-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathanalysis_summary.txt
336 lines (240 loc) · 11 KB
/
analysis_summary.txt
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
Summary:
We compared two methods of estimate true Bd load: first, we used strict cut-offs where Bd values must be over 5 individually, but also to have at least one (of three) tested values over 50.
The second method for estimating Bd load was to model Bd load using a Poisson process, and then ask whether the expected Bd load is statistically different from zero.
We find both models yield the approximate same amount; the Poisson process was more liberal in assigning "infected" status, but these discrepancies make sense logically since the values for those are generally small but consistent between the three measurements.
We adjusted the time variable so that time is continuous from time point 1-16.
Osse was not sampled the first week, so we started its time at time point 2.
Bubo was also strange because it was the only species sampled at time point 10; all other species skipped this time point. So we adjusted for this.
We found a small error where Raca10 had 2 time point 3's and Buma11 has 2 time point 2's; we believe this is a data entry mistake and have changed one fo the Buma11's to Buma9 and one of the Raca 10's to Raca 9. This change is supported by the other metadata columns.
There were also 14 individuals that were contaminated upon arrival. We removed these individuals from analyses.
The final OTU table was filtered by:
(1) changing all read values less than 5 to 0
(2) removing OTUs that had less than 100 total reads in the entire dataset
(3) removing all samples with less than 5000 reads
PRELIMINARY FINDINGS:
- Among control individuals, composition of microbiome is different between species, time, and interaction (ADONIS; all p < 0.001; sp R2 = 0.551, timeR2 = 0.051, sp:timeR2 = 0.053)
-shows community composition is changing, and that different species are changing differently over time.
Call:
adonis(formula = dist(dm.filt.con) ~ species * time, data = mf_con_without_init_infect)
Permutation: free
Number of permutations: 999
Terms added sequentially (first to last)
Df SumsOfSqs MeanSqs F.Model R2 Pr(>F)
species 4 424.12 106.030 78.785 0.55117 0.001 ***
time 1 39.03 39.028 28.999 0.05072 0.001 ***
species:time 4 41.22 10.305 7.657 0.05357 0.001 ***
Residuals 197 265.12 1.346 0.34454
Total 206 769.49 1.00000
---
- Among treatment individuals, time and species, again, effects composition but PABD as an added term only affects along (0.015, R2 = 0.007) but not in interactions
Call:
adonis(formula = dist(dm.filt.treat) ~ species * time * PABD, data = mf_treat_without_init_infect_post)
Permutation: free
Number of permutations: 999
Terms added sequentially (first to last)
Df SumsOfSqs MeanSqs F.Model R2 Pr(>F)
species 4 344.67 86.168 43.631 0.44148 0.001 ***
time 1 24.55 24.550 12.431 0.03145 0.001 ***
PABD 1 5.91 5.913 2.994 0.00757 0.015 *
species:time 4 28.21 7.052 3.571 0.03613 0.001 ***
species:PABD 3 6.42 2.140 1.084 0.00822 0.327
time:PABD 1 9.80 9.795 4.960 0.01255 0.001 ***
species:time:PABD 2 5.67 2.835 1.435 0.00726 0.140
Residuals 180 355.48 1.975 0.45533
Total 196 780.71 1.00000
---
- shannon and logRich are visually well-described by a normal distribution
- Want to see if richness is changing over time:
Type I to test for interaction in control group
Analysis of Variance Table
Response: shannon
Df Sum Sq Mean Sq F value Pr(>F)
species 4 39.401 9.8503 24.3888 <2e-16 ***
time 1 1.060 1.0605 2.6256 0.1067
species:time 4 2.742 0.6855 1.6973 0.1521
Residuals 197 79.565 0.4039
---
Type II to test for main effects in control group
Anova Table (Type II tests)
Response: shannon
Sum Sq Df F value Pr(>F)
species 39.598 4 24.1750 2.417e-16 ***
time 1.060 1 2.5897 0.1091
Residuals 82.307 201
---
Type I to test for interaction in treatment group
Analysis of Variance Table
Response: shannon
Df Sum Sq Mean Sq F value Pr(>F)
species 4 29.920 7.4801 12.3760 2.884e-09 ***
time 1 1.654 1.6539 2.7364 0.09923 .
species:time 4 23.685 5.9213 9.7969 2.020e-07 ***
Residuals 274 165.607 0.6044
---
Type II to test for main effects in treatment group
Anova Table (Type III tests)
Response: shannon
Sum Sq Df F value Pr(>F)
(Intercept) 637.40 1 1054.5859 < 2.2e-16 ***
species 14.19 4 5.8692 0.0001517 ***
time 0.98 1 1.6290 0.2029166
species:time 23.69 4 9.7969 2.02e-07 ***
Residuals 165.61 274
---
RICHNESS
Type I to test for interaction in control group
Analysis of Variance Table
Response: logRich
Df Sum Sq Mean Sq F value Pr(>F)
species 4 9.8448 2.46120 17.3532 3.263e-12 ***
time 1 0.1974 0.19738 1.3916 0.2396
species:time 4 0.8054 0.20135 1.4197 0.2288
Residuals 197 27.9404 0.14183
---
Type II to test for main effects in control group
Anova Table (Type II tests)
Response: logRich
Sum Sq Df F value Pr(>F)
species 9.9684 4 17.4255 2.724e-12 ***
time 0.1974 1 1.3801 0.2415
Residuals 28.7458 201
---
Type I to test for interaction in treatment group
Analysis of Variance Table
Response: logRich
Df Sum Sq Mean Sq F value Pr(>F)
species 4 11.768 2.94196 18.1367 3.14e-13 ***
time 1 0.000 0.00048 0.0029 0.9567623
species:time 4 3.364 0.84109 5.1852 0.0004847 ***
Residuals 274 44.446 0.16221
---
Type III to test for main effects in treatment group
Anova Table (Type III tests)
Response: logRich
Sum Sq Df F value Pr(>F)
(Intercept) 916.28 1 5648.7368 < 2.2e-16 ***
species 2.44 4 3.7548 0.0054249 **
time 0.02 1 0.1238 0.7252123
species:time 3.36 4 5.1852 0.0004847 ***
Residuals 44.45 274
---
Summary [ Note; richness here is same as shannon ]:
- Species differ in richness, but richness does not change over time in control group.
- Richness changes over time doesn't change uniformly when infected, BUT different species change in different ways once infected.
BETA DIVERSITY (instability)
Type I to test for interaction in control group
Analysis of Variance Table
Response: distance_bray_curtis
Df Sum Sq Mean Sq F value Pr(>F)
species 4 0.50309 0.125772 7.6486 1.161e-05 ***
time 1 0.00416 0.004155 0.2527 0.6159
species:time 4 0.05099 0.012746 0.7752 0.5429
Residuals 159 2.61456 0.016444
---
Type II to test for main effects in control group
Anova Table (Type II tests)
Response: distance_bray_curtis
Sum Sq Df F value Pr(>F)
species 0.47511 4 7.2633 2.086e-05 ***
time 0.00416 1 0.2541 0.6149
Residuals 2.66554 163
---
Type I to test for interaction in treatment group
Analysis of Variance Table
Response: distance_bray_curtis
Df Sum Sq Mean Sq F value Pr(>F)
species 4 0.6596 0.164896 6.9084 2.99e-05 ***
time 1 0.0775 0.077468 3.2455 0.07301 .
species:time 4 0.0392 0.009805 0.4108 0.80078
Residuals 216 5.1557 0.023869
---
Type II to test for main effects in treatment group
Anova Table (Type II tests)
Response: distance_bray_curtis
Sum Sq Df F value Pr(>F)
species 0.6256 4 6.6233 4.742e-05 ***
time 0.0775 1 3.2807 0.07146 .
Residuals 5.1949 220
---
Summary:
- Species differ in their microbiome instability, but this signal is really small relative to noise from sampling effort
- No effect of time on microbiome instability
PERCENT INHIBITORY
Type I to compare interactions in control
Analysis of Deviance Table
Model: binomial, link: logit
Response: percInhib
Terms added sequentially (first to last)
Df Deviance Resid. Df Resid. Dev Pr(>Chi)
NULL 206 505798
species 4 73171 202 432627 < 2.2e-16 ***
time 1 4778 201 427849 < 2.2e-16 ***
species:time 4 80748 197 347101 < 2.2e-16 ***
---
Type III to compare main effects in control
Analysis of Deviance Table (Type III tests)
Response: percInhib
LR Chisq Df Pr(>Chisq)
species 81206 4 <2e-16 ***
time 0 1 0.7259
species:time 80748 4 <2e-16 ***
---
Type I to compare interactions in treatment
Analysis of Deviance Table
Model: binomial, link: logit
Response: percInhib
Terms added sequentially (first to last)
Df Deviance Resid. Df Resid. Dev Pr(>Chi)
NULL 283 1273299
species 4 163434 279 1109865 < 2.2e-16 ***
time 1 107035 278 1002830 < 2.2e-16 ***
species:time 4 50701 274 952129 < 2.2e-16 ***
---
Type III to compare main effects in treatment
Analysis of Deviance Table (Type III tests)
Response: percInhib
LR Chisq Df Pr(>Chisq)
species 42348 4 < 2.2e-16 ***
time 57575 1 < 2.2e-16 ***
species:time 50701 4 < 2.2e-16 ***
---
InhibRich
- Fits a poisson distribution well
Type I to test for interactions in control group
Analysis of Deviance Table
Model: poisson, link: log
Response: inhibRich
Terms added sequentially (first to last)
Df Deviance Resid. Df Resid. Dev Pr(>Chi)
NULL 206 232.35
species 4 29.2924 202 203.06 6.818e-06 ***
time 1 0.6686 201 202.39 0.4135
species:time 4 31.3583 197 171.03 2.587e-06 ***
---
Type III test to test main effects in control group
Analysis of Deviance Table (Type III tests)
Response: inhibRich
LR Chisq Df Pr(>Chisq)
species 42.494 4 1.318e-08 ***
time 3.141 1 0.07633 .
species:time 31.358 4 2.587e-06 ***
---
Type I to test for interactions in treatment group
Analysis of Deviance Table
Model: poisson, link: log
Response: inhibRich
Terms added sequentially (first to last)
Df Deviance Resid. Df Resid. Dev Pr(>Chi)
NULL 283 361.86
species 4 130.642 279 231.22 < 2.2e-16 ***
time 1 0.033 278 231.19 0.856329
species:time 4 18.271 274 212.92 0.001092 **
---
Type III test to test for main effects
Analysis of Deviance Table (Type III tests)
Response: inhibRich
LR Chisq Df Pr(>Chisq)
species 23.8941 4 8.387e-05 ***
time 0.7994 1 0.371281
species:time 18.2715 4 0.001092 **
---