents/Wilkins Lab/Ohio EPA Project/Sequencing Data/16S Data/Ecological Modeling Files/")
data = read.csv("Total_0.2_Cohesion.csv", row.names = 1)
comm = read.table("Raw_EPA_no_pynast.txt", row.names = 1, header = T)
fac.file = read.csv("Family_Level_OTU_edit.csv")
w.uni = read.csv("../Weighted_Uni_output.csv", row.names = 1)
uw.uni = read.csv("../Unweighted_Uni_output.csv", row.names = 1)
bnti = read.csv("MC_EPA_bNTI.csv", row.names = 1)
# Storing the factors
factors = t(fac.file[1:5,])
colnames(factors) = factors[1,]
factors = factors[-1,]
# General data scrubbing
comm = comm[,-which(colnames(comm) %in% c("YB04","BL01","BL02"))] # Removing YB04 becuase it was a failed sequencing run
colnames(comm) = colnames(fac.file[,-1])
# # Removing the factors from the community matrix and bnti dataset
# comm = comm[-1:-5,]
bnti = bnti[,-1:-4]
# # Fixing the data to be numbers
# comm[,-1] = apply(comm[,-1], 2, function(x) as.numeric(as.character(x)))
#
# # Aggregating the data because duplicates need to be removed
# comm = aggregate(.~OTU.ID, data = comm, FUN = sum)
#
# # Setting row names on the newly aggregated data
# row.names(comm) = comm[,1]
# comm = comm[,-1]
#
# # Transposing the community matrix
comm = t(comm)
# Fixing Dates
data$Time = gsub("_","/1/",data$Time)
data$Time = as.Date(data$Time, "%m/%d/%Y")
rm("fac.file")
######################################
#### Averaging duplicated samples ####
######################################
x = unique(factors[,1])
comm.avg = matrix(nrow = length(x), ncol = length(comm[1,]))
# Generating the averaged Family level table
for(i in 1:length(x)){
w = which(factors[,1] %in% x[i])
if(length(w)==1){
comm.avg[i,] = comm[w,]
} else{
comm.avg[i,] = apply(comm[w,], 2, mean)
}
}
row.names(comm.avg) = x
colnames(comm.avg) = colnames(comm)
comm.avg = as.data.frame(comm.avg)
comm = comm.avg
# Generating the corresponding factors sheet
w = which(factors[,1] %in% x[1])[1]
factors.avg = factors[w,]
for(i in 2:length(x)){
w = which(factors[,1] %in% x[i])[1]
factors.avg = rbind(factors.avg, factors[w,])
}
row.names(factors.avg) = factors.avg[,1]
factors.avg = factors.avg[,-1]
factors = as.data.frame(factors.avg)
rm('comm.avg', 'factors.avg')
##################################
### Plotting the cohesion data ###
##################################
# Plotting the data by time
ggplot(data = data, aes(x = Time, y = Positive_Cohesion))+
geom_point(aes(colour = Well))+
geom_line(aes(colour = Well))+
theme_bw()+
ggtitle("Positive Cohesion thru Time")
ggplot(data = data, aes(x = Time, y = Negative_Cohesion))+
geom_point(aes(colour = Well))+
geom_line(aes(colour = Well))+
theme_bw()+
ggtitle("Negative Cohesion thru Time")
# Generating summary boxplots for Cohesion data
ggplot(data = data, aes(x = Well, y = Positive_Cohesion))+
geom_boxplot(aes(group = Well, fill = Well))+
theme_bw()+
ggtitle("Positive Cohesion by Well")
ggplot(data = data, aes(x = Well, y = Negative_Cohesion))+
geom_boxplot(aes(group = Well, fill = Well))+
theme_bw()+
ggtitle("Negative Cohesion by Well")
#########################################################
### Relating the cohesion value to community turnover ###
#########################################################
# Removing the 0.1 fraction of the data
w = c(which(factors$Filter %in% "0.1"), which(factors$Well %in% "Licking-PW"))
comm = comm[-w,]
factors = factors[-w,]
bnti = bnti[-w,-w]
# Generating a Bray-Curtis distance matrix
comm = as.matrix(vegdist(x = comm, method = "bray"))
#--------------------------------------------------------------------#
### Attempting to plot Bray-Curtis and bNTI by Cohesion
# Plotting Positive
x = NULL
y = NULL
w = grep("Athens", row.names(data))
ww = grep("Athens", row.names(comm))
for(i in 1:length(ww)){
x = rbind(x, cbind(data$Positive_Cohesion[w], comm[ww,ww[i]]))
y = rbind(y, cbind(data$Positive_Cohesion[w], bnti[ww,ww[i]]))
}
w = grep("Greene", row.names(data))
ww = grep("Greene", row.names(comm))
for(i in 1:length(ww)){
x = rbind(x, cbind(data$Positive_Cohesion[w], comm[ww,ww[i]]))
y = rbind(y, cbind(data$Positive_Cohesion[w], bnti[ww,ww[i]]))
}
w = grep("Licking", row.names(data))
ww = grep("Licking", row.names(comm))
for(i in 1:length(ww)){
x = rbind(x, cbind(data$Positive_Cohesion[w], comm[ww,ww[i]]))
y = rbind(y, cbind(data$Positive_Cohesion[w], bnti[ww,ww[i]]))
}
x = x[-which(x[,2] == 0),]
y = y[-is.na(y[,2]),]
plot(x[,1], x[,2], main = "Bray-Curtis by Positive Cohesion")
abline(lm(x[,2]~x[,1]))
corr1 = rcorr(x[,1], x[,2])
legend("bottomright", sprintf("r-value: %s, p-value: %s", corr1$r[1,2], corr1$P[1,2]))
plot(y[,1], y[,2], main = "bNTI by Positive Cohesion")
abline(lm(y[,2]~y[,1]))
corr2 = rcorr(y[,1], y[,2])
legend("bottomright", sprintf("r-value: %s, p-value: %s", corr2$r[1,2], corr2$P[1,2]))
boxplot(x[,2]~x[,1])
boxplot(y[,2]~y[,1])
# Plotting Negative
x = NULL
y = NULL
w = grep("Athens", row.names(data))
ww = grep("Athens", row.names(comm))
for(i in 1:length(ww)){
x = rbind(x, cbind(data$Negative_Cohesion[w], comm[ww,ww[i]]))
y = rbind(y, cbind(data$Negative_Cohesion[w], bnti[ww,ww[i]]))
}
w = grep("Greene", row.names(data))
ww = grep("Greene", row.names(comm))
for(i in 1:length(ww)){
x = rbind(x, cbind(data$Negative_Cohesion[w], comm[ww,ww[i]]))
y = rbind(y, cbind(data$Negative_Cohesion[w], bnti[ww,ww[i]]))
}
w = grep("Licking", row.names(data))
ww = grep("Licking", row.names(comm))
for(i in 1:length(ww)){
x = rbind(x, cbind(data$Negative_Cohesion[w], comm[ww,ww[i]]))
y = rbind(y, cbind(data$Negative_Cohesion[w], bnti[ww,ww[i]]))
}
x = x[-which(x[,2] == 0),]
y = y[-which(is.na(y[,2]) %in% TRUE),]
plot(x[,1], x[,2], main = "Bray-Curtis by Negative Cohesion")
abline(lm(x[,2]~x[,1]))
corr1 = rcorr(x[,1], x[,2])
legend("bottomright", sprintf("r-value: %s, p-value: %s", corr1$r[1,2], corr1$P[1,2]))
plot(y[,1], y[,2], main = "bNTI by Negative Cohesion")
abline(lm(y[,2]~y[,1]))
corr2 = rcorr(y[,1], y[,2])
legend("bottomright", sprintf("r-value: %s, p-value: %s", corr2$r[1,2], corr2$P[1,2]))
boxplot(x[,2]~x[,1])
boxplot(y[,2]~y[,1])
rm("x", "y")
#######################################################
### Working on some different statistics approaches ###
#######################################################
### Attempting to use Mantel to solve my problem ###
# Using difference in cohesion between samples to relate it to bNTI and Bray-Curtis
bnti[upper.tri(bnti)]=t(bnti)[upper.tri(bnti)]
data = data[match(row.names(bnti), row.names(data)),] # Reordering data to match bNTI and Bray-Curtis
pos.coh = abs(data$Positive_Cohesion) # Absolute value for positive cohesion values (shouldn't matter but doing this anyway)
pos.coh = abs(outer(pos.coh, pos.coh, "-")) # Generating matrix of differences
row.names(pos.coh) = row.names(data)
colnames(pos.coh) = row.names(data)
neg.coh = abs(data$Negative_Cohesion) # Absolute value for positive cohesion values (shouldn't matter but doing this anyway)
neg.coh = abs(outer(neg.coh, neg.coh, "-")) # Generating matrix of differences
row.names(neg.coh) = row.names(data)
colnames(neg.coh) = row.names(data)
# Total Mantel analysis
mant = mantel(formula = as.dist(bnti)~as.dist(neg.coh), nperm = 9999, mrank = T)
mant = rbind(mant, mantel(formula = as.dist(bnti)~as.dist(pos.coh), nperm = 9999, mrank = T))
mant = rbind(mant, mantel(formula = as.dist(uw.uni)~as.dist(neg.coh), nperm = 9999, mrank = T))
mant = rbind(mant, mantel(formula = as.dist(uw.uni)~as.dist(pos.coh), nperm = 9999, mrank = T))
mant = rbind(mant, mantel(formula = as.dist(w.uni)~as.dist(neg.coh), nperm = 9999, mrank = T))
mant = rbind(mant, mantel(formula = as.dist(w.uni)~as.dist(pos.coh), nperm = 9999, mrank = T))
mant = rbind(mant, mantel(formula = as.dist(comm)~as.dist(neg.coh), nperm = 9999, mrank = T))
mant = rbind(mant, mantel(formula = as.dist(comm)~as.dist(pos.coh), nperm = 9999, mrank = T))
# By-well Mantel analysis
n = c("Athens", "Greene", "Licking")
r.names = NULL
for(i in 1:length(n)){
w = grep(n[i], row.names(bnti))
temp = mantel(formula = as.dist(bnti[w,w])~as.dist(neg.coh[w,w]), nperm = 9999, mrank = T)
temp = rbind(temp, mantel(formula = as.dist(bnti[w,w])~as.dist(pos.coh[w,w]), nperm = 9999, mrank = T))
temp = rbind(temp, mantel(formula = as.dist(uw.uni[w,w])~as.dist(neg.coh[w,w]), nperm = 9999, mrank = T))
temp = rbind(temp, mantel(formula = as.dist(uw.uni[w,w])~as.dist(pos.coh[w,w]), nperm = 9999, mrank = T))
temp = rbind(temp, mantel(formula = as.dist(w.uni[w,w])~as.dist(neg.coh[w,w]), nperm = 9999, mrank = T))
temp = rbind(temp, mantel(formula = as.dist(w.uni[w,w])~as.dist(pos.coh[w,w]), nperm = 9999, mrank = T))
temp = rbind(temp, mantel(formula = as.dist(comm[w,w])~as.dist(neg.coh[w,w]), nperm = 9999, mrank = T))
temp = rbind(temp, mantel(formula = as.dist(comm[w,w])~as.dist(pos.coh[w,w]), nperm = 9999, mrank = T))
mant = rbind(mant, temp)
r.names = c(r.names, paste(n[i],"bNTI-Neg"), paste(n[i],"bNTI-Pos"), paste(n[i],"UWU-Neg"), paste(n[i],"UWU-Pos"),
paste(n[i],"WU-Neg"), paste(n[i],"WU-Pos"), paste(n[i],"BC-Neg"), paste(n[i],"BC-Pos"))
}
row.names(mant) = c("Total bNTI-Neg", "Total bNTI-Pos", "Total UWU-Neg", "Total UWU-Pos",
"Total WU-Neg", "Total WU-Pos", "Total BC-Neg", "Total BC-Pos",
r.names)
### Multivariate comparisons between cohesion and turnover measurements ###
col = c(rgb(190,38,37, maxColorValue = 255), rgb(0,97,28, maxColorValue = 255), rgb(13,79,139, maxColorValue = 255)) # Generates vector of consistent colors for the NMDS
shape = c(15,17,16) # Generating consistent shapes for use in the NMDS plot
n = c("Athens", "Greene", "Licking")
# Plotting BC
y = prcomp(as.dist(comm)) # Generating principle component analysis for Bray-Curtis
ordiplot(y, type = "n", xlab = paste("PC1 (", (summary(y)$importance[2,1]*100), "% )", sep = ""),
ylab = paste("PC2 (", (summary(y)$importance[2,2]*100), "%)", sep = "")) # Blank plot
for(i in 1:length(n)){
w = grep(n[i], row.names(y$x))
points(x = y$x[w,1], y =y$x[w,2], col = col[i], pch = shape[i])
# text(x = y$x[w,1], y =y$x[w,2], labels = row.names(y$x)[w]) # Comment this out usually, just making sure my points are colored correctly
} # Plotting graph
x = envfit(y, data[,3:4], permutations = 999) # Generating the environmental fit for cohesion values
plot(x, p.max = 0.05) # Plotting only significant envfit results
multi.coh = cbind(x$vectors$r, x$vectors$pvals) # Storing all envfit results
# Plotting Unweighted UniFrac
y = prcomp(as.dist(uw.uni)) # Generating principle component analysis for Bray-Curtis
ordiplot(y, type = "n", xlab = paste("PC1 (", (summary(y)$importance[2,1]*100), "% )", sep = ""),
ylab = paste("PC2 (", (summary(y)$importance[2,2]*100), "%)", sep = "")) # Blank plot
for(i in 1:length(n)){
w = grep(n[i], row.names(y$x))
points(x = y$x[w,1], y =y$x[w,2], col = col[i], pch = shape[i])
# text(x = y$x[w,1], y =y$x[w,2], labels = row.names(y$x)[w])
} # Plotting graph
x = envfit(y, data[,3:4], permutations = 999) # Generating the environmental fit for cohesion values
plot(x, p.max = 0.05) # Plotting only significant envfit results
temp = cbind(x$vectors$r, x$vectors$pvals) # Storing all envfit results
multi.coh = rbind(multi.coh, temp) # Merging them with the previous results
# Plotting Weighted UniFrac
y = prcomp(as.dist(w.uni)) # Generating principle component analysis for Bray-Curtis
ordiplot(y, type = "n", xlab = paste("PC1 (", (summary(y)$importance[2,1]*100), "% )", sep = ""),
ylab = paste("PC2 (", (summary(y)$importance[2,2]*100), "%)", sep = "")) # Blank plot
for(i in 1:length(n)){
w = grep(n[i], row.names(y$x))
points(x = y$x[w,1], y =y$x[w,2], col = col[i], pch = shape[i])
# text(x = y$x[w,1], y =y$x[w,2], labels = row.names(y$x)[w])
} # Plotting graph
x = envfit(y, data[,3:4], permutations = 999) # Generating the environmental fit for cohesion values
plot(x, p.max = 0.05) # Plotting only significant envfit results
temp = cbind(x$vectors$r, x$vectors$pvals) # Storing all envfit results
multi.coh = rbind(multi.coh, temp) # Merging them with the previous results
# Plotting bNTI
y = prcomp(as.dist(bnti)) # Generating principle component analysis for Bray-Curtis
ordiplot(y, type = "n", xlab = paste("PC1 (", (summary(y)$importance[2,1]*100), "% )", sep = ""),
ylab = paste("PC2 (", (summary(y)$importance[2,2]*100), "%)", sep = "")) # Blank plot
for(i in 1:length(n)){
w = grep(n[i], row.names(y$x))
points(x = y$x[w,1], y =y$x[w,2], col = col[i], pch = shape[i])
# text(x = y$x[w,1], y =y$x[w,2], labels = row.names(y$x)[w])
} # Plotting graph
x = envfit(y, data[,3:4], permutations = 999) # Generating the environmental fit for cohesion values
plot(x, p.max = 0.05) # Plotting only significant envfit results
temp = cbind(x$vectors$r, x$vectors$pvals) # Storing all envfit results
multi.coh = rbind(multi.coh, temp) # Merging them with the previous results
row.names(multi.coh) = c("Bray-Curtis Negative", "Bray-Curtis Postive", "Unweighted Negative", "Unweighted Postive",
"Weighted Negative", "Weighted Postive", "bNTI Negative", "bNTI Postive")
colnames(multi.coh) = c("r-value", "pval")
### Selecting the previous timepoint comparisons to correlate, similar to Herren and McMahon ###
data = data[order(data$Well, data$Time),]
bnti = bnti[match(row.names(data), row.names(bnti)), match(row.names(data), row.names(bnti))]
comm = comm[match(row.names(data), row.names(comm)), match(row.names(data), row.names(comm))]
n = c("Athens", "Greene", "Licking")
bnti.vec = NULL
comm.vec = NULL
for(i in 1:length(n)){
w = grep(n[i], row.names(bnti))
temp.bnti = matrix(nrow = length(w)-1, ncol = 3)
temp.comm = matrix(nrow = length(w)-1, ncol = 3)
for(j in 1:length(w)-1){
temp.bnti[j,1] = as.numeric(bnti[j+1,j])
temp.bnti[j,2] = as.numeric(data[j+1,3])
temp.bnti[j,3] = as.numeric(data[j+1,4])
temp.comm[j,1] = as.numeric(comm[j+1,j])
temp.comm[j,2] = as.numeric(data[j+1,3])
temp.comm[j,3] = as.numeric(data[j+1,4])
}
bnti.vec = rbind(bnti.vec, temp.bnti)
comm.vec = rbind(comm.vec, temp.comm)
rm("temp.comm", "temp.bnti")
}
comm.mod = lm(comm.vec[,1]~comm.vec[,2])
bnti.mod = lm(bnti.vec[,1]~bnti.vec[,2])
### Looking at differences between cohesion distributions
sum.stat = NULL
r.names = NULL
temp = kruskal.test(data$Positive_Cohesion ~ data$Well)
sum.stat = rbind(sum.stat, c(temp$statistic, temp$p.value))
temp = kruskal.test(data$Negative_Cohesion ~ data$Well)
sum.stat = rbind(sum.stat, c(temp$statistic, temp$p.value))
for(i in 1:(length(n)-1)){
for(j in 1:length(n)){
if(i ==j){
print("Skipping self-comparison")
} else {
w = c(grep(n[i], row.names(data)), grep(n[j], row.names(data)))
temp = wilcox.test(data$Positive_Cohesion[w] ~ data$Well[w])
sum.stat = rbind(sum.stat, c(temp$statistic, temp$p.value))
temp = wilcox.test(data$Negative_Cohesion[w] ~ data$Well[w])
sum.stat = rbind(sum.stat, c(temp$statistic, temp$p.value))
r.names = c(r.names, paste(n[i], "-", n[j], "Pos"), paste(n[i], "-", n[j], "Neg"))
}
}
}
row.names(sum.stat) = c("Overall Pos", "Overall Neg", r.names)