-
Notifications
You must be signed in to change notification settings - Fork 16
/
temporal-dpc.qmd
261 lines (210 loc) · 12.6 KB
/
temporal-dpc.qmd
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
---
output: html_document
editor_options:
chunk_output_type: console
---
# Disease progress curves
## How epidemics occur
Before knowing how epidemics develop in time, it is important to understand how an epidemic occur. An epidemic begins when the **primary inoculum** (a variable number of propagules able to infect the plant) that is *surviving* somewhere establishes an intimate contact with individuals of the host population - this process is called *infection*. These inocula are usually surviving externally to the plant host and need to *disperse* (move), passively or by means of a vector, to reach the plant. It can also be that a growing host encounter a localized (static) source of inoculum.
Once the infection is established, the pathogen *colonizes* the plant tissues and disease symptoms are noticed. When this happens, the **incubation period** can be measured in time units. A successful colonization will lead to *reproduction* of the pathogen inside and/or external to the crop, and so the **latent period** is completed, and can also be measure in time units. Finally, the **infectious period** takes place and continues until the pathogen is not capable of producing the **secondary inoculum** on the infected site.
```{mermaid}
%%| label: fig-diagram
%%| fig-cap: "Five main processes of the disease cycle"
flowchart
A[Infection] --> B[Colonization]
B --> C[Reproduction]
C -. New inoculum .-> D[Dispersal]
E[Survival] -- Primary inoculum --> D
D --> A
D -.-> A
C --> E
```
Epidemiologists are generally interested in determining the length of the incubation, latent, and infectious periods as influenced by factors related to the host, pathogen, or environment. This is relevant because the longer it takes for the completion of the incubation and latent periods, the lower the potential number of repeated cycles. In summary, a single "infection cycle" represents all events that occur from infection to dispersal, and this occurs only once for many diseases, while for others there may be multiple cycles, which are defined as an "infection chain."
```{r}
#| warning: false
#| message: false
#| code-fold: true
library(tidyverse)
library(r4pde)
periods <- tibble::tribble(
~period, ~length, ~color, ~order,
"Incubation", 10, 0, 1,
"Latent" , 15, 0, 2,
"Infectious", 25, 15, 3
)
p <- periods |>
ggplot(aes(reorder(period, order), length, fill = period))+
geom_col()+
geom_col(aes(period, color), color = "white", fill = "white")+
coord_flip()+
theme_void()+
theme(legend.position = "none")+
annotate(geom = "text", x = 0.5, y = 15, label = "----- Time --->")+
annotate(geom = "text", x = 1, y = 5, label = "Incubation", color = "white")+
annotate(geom = "text", x = 2, y = 8, label = "Latent", color = "white")+
annotate(geom = "text", x = 3, y = 20, label = "Infectious", color = "white")+
annotate(geom = "text", x = 1, y = 10.5, label = "Visible symptoms", angle = 90, size = 1.7)+
annotate(geom = "text", x = 2, y = 15.5, label = "Reproduction starts", angle = 90, size =1.7)+
annotate(geom = "text", x = 3, y = 25.5, label = "Reproduction ends", angle = 90, size =1.7)+
scale_fill_manual(values = c("darkgreen", "brown", "darkorange"))+
geom_segment(mapping=aes(x=0.6, y=0, xend=0.6, yend=10), arrow=arrow(ends='both'), size=1, color = "black")+
geom_segment(mapping=aes(x=1.6, y=0, xend=1.6, yend=15), arrow=arrow(ends='both'), size=1, color = "black") +
geom_segment(mapping=aes(x=2.6, y=15, xend=2.6, yend=25), arrow=arrow(ends='both'), size=1, color = "black")
library(png)
library(cowplot)
incubation <- readPNG("imgs/incubation3.png", native = TRUE)
latent <- readPNG("imgs/latent3.png", native = TRUE)
p2 <- p + draw_image(incubation , x = 0.5, y = 13, scale = 5)+
draw_image(latent , x = 1.5, y = 20, scale = 5)
ggsave("imgs/periods.png", width =6, height =2, bg = "white")
```
{#fig-periods fig-align="center"}
## Disease curves
A key understanding of the epidemics relates to the knowledge of rates and patterns. Epidemics can be viewed as dynamic systems that change their state as time goes. The first and simplest way to characterize such changes in time is to produce a graphical plot called disease progress curve (DPC). This curve can be obtained as long as the intensity of the disease (*y*) in the host population is assessed sequentially in time (*t*).
A DPC summarizes the interaction of the three main components of the disease triangle occurring during the epidemic. The curves can vary greatly in shape according to variations in each of the components, in particular due to management practices that alter the course of the epidemics and for which the goal is to stop disease increase. We can create a data frame in R for a single DPC and make a plot using ggplot. By convention we use `t` for time and `y` for disease intensity, expressed in percentage (0 to 100%).
Firstly, let's load the essential R packages and set up the environment.
```{r}
#| warning: false
#| message: false
library(tidyverse) # essential packages
theme_set(theme_r4pde()) # set global theme
```
There are several ways to create a data frame in R. I like to use the `tribble` function as below. The entered data will be assigned to a dataframe called `dpc`.
```{r}
dpc <-
tribble(
~t, ~y,
0, 0.1,
7, 1,
14, 9,
21, 25,
28, 80,
35, 98,
42, 99,
49, 99.9
)
```
Now the plot
```{r}
#| message: false
dpc1 <- dpc |>
ggplot(aes(t, y)) +
theme_r4pde()+
geom_line(size = 1)+
geom_point(size = 3, shape = 16)+
labs(x = "Assessment time (days)",
y = "Disease intensity (%)")
ggsave("imgs/dpc1.png", dpc1)
```
{#fig-dpc1}
## Epidemic classification
Vanderplank analysed the shapes of great number of epidemic curves and classified the epidemics into two basic types: monocyclic or polycyclic [@plantdi1963]. In **monocyclic** epidemics, inoculum capable of infecting the crop is not produced during the epidemics. These epidemics are initiated and maintained only by the primary inoculum. There is no secondary infection and hence no further spread of newly produced inoculum among the host individuals. Tipically, the progress curves for monocyclic epidemics have a saturation type shape. An example of a monocyclic epidemic is the disease known as wheat smut caused by *Tilletia caries*. In this disease, although two types of spores are produced during a single cycle (teliospores and basidiospores), no additional infection cycles occur within the same growing season.
Conversely, when the secondary inoculum produced during the epidemics is capable of infecting the host during the same crop cycle, a **polycyclic** epidemic is established. The number of repeated cycles just depends on how long it takes to complete a single infection cycle. These epidemics most commonly present a sigmoid shape @fig-cycles. An example of a polycyclic epidemic is soybean rust as it undergoes repeated infection cycles with uredospores on soybean.
```{r}
#| warning: false
#| message: false
#| code-fold: true
library(tidyverse)
theme_set(theme_bw(base_size = 16))
library(epifitter)
polyc <- sim_logistic(N = 50, dt = 5,
y0 = 0.01, r = 0.2,
K = 0.8, n = 1,
alpha =0)
p <- polyc |>
ggplot(aes(time, y))+
geom_point(aes(time, y), size =19, shape =1)+
geom_line()+
ylim(0,1)+
theme_r4pde()+
labs(x = "Time", y = "Disease intensity")
monoc <- sim_monomolecular(N = 50, dt = 5,
y0 = 0.01, r = 0.1,
K = 0.8, n = 1,
alpha =0)
library(ggforce)
m <- monoc |>
ggplot(aes(time, y))+
geom_point(aes(x = 25, y = 0.5), size =90, shape = 1)+
geom_line()+
theme_r4pde()+
ylim(0,1)+
labs(x = "Time", y = "Disease intensity")
library(patchwork)
cycles <- m | p
ggsave("imgs/cycles.png", bg = "white", width = 8, height =4)
```
{#fig-cycles fig-align="center"}
## Curve descriptors and AUDPC
The depiction and analysis of disease progress curves can provide useful information for gaining understanding of the underlying epidemic process. The curves are extensively used to evaluate how disease control measures affect epidemics. When characterizing DPCs, a researcher may be interested in describing and comparing epidemics that result from different treatments, or simply in their variations as affected by changes in environment, host or pathogen.
The precision and complexity of the analysis of progress curve data depends on the objective of the study. In general, the goal is to synthesize similarities and differences among epidemics based on common descriptors of the disease progress curves. For example, the simple appraisal of the disease intensity at any time during the course of the epidemic should be sufficient for certain situations. Furthermore, a few quantitative and qualitative descriptors can be extracted including:
- Epidemic duration
- Maximum disease
- Curve shape
- Area under the area under the disease progress curve (AUDPC).
Let's visualize the AUDPC in the same plot that we produced above.
```{r}
#| message: false
dpc2 <- dpc |>
ggplot(aes(t, y)) +
labs(x = "Assessment time (days)",
y = "Disease intensity (%)")+
geom_area(fill = "#339966")+
geom_line(linewidth = 1)+
theme_r4pde()+
geom_point(size = 3, shape = 16)+
scale_x_continuous(breaks = c(0, 7, 14, 21, 28, 35, 42))
ggsave("imgs/dpc2.png")
```
{#fig-audpc2 fig-align="center"}
The AUDPC summarizes the "total measure of disease stress" and is largely used to compare epidemics [@jeger2001]. The most common approach to calculate AUDPC is the trapezoidal method, which splits the disease progress curves into a series of rectangles, calculating the area of each of them and then summing the areas. Let's extend the plot code to show those rectangles using the `annotate` function.
```{r}
#| message: false
#| code-fold: true
dpc3 <- dpc |>
ggplot(aes(t, y)) +
theme_r4pde()+
labs(x = "Assessment time (days)",
y = "Disease intensity (%)")+
annotate("rect", xmin = dpc$t[1], xmax = dpc$t[2],
ymin = 0, ymax = (dpc$y[1]+ dpc$y[2])/2,
color = "white", fill = "#339966")+
annotate("rect", xmin = dpc$t[2], xmax = dpc$t[3],
ymin = 0, ymax = (dpc$y[2]+ dpc$y[3])/2,
color = "white", fill = "#339966")+
annotate("rect", xmin = dpc$t[3], xmax = dpc$t[4],
ymin = 0, ymax = (dpc$y[3]+ dpc$y[4])/2,
color = "white", fill = "#339966")+
annotate("rect", xmin = dpc$t[4], xmax = dpc$t[5],
ymin = 0, ymax = (dpc$y[4]+ dpc$y[5])/2,
color = "white", fill = "#339966")+
annotate("rect", xmin = dpc$t[5], xmax = dpc$t[6],
ymin = 0, ymax = (dpc$y[5]+ dpc$y[6])/2,
color = "white",fill = "#339966")+
annotate("rect", xmin = dpc$t[6], xmax = dpc$t[7],
ymin = 0, ymax = (dpc$y[6]+ dpc$y[7])/2,
color = "white", fill = "#339966")+
annotate("rect", xmin = dpc$t[7], xmax = dpc$t[8],
ymin = 0, ymax = (dpc$y[7]+ dpc$y[8])/2,
color = "white", fill = "#339966")+
geom_line(linewidth = 1)+
geom_point(size = 3, shape = 16)+
annotate(geom = "text", x = 36.5, y = 50,
label = "AUDPC = 2534" , size = 4)+
scale_x_continuous(breaks = c(0, 7, 14, 21, 28, 35, 42, 49))
ggsave("imgs/dpc3.png")
```
{#fig-dpc3 fig-align="center"}
In R, we can obtain the AUDPC for the DPC we created earlier using the `AUDPC` function offered by the *epifitter* package. Because we are using the percent data, we need to set the argument `y_proportion = FALSE`. The function returns the absolute AUDPC. If one is interested in relative AUDPC, the argument `type` should be set to `"relative"`. There is also the alternative to AUDPC, the area under the disease progress stairs (AUDPS) [@simko2012].
```{r}
library(epifitter)
AUDPC(dpc$t, dpc$y,
y_proportion = FALSE)
# The relative AUDPC
AUDPC(dpc$t, dpc$y,
y_proportion = FALSE,
type = "relative")
# To calculate AUDPS, the alternative to AUDPC
AUDPS(dpc$t, dpc$y,
y_proportion = FALSE)
```