PreprintPDF Available
Preprints and early-stage research may not have been peer reviewed yet.

Abstract and Figures

In the last years, the emergence of zoonotic diseases and the frequency of disease outbreaks have increased substantially, fuelled by habitat encroachment and asynchrony of biological cycles due to global change. The virulence of these diseases is a key aspect for their success. In order to understand the complex processes of pathogen virulence evolution in the global change context, we adapted an established individual-based model of host-pathogen dynamics. Our model simulates a population of social hosts affected by an evolving pathogen in a dynamic landscape. Pathogen virulence evolution is explored by the inclusion of multiple strains in the model that differ in their transmission capability and lethality. Simultaneously, the host’s resource landscape is subjected to spatial and temporal dynamics, emulating effects of global change. We found an increase in pathogenic virulence and a shift in strain dominance with increasing landscape homogenisation. Our model further shows a trend to lower virulence pathogens being dominant in fragmented landscapes, although pulses of highly virulent strains are expected under resource asynchrony. While all landscape scenarios favour coexistence of low and high virulent strains, when host density increases, the high virulence strains capitalize on the high possibility for transmission and are likely to become dominant. Author Summary Disease outbreaks primarily caused by contact with animals are increasing in recent years, related to habitat destruction and altered biological cycles due to climate change. Pathogens associated with such outbreaks will be more successful the more effectively they can spread in a population. Thus, understanding the conditions over which those pathogens evolve will help us to limit the impact of disease outbreaks in the future. To this end, we used an individual based model that allowed us to study different scenarios. Our model had three main components: a host-pathogen system, a dynamic resource landscape with different degrees of fragmentation and temporal resource mismatches. We used dynamic landscapes with varying resource amounts over the years and consisting of multiple large or smaller habitat clusters. Our simulations showed that homogenous landscapes resulted in higher virulent pathogens and fragmented landscapes in lesser virulent pathogens. However, across all scenarios, high and low virulent pathogen strains were able to coexist.
Content may be subject to copyright.
Resource asynchrony and landscape homogenization as drivers of virulence
1
evolution
2
Tobias Kürschner1,2, Cédric Scherer2, Viktoriia Radchuk2, Niels Blaum1, Stephanie Kramer-
3
Schadt2,3
4
1 University of Potsdam, Plant Ecology and Nature Conservation, Potsdam, Germany
5
2 Leibniz Institute for Zoo and Wildlife Research, Department of Ecological Dynamics, Berlin, Germany
6
3 Technische Universität Berlin, Institute of Ecology, Berlin, Germany
7
Keywords: virulence, evolution, host-pathogen dynamics, dynamic landscapes, global change
8
9
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
Abstract
10
In the last years, the emergence of zoonotic diseases and the frequency of disease outbreaks
11
have increased substantially, fuelled by habitat encroachment and asynchrony of biological
12
cycles due to global change. The virulence of these diseases is a key aspect for their success.
13
In order to understand the complex processes of pathogen virulence evolution in the global
14
change context, we adapted an established individual‐based model of host-pathogen
15
dynamics. Our model simulates a population of social hosts affected by an evolving pathogen
16
in a dynamic landscape. Pathogen virulence evolution is explored by the inclusion of multiple
17
strains in the model that differ in their transmission capability and lethality. Simultaneously,
18
the host’s resource landscape is subjected to spatial and temporal dynamics, emulating effects
19
of global change.
20
We found an increase in pathogenic virulence and a shift in strain dominance with increasing
21
landscape homogenisation. Our model further shows a trend to lower virulence pathogens
22
being dominant in fragmented landscapes, although pulses of highly virulent strains are
23
expected under resource asynchrony. While all landscape scenarios favour coexistence of low
24
and high virulent strains, when host density increases, the high virulence strains capitalize on
25
the high possibility for transmission and are likely to become dominant.
26
27
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
Author Summary
28
Disease outbreaks primarily caused by contact with animals are increasing in recent years,
29
related to habitat destruction and altered biological cycles due to climate change. Pathogens
30
associated with such outbreaks will be more successful the more effectively they can spread
31
in a population. Thus, understanding the conditions over which those pathogens evolve will
32
help us to limit the impact of disease outbreaks in the future. To this end, we used an
33
individual based model that allowed us to study different scenarios. Our model had three
34
main components: a host-pathogen system, a dynamic resource landscape with different
35
degrees of fragmentation and temporal resource mismatches. We used dynamic landscapes
36
with varying resource amounts over the years and consisting of multiple large or smaller
37
habitat clusters. Our simulations showed that homogenous landscapes resulted in higher
38
virulent pathogens and fragmented landscapes in lesser virulent pathogens. However, across
39
all scenarios, high and low virulent pathogen strains were able to coexist.
40
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
Introduction
41
42
A key aspect of the invasive success of infectious pathogens such as Ebola, SARS-CoV-2 or
43
Avian Influenza in a host population is the mastering of the delicate interplay of transmission
44
and host exploitation, also termed virulence. To persist, a pathogen must find the balance
45
between quick replication and growth in the host often resulting in severe infections killing
46
its host while still being able to spread across timescales (Visher et al. 2021). This intricate
47
balance can only be kept up by an arms race between hosts immune reactions and strategies
48
of the pathogen to evade and counteract host resistance, termed adaptive evolution of
49
virulence (Cressler et al. 2016). This leads to the emergence of ever-new pathogenic strains
50
from the wild strain with modulated pathogenic traits, and if the new strain manages to
51
establish, this might have unforeseeable effects on host population and disease dynamics.
52
53
Both transmission and virulence are integrally tied to density and spatiotemporal distribution
54
of host individuals (Alizon et al. 2009, Cressler et al. 2016), which in return are subject to
55
habitat configuration and spatiotemporal variation in resource availability.
56
Global change might exacerbate disease dynamics in the near future, facilitated by land-use
57
change, habitat encroachment, or climate warming (Patz et al. 2004, Wilcox and Gubler
58
2005). These disturbances will severely influence disease outbreaks by changes in the life
59
history, density and availability of hosts as well as feedbacks on the landscape level due to
60
asynchrony in timescales (Kürschner et al. 2021). In this context, it is particularly important
61
to not only understand factors that govern the spread and the persistence of pathogens in
62
changing landscapes to put counteractive measures in place (Griette et al. 2015), but to also
63
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
understand how these factors reciprocally influence the adaptive potential of pathogenic
64
traits.
65
66
Virulence evolution is often highly accelerated during the emergence or invasion stage of an
67
epidemic (Griette et al. 2015, Geoghegan and Holmes 2018). The emergence stage is
68
characterized by a high number of susceptible and later infected host individuals
69
associated with a high number of mutations due to the steep increase of infected individuals
70
(Galvani 2003). Since the distribution of host individuals in a landscape determines the
71
number of available susceptible individuals, local and regional host densities are important
72
factors in the evolution of virulence (Boots 2004). With global change further altering the
73
resource distribution in space and time, subsequent changes in the spatiotemporal density and
74
distribution of host individuals (Galvani 2003, Boots 2004, Geoghegan and Holmes 2018)
75
could influence the evolution of virulence. Density changes could for example be induced via
76
mismatches between the host’s life history such as reproduction and host resource availability
77
at that time.
78
79
Theory predicts an evolution towards low virulence through altered habitat configuration or
80
host density distribution (Boots and Mealor 2007, Cressler et al. 2016). Virulence has been
81
shown to be adaptive if there is a correlation with other pathogenic traits such as
82
contagiousness, which is known as virulence-transmission trade-off hypothesis (Day 2003).
83
The transmission-virulence trade-off hypothesis states that an increase in strain transmission
84
causes shorter infections through higher lethality (Anderson and May 1982, Alizon and
85
Michalakis 2015). In other words, pathogen virulence is subject to a variety of evolutionary
86
trade-offs (Kamo et al. 2007, Messinger and Ostling 2009, Cressler et al. 2016).
87
88
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
The theoretical models of virulence evolution, particularly the classical adaptive dynamics
89
framework, rely on the assumption that mutation of pathogens happens very slowly and that
90
mutations towards new strains can only occur after the dominant strain has reached
91
equilibrium (Dieckmann et al. 2005). However, such simplified assumptions are rarely
92
applicable to pathogens in nature, which often undergo transient dynamics, for example due
93
to temporal and spatial changes in the landscape structure. Due to temporal variation in the
94
landscape, the formation of spatial (figure 1 a) and or temporal (figure 1 b) host niches can
95
cascade through the density distribution of potential hosts onto host-pathogen interactions
96
(figure 1 c, d). The formation of niches with varying beneficial or detrimental properties for
97
host and pathogen could facilitate the appearance of different pathogenic strains at specific
98
times or locations. The result can be a complex system of different competing and coexisting
99
pathogen strains (figure 1 e) with their own spatial and temporal dynamics. The constant
100
emergence, re-emergence, and extinction of pathogenic strains will result in an overlap and
101
possible coexistence between different strains, all competing for the same resource (Choua
102
and Bonachela 2019).
103
While theoretical studies focus on long-term predictions of pathogenic strains with
104
evolutionarily stable virulence at equilibrium (Lenski and May 1994, Day and Gandon 2007),
105
there is a lack of knowledge linking complex dynamics arising from global change to the
106
evolution of virulence through space and time during an epidemic (Lebarbenchon et al.
107
2008). Also, links between resources and host density are rarely incorporated into
108
evolutionary models, which typically assume that host density remains at equilibrium [3,28 in
109
Hite&Cressler2018]. A prominent example tackling the evolution of virulence in changing
110
host densities due to changes in resources is the work of Hite & Cressler (2018). They
111
revealed complex effects of host population dynamics on parasite evolution, including
112
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
regions of evolutionary bistability, where parasites ‘rode the cycles’ of their hosts and phases
113
with high host exploitation superseded phases of low virulence (Hite & Cressler 2018).
114
115
Here, we go one step beyond the important link between host ecology and parasite evolution
116
by asking what effect heterogeneously distributed and dynamic resources will pose on the
117
evolution of virulence, particularly how temporal mismatches between optimal resource
118
availability and biological events, such as reproduction, affect host-pathogen coexistence and
119
pathogen spread through adaptive virulence dynamics. To this end, we modified an existing
120
spatially-explicit individual-based host-pathogen model of a group-living social herbivore
121
(Kramer-Schadt et al. 2009, Lange et al. 2012a, b, Scherer et al. 2020, Kürschner et al. 2021)
122
and added evolution in pathogen traits leading to multi-strain outbreak scenarios. In
123
accordance with theory, we have already shown for a static host exploitation rate that
124
pathogen extinction is higher in landscapes with randomly distributed and fluctuating
125
resources, but that the formation of disease hotspots form an epidemic rescue for the
126
pathogen when hosts are mobile (Kürschner et al. 2021).
127
128
We here hypothesized that dynamic landscapes induce evolution in pathogenic virulence to
129
facilitate host-pathogen coexistence (H1). In more detail, we expect pathogenic virulence to
130
evolve into a system of different viral strains that will coexist and persist within the host
131
population in parallel (prediction 1). We also predict that the frequency of ‘host cycle riding’
132
pathogenic strain emergence will be larger under environmental uncertainty, hence global
133
change effects might lead to higher pathogenic strain emergence (prediction 2), i.e. with a
134
higher chance for spill-over events.
135
136
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
We further hypothesize that due to the destabilisation of the host population under
137
asynchronous dynamics, virulence will evolve to lower levels than under homogeneous and
138
stable resource availability (H2). We expect increasing landscape homogenization and related
139
contact homogenization to facilitate evolution towards higher pathogenic virulence by
140
increasing the availability of hosts for highly virulent strains (prediction 3), with few
141
dominant strains governing the dynamics for a long time (prediction 4).
142
143
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
144
Figure 1: Conceptual figure: Landscape homogenization (A) and synchrony/asynchrony (tlag)
145
of host life-history and host-resource availability (B) influence host-pathogen dynamics (C)
146
and subsequently the evolution of pathogenic traits (D) that will affect strain occurrence over
147
time where gaps in the background line are times when the strain did not occur in the
148
landscape (E).
149
150
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
Results
151
Host-pathogen coexistence
152
Overall, host pathogen coexistence Pcoex was very high in almost all tested scenarios. Due to a
153
collapse of the host population in asynchronous scenarios in homogeneous landscapes, host-
154
pathogen coexistence was not achievable in the current model-framework and this single
155
scenario therefore excluded. We did not find notable differences in the other scenarios
156
(Appendix Fig. B2).
157
Categorized infection trends
158
Our model showed that in synchronous scenarios, highly virulent strains were the least
159
abundant ones among the three strain categories during the early stages of the epidemic.
160
However, these strains became dominant in the later stages of the epidemic in homogeneous
161
and large clustered landscapes (figure 2, left). With increasing landscape homogenization,
162
medium virulence strains in the later stages of the epidemic were usually dominating along
163
with high virulence strains. Across all landscapes, low virulent strains only occurred in high
164
prevalence in the early stages of the epidemic but reached higher prevalence in less
165
heterogeneous landscapes.
166
167
In scenarios with asynchrony, low virulent strains occurred over a longer time period and
168
were more prevalent in the host population, while medium and highly virulent strains
169
occurred later at high prevalence (figure 2, right). Furthermore, prevalence of all strain
170
categories was lower throughout the simulations when directly compared to the
171
‘synchronous’ scenarios. A clear shift towards a dominance of highly virulent strains only
172
occurred in the less heterogeneous, large clustered landscapes.
173
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
174
175
Figure 2: Temporal trends of the number of hosts infected with the strains of three virulence
176
categories, low (blue) medium (black) and high (red) virulence over time. The left half shows
177
the trends for synchronous host reproduction (tlag = 0) and the right half for asynchronous
178
host reproduction (tlag = 100) scenarios.
179
Strain specific occurrence and dominance over time
180
Strain occurrence and dominance of high virulent strains decreased with increasing landscape
181
heterogenization in dynamic landscapes in combination with synchronous scenarios (figure 3
182
top). In the highly heterogeneous random landscape, low virulent strains persisted longer
183
while higher virulent strains occurred much later in time (around year 40), compared to the
184
same landscape in synchronous scenarios. After its appearance through mutations around
185
year 45, the medium virulent strain 7 became dominant for the course of the epidemic. As
186
landscape homogeneity increased (from random landscapes to small clustered landscapes),
187
lower virulent strains occurred on much shorter time spans. The high virulent strains
188
appeared much earlier around year 10 of the epidemic and remained present in the landscape
189
for the duration of the simulation. A further increase in homogenization towards medium-
190
sized patches showed overall similar patterns as the small clustered landscape with the
191
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
exception of the strain 7 peak prevalence, which shifted towards the end of the epidemic. In
192
the highly homogenous large clustered landscapes, the lower virulent strains persisted for a
193
longer period, while medium virulent strains were represented during the full period of the
194
epidemic. Contrary to the more heterogeneous landscapes, strain 7 did not become the
195
dominant strain despite high prevalence in the host population. As indicated by the larger
196
proportion of hosts infected with higher virulent strains, overall, in synchronous scenarios
197
(figure 3), virulence of occurring strains increased over time and with increasing landscape
198
homogenization. In asynchronous scenarios we observed a similar increase in strain
199
occurrence with landscape homogenization, even though the temporal rate of increase was
200
smaller compared to synchronous scenarios. Furthermore, there was a temporal variation of
201
strain occurrence within the more homogenous landscape between synchronous and
202
asynchronous scenarios, as can be seen in a direct comparison of the per-strain infection
203
counts and wavelets of synchronous and asynchronous scenarios over time (figure 4). A
204
general comparison of the proportional strain contribution in asynchronous vs synchronous
205
scenarios further showed that the occurring strains were of lower virulence in asynchronous
206
scenarios in case of the more heterogeneous landscapes (Appendix Fig.B3).
207
208
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
209
210
Figure 3: Occurrence and dominance of the different virulence strains in synchronous (tlag =
211
0, top row) and asynchronous (tlag = 100, bottom row) scenarios. Colour gradient represents
212
the proportion of infected individuals with each strain in the landscape. Grey areas represent
213
zero occurrence of the strains.
214
215
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
216
Figure 4: A-Muller plot for a single example run in a large clustered landscape in
217
synchronous (left) and asynchronous (right) scenarios, showing the number of infected
218
individuals for each strain (colour) over time aggregated as annual mean. B- Wavelet
219
analysis of high virulent strains in synchronous (left) and asynchronous (right) scenarios of a
220
single example run in a large clustered landscape using the R package “wsyn” [49].
221
Discussion
222
To extend the understanding of pathogen evolution and spread during epidemics, we
223
implemented virulence evolution in an individual-based model simulating an interdependent,
224
tri-trophic system (landscape resources - host - pathogen) under the effects of global change.
225
In accordance with our hypotheses, we found an increase in pathogenic virulence and a shift
226
in strain dominance with increasing landscape homogenisation.
227
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
Landscape homogenisation alters the density distribution of susceptible host individuals by
228
increasing host connectivity, which subsequently can lead to more infection events and viral
229
mutations. The detrimental effect that density, connectivity and contact rates can have on
230
viral mutations and infection events can also be observed in “superspreader”-events of the
231
current SARS-CoV-2 pandemic (Tasakis et al. 2021). Our results support that host density
232
and connectivity are the most important factors that affect the emergence of high virulence in
233
directly transmitted diseases under classical transmission-virulence trade-offs (Castillo-
234
Chavez and Velasco-Hernández 1998).While we found lower mean virulence in scenarios
235
with asynchronous host resources, the landscape heterogeneity was the main driver of
236
virulence evolution. Interestingly, under asynchrony, we found higher proportions of low and
237
high strains coexisting in homogeneous landscapes, indicating that isolated disease hotspots
238
(Kürschner et al. 2021) could facilitate the persistence of different viral strains.
239
As long as host populations in our model are distributed heterogeneously, mean pathogenic
240
virulence remains similar, with little change from completely heterogeneous, i.e., random
241
landscapes, to the less heterogeneous medium habitat clusters. However, in large clusters, a
242
clear increase in mean virulence was apparent, showing that there is a threshold in landscape
243
homogeneity not only enhancing disease spread, but also evolution towards higher virulence.
244
These modelling findings are consistent with previous research on thresholds in disease
245
transmission and functional connectivity. For example, homogenous landscapes have been
246
shown to facilitate the spread of rabies in raccoons (Procyon lotor) (Brunker et al. 2012) or
247
tuberculosis in badgers (Meles meles) (Acevedo et al. 2019), while more heterogeneous
248
landscapes have been shown to limit the spread of highly virulent pathogens (Lane-deGraaf
249
et al. 2013 p.). Host-pathogen interactions in directly transmitted diseases occur at
250
specific locations and points in time, with the spatial and temporal variability in the
251
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
availability of susceptible hosts being one of the governing factors of a successful
252
transmission (Hudson 2002, Ostfeld et al. 2005, Real and Biek 2007). Consequently,
253
homogenous landscapes and their lack of barriers allow more virulent pathogen strains to
254
infect a sufficient number of hosts to persist in those landscapes. On the contrary, in
255
heterogeneous landscapes, small clusters of high host density in a matrix of low density cause
256
the extinction of highly virulent strains. This ‘dilution’ pattern can be explained by the short
257
survival time of individuals in the matrix that form an immunity belt around the clusters and
258
prevent spread between clusters (Marescot et al. 2021). Hence, in parallel with the dilution
259
hypotheses at the community scale, heterogeneous or ‘diverse’ landscapes provide less
260
competent hosts for an epidemic (Patz et al. 2004, Civitello et al. 2015).
261
Increasing landscape homogenisation also resulted in higher mean virulence in scenarios with
262
asynchrony between host life-history and resource availability (prediction 3). Even though
263
overall susceptible host density was lower in asynchronous scenarios, the homogenous
264
landscape’s increased connectivity allowed for higher virulent strains to persist at high
265
prevalence. In the more homogenous, but still clustered, landscape, composed of large areas
266
of high habitat suitability, the virulence of occurring strains was similar between the
267
scenarios with and without synchrony. This indicates a strong effect of landscape
268
configuration.
269
Interestingly, in our previous study (Kürschner et al. 2021) , we showed that increasing
270
spatial homogeneity of the landscape affected pathogen persistence negatively without
271
pathogen virulence evolution. One reason behind this difference lies in the temporal
272
differentiation of the strains within the landscapes. During the beginning of an outbreak, the
273
pathogen strains with low virulence are able to spread across the landscape into larger habitat
274
clusters due to the long survival times they impose on their hosts. Once the susceptible host
275
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
density in one of the neighbouring areas is high enough, highly virulent strains that
276
previously only occurred in low prevalence outcompete the low virulent strains and increase
277
in prevalence. In other words, when host density increases, the high virulence strains
278
capitalize on the high possibility for transmission and are likely to become dominant (Altizer
279
et al. 2006, Hite and Cressler 2018). However, although highly virulent strains became more
280
dominant, lower virulent strains continued to persist within the host population. In line with
281
our findings, the coexistence of high and low virulent strains was also shown for rabbit
282
haemorrhagic disease in the United Kingdom (Forrester et al. 2009) as well as influenza A in
283
wild birds (Olsen et al. 2006).
284
Furthermore, our results show that, independent of landscape heterogeneity, a single, low
285
virulent strain of a pathogen is able to evolve into a complex system of multiple coexisting
286
strains with varying virulence (prediction 1). However, while multiple strains coexisted at
287
any given time throughout all tested scenarios, we demonstrated that some strains likely
288
become dominant (prediction 2). Similarly, a system of coexisting low and highly virulent
289
strains were reported by empirical studies of the African swine fever virus in wild boar
290
(Portugal et al. 2015), a pathogen causing severe diseases with huge economic impact (Artois
291
et al. 2002). In this system the carriers of low virulent strains could remain infectious over
292
long periods of time (de Carvalho Ferreira et al. 2012) increasing the chance of the pathogen
293
transmission and its mutation into higher virulent strains, which could become dominant over
294
time. In our study, the virulence of the dominant strain was intrinsically linked to the degree
295
of landscape homogenisation but was also variable in time. Our findings are consistent with
296
theoretical models that showed an increase of pathogenic virulence over time (Osnas et al.
297
2015). However, while Osnas et al. (2015) assumed a direct trade-off between virulence and
298
host movement in homogenous landscapes, here we show that different landscape
299
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
configurations may lead to the same patterns of increasing virulence without the necessity of
300
such a trade-off.
301
On the one hand, our results show that with natural landscapes becoming more fragmented
302
and resources becoming more asynchronous due to global change, a shift towards lower
303
virulent pathogens could be expected. As a consequence, some diseases may become
304
endemic in their respective host populations. The longer a pathogen is able to persist within
305
its host population the higher the risk for spontaneous mutations and the possibility of
306
spillovers to other species. On the other hand, global change will lead to increasing
307
homogenisation within those fragments (Patz et al. 2004) and has the potential to increase the
308
average pathogenic virulence with possibly catastrophic effects on wildlife communities. A
309
large variance in virulence has been shown among infected host individuals, where the
310
infection can range from severe to asymptomatic. This variation can be the result of a variety
311
of factors, including genetic variation or intraspecific host interactions but also environmental
312
conditions (Ebert and Bull 2003). Furthermore, an increase in virulence will go hand in hand
313
with higher transmission rates in many diseases (Messinger and Ostling 2009, Alizon and
314
Michalakis 2015) that will increase the probability of pathogen spillovers even more. While
315
pathogen spillovers to other wild or domestic animal populations can have profound social or
316
economic effects (Kamo et al. 2007), the possibly detrimental effects on human health cannot
317
be underestimated. The current SARS-CoV-2 pandemic clearly highlights the importance of
318
understanding which factors govern the spread of diseases in wildlife populations and how
319
anthropogenic changes may alter those in the future.
320
321
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
Methods
322
Model overview
323
We modified a spatially explicit individual-based, eco-epidemiological model developed by
324
Kürschner et al. ( 2021). It is based on earlier models considering neighbourhood infections
325
only that was developed by Kramer-Schadt et al. (2009), Lange et al. ( 2012a, b) and Scherer
326
et al. ( 2020) and includes spatiotemporal landscape dynamics representing changing resource
327
availability, coupled with resource-based mortality. We incorporated evolution of viral traits
328
such as virulence and corresponding trade-offs with viral transmission (see below). A
329
complete and detailed model description following the ODD (Overview, Design concepts,
330
Detail) protocol (Grimm et al. 2006, 2010) is provided in the supplementary material and the
331
model (implementation) in the Zenodo Database and on GitHub [links provided on
332
acceptance].
333
334
The model comprises three main components, a host model depending on underlying
335
landscape features, an epidemiological pathogen model and a pathogen evolutionary model.
336
Host individuals are characterised by sex, age, location, demographic status (residential,
337
dispersing) and epidemiological status (susceptible, infected, immune). The epidemiological
338
status of the individuals is defined by an SIR epidemiological classification (susceptible,
339
infected, and recovered; Kermack and McKendrick 1927)). The pathogen is characterized by
340
strain type, virulence and transmission. The pathogen model alters host survival rates and
341
infection length depending on the pathogen’s virulence, while the dynamic landscape features
342
determine host reproductive success. We record strain occurrences as the number of infected
343
individuals carrying a specific strain and pathogen persistence, measured at the level of
344
simulation runs (see below).
345
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
Pathogen dynamics
346
We determined the course of the disease by an age-specific case fatality rate and a strain-
347
specific infectious period. Highly virulent strains are characterized by a short infectious
348
period and low virulent strains by a long infectious period. Transiently infected hosts shed the
349
pathogen for one week and gain lifelong immunity (Dahle and Liess 1992). Infection
350
dynamics emerge from multiple processes: within-group transmission and individual age-
351
dependent courses of infection. Within groups, the density-dependent infection pressure (i.e.
352
the chance of a host individual to become infected) is determined by a transmission chance
353
and the number of infectious group members carrying the same strain. In this model we
354
included the dependence of the transmission chance on the strain’s virulence, so that the
355
strains with higher virulence have higher transmission chance. Furthermore, we modified the
356
density dependence of the infection pressure to be strain-specific to accommodate a lower
357
per-strain infection density for the following reason: The original model based on a single
358
pathogen strain used the density of infected individuals in a cell to infer the likelihood for a
359
susceptible host in that cell to become infected based on a binomial model.. Our model
360
allows the evolution into 12 (arbitrarily categorized) different viral strains. The infection
361
pressure λ, i.e. the probability of pathogen transmission to a susceptible host individual, is
362
determined for each strain individually. Differences in strain transmissibility are added to the
363
strain specific infection pressure through Ts (1). The probability λis of an individual i of being
364
infected by a specific strain s is calculated as
365
 1 󰇛1 󰇜 󰇛1
10 󰇜  (1)
366
with being the individual probability of transmission to the power of all infected
367
individuals Ijs in a group j per strain s as well as a reduced transmission probability between
368
groups (i.e. cells)
10 to the power of all infected individuals in neighboring groups .
369
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
The strain virulence translates directly into infection length, i.e., host survival time, where a
370
high virulence results in shorter survival times for the host compared to low-virulence.
371
Consequently, the shorter lifetime of a highly virulent pathogen results in a shorter
372
reproductive time span, while making the pathogen highly infective.
373
Evolution of pathogenic traits Virulence and transmission are emergent properties and are
374
evolving in the model. This means, while the position of each of the 12 strains on the
375
transmission trade-off-curve is fixed, the selection of each strain during a transmission event
376
is variable. Our trade-off curve is modelled to follow theoretical transmission-virulence
377
trade-off curves (Alizon et al. 2009) and is applied for each infected host individually.
378
During a transmission event, a strain can, with a mutation rate of 0.01, mutate into a new
379
strain with a different virulence. The virulence of the new strain is selected from a normal
380
distribution with a standard deviation σ = 1 around the virulence value of the originally
381
transmitted strain, meaning that the new strain will be closely related to the parental strain.
382
Landscape structure and dynamics
383
The tested landscapes consist of a spatial grid of 1.250 2 km x 2 km cells, each representing
384
the average home range of a social host, e.g. a wild boar group (Kramer-Schadt et al. 2009) ,
385
totalling a 100 km x 50 km landscape. The landscapes are self-contained systems without any
386
outside interaction. Each cell is characterized by a variable resource availability that
387
represents host breeding capacity and translates directly into host group size, with the
388
minimum being one breeding female per group to a maximum of nine. Resource availability
389
was adapted to achieve the average wild boar density of five breeding females per km2
390
(Howells and Edwards-Jones 1997, Sodeikat and Pohlmeyer 2003, Melis et al. 2006). We
391
investigated several landscape scenarios of varying spatial complexity, ranging from a fully
392
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
random landscape structure to different degrees of random landscape clusters generated in R
393
(R Core Team 2020) using the NLMR package (Sciaini et al. 2018) up to a fully
394
homogeneous landscape. To exclude any biases that could stem from different host densities,
395
the mean female breeding capacity was kept constant at five females per km2 across the
396
different landscape types (Supplementary material Appendix Fig. B1). The spatiotemporal
397
landscape dynamics that were designed to mimic seasonal changes in resource availability by
398
gradually increasing and decreasing resource availability were kept unchanged from the
399
previous model implementation by Kürschner et al. ( 2021).
400
Process overview and scheduling
401
The temporal resolution of the model equals the approximate pathogen incubation time of
402
one week (Artois et al. 2002). The model procedures were scheduled each step in the
403
following order: pathogen transmission, pathogen evolution, natal host group split of subadult
404
males and females, resource-based host dispersal, host reproduction, baseline host mortality,
405
strain-based host mortality, resource-based host mortality, host ageing and landscape
406
dynamics. Natal group split of males and females was limited to week 17 and week 29 of
407
each year, respectively, representing the observed dispersal time for each sex.
408
Host mortality Mortality in response to resource availability remained unchanged to the
409
previous model implementation (for details see ODD in the supplementary material).
410
Additionally, we added a fixed, strain-specific mortality for each strain that affects the host
411
population.
412
Landscape dynamics with temporal lag We modelled two levels of temporal lag (tlag)
413
implemented in Kürschner et al. (2021) . We focus on the level 0% (synchrony between host
414
population dynamics and resource availability) to 100% (asynchrony between host population
415
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
dynamics and resource availability), with the latter simulating phenological mismatch
416
between the resources and hosts reproduction potentially due to climate change. The extreme
417
values were chosen because previous studies investigating temporal lag did not show strong
418
effects in the intermediary steps (Kürschner et al. 2021).
419
Model analysis
420
Each simulation was run for 100 years in total, with the virus released in a randomly taken
421
week of the second year (week 53104), to allow the population to stabilize after
422
initialization. The virus was introduced to a set of multiple predefined cells in the centre of
423
the landscape to ensure an outbreak. The virus was released in a low virulence variant. We
424
ran 25 repetitions per combination of landscape scenarios (5 levels: small clusters, medium
425
clusters, large clusters, homogenous landscape and random) and asynchrony (2 levels: tlag 0%,
426
tlag 100%). We also analysed the strain occurrence (i.e., if a strain was present in any
427
landscape cell, recorded at every timestep) and number of infected hosts per strain at every
428
timestep to measure strain extinction as well as reappearance through mutation. We further
429
recorded the proportion that each strain contributed to the pool of infected hosts by
430
calculating the ratio of the hosts infected with each strain to the total number of hosts infected
431
with all strains, at each time step. To highlight differences in strain composition in those
432
scenarios, we subtracted the mean strain proportion in asynchronous scenarios from the mean
433
proportion in synchronous scenarios. We categorized all viral strains into three categories:
434
low virulence strains; medium virulence strains; high virulence strains, each compartment
435
summing the outcomes of 4 of the 12 strains modelled.
436
437
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
Acknowledgements
438
This work was supported by the German Research Foundation (DFG) in the framework of the
439
BioMove Research Training Group (DFG-GRK 2118/1). We thank Volker Grimm for
440
valuable comments on earlier drafts of this manuscript and Florian Jeltsch and Heribert Hofer
441
for helpful discussions.
442
Statement of authorship
443
All authors agree to submission of the manuscript, and each author carries a degree of
444
responsibility for the accuracy, integrity and ethics of the manuscript and works described
445
therein.
446
Author contributions
447
TK and SKS developed the core idea and designed the study. TK rewrote and modified the
448
simulation model together with CS and SKS. TK, VR, and SKS analysed the simulation
449
results. TK is the lead author and CS, VR, NB and SKS contributed substantially to the
450
writing. All authors agreed to submission of the manuscript, and each author is accountable
451
for the aspects of the conducted work and ensures that questions related to the accuracy or
452
integrity of any part of the work are appropriately investigated and resolved.
453
454
References
455
Acevedo, P. et al. 2019. Tuberculosis Epidemiology and Badger (Meles meles) Spatial
456
Ecology in a Hot-Spot Area in Atlantic Spain. - Pathogens 8: 292.
457
Alizon, S. and Michalakis, Y. 2015. Adaptive virulence evolution: the good old fitness-based
458
approach. - Trends in Ecology & Evolution 30: 248254.
459
Alizon, S. et al. 2009. Virulence evolution and the trade-off hypothesis: history, current state
460
of affairs and the future. - Journal of evolutionary biology 22: 245259.
461
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
Altizer, S. et al. 2006. Seasonality and the dynamics of infectious diseases. - Ecology letters
462
9: 467484.
463
Anderson, R. M. and May, R. M. 1982. Coevolution of hosts and parasites. - Parasitology 85:
464
411426.
465
Artois, M. et al. 2002. Classical swine fever (hog cholera) in wild boar in Europe. - Rev. Sci.
466
Tech. OIE 21: 287303.
467
Boots, M. 2004. Large Shifts in Pathogen Virulence Relate to Host Population Structure. -
468
Science 303: 842844.
469
Boots, M. and Mealor, M. 2007. Local Interactions Select for Lower Pathogen Infectivity. -
470
Science 315: 12841286.
471
Brunker, K. et al. 2012. Integrating the landscape epidemiology and genetics of RNA
472
viruses: rabies in domestic dogs as a model. - Parasitology 139: 18991913.
473
Castillo-Chavez, C. and Velasco-Hernández, J. X. 1998. On the Relationship Between
474
Evolution of Virulence and Host Demography. - Journal of Theoretical Biology 192:
475
437444.
476
Choua, M. and Bonachela, J. A. 2019. Ecological and Evolutionary Consequences of Viral
477
Plasticity. - The American Naturalist 193: 346358.
478
Civitello, D. J. et al. 2015. Biodiversity inhibits parasites: Broad evidence for the dilution
479
effect. - Proc. Natl. Acad. Sci. U.S.A. 112: 86678671.
480
Cressler, C. E. et al. 2016. The adaptive evolution of virulence: a review of theoretical
481
predictions and empirical tests. - Parasitology 143: 915930.
482
Dahle, J. and Liess, B. 1992. A review on classical swine fever infections in pigs:
483
Epizootiology, clinical disease and pathology. - Comparative Immunology,
484
Microbiology and Infectious Diseases 15: 203211.
485
Day, T. 2003. Virulence evolution and the timing of disease life-history events. - Trends in
486
Ecology & Evolution 18: 113118.
487
Day, T. and Gandon, S. 2007. Applying population-genetic models in theoretical evolutionary
488
epidemiology. - Ecol Letters 10: 876888.
489
de Carvalho Ferreira, H. C. et al. 2012. African swine fever virus excretion patterns in
490
persistently infected animals: A quantitative approach. - Veterinary Microbiology 160:
491
327340.
492
Adaptive dynamics of infectious diseases: in pursuit of virulence management 2005. (U
493
Dieckmann, JAJ Metz, M Sabelis W, and K Sigmund, Eds.). - Cambridge University
494
Press.
495
Ebert, D. and Bull, J. J. 2003. Challenging the trade-off model for the evolution of virulence:
496
is virulence management feasible? - Trends in Microbiology 11: 1520.
497
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
Forrester, N. L. et al. 2009. Co-circulation of widely disparate strains of Rabbit haemorrhagic
498
disease virus could explain localised epidemicity in the United Kingdom. - Virology
499
393: 4248.
500
Galvani, A. P. 2003. Epidemiology meets evolutionary ecology. - Trends in Ecology &
501
Evolution 18: 132139.
502
Geoghegan, J. L. and Holmes, E. C. 2018. The phylogenomics of evolving virus virulence. -
503
Nat Rev Genet 19: 756769.
504
Griette, Q. et al. 2015. Virulence evolution at the front line of spreading epidemics. -
505
Evolution 69: 28102819.
506
Grimm, V. et al. 2006. A standard protocol for describing individual-based and agent-based
507
models. - Ecological Modelling 198: 115126.
508
Grimm, V. et al. 2010. The ODD protocol: A review and first update. - Ecological Modelling
509
221: 27602768.
510
Hite, J. L. and Cressler, C. E. 2018. Resource-driven changes to host population stability
511
alter the evolution of virulence and transmission. - Phil. Trans. R. Soc. B 373:
512
20170087.
513
Howells, O. and Edwards-Jones, G. 1997. A feasibility study of reintroducing wild boar Sus
514
scrofa to Scotland: Are existing woodlands large enough to support minimum viable
515
populations. - Biological Conservation 81: 7789.
516
The ecology of wildlife diseases 2002. (PJ Hudson, Ed.). - Oxford University Press.
517
Kamo, M. et al. 2007. The role of trade-off shapes in the evolution of parasites in spatial host
518
populations: An approximate analytical approach. - Journal of Theoretical Biology
519
244: 588596.
520
Kermack, W. O. and McKendrick, A. G. 1927. A Contribution to the Mathematical Theory of
521
Epidemics. - Proceedings of the Royal Society A: Mathematical, Physical and
522
Engineering Sciences 115: 700721.
523
Kramer-Schadt, S. et al. 2009. Individual variations in infectiousness explain long-term
524
disease persistence in wildlife populations. - Oikos 118: 199208.
525
Kürschner, T. et al. 2021. Movement can mediate temporal mismatches between resource
526
availability and biological events in hostpathogen interactions. - Ecol. Evol. 11:
527
57285741.
528
Lane-deGraaf, K. E. et al. 2013. A test of agent-based models as a tool for predicting
529
patterns of pathogen transmission in complex landscapes. - BMC Ecol 13: 35.
530
Lange, M. et al. 2012a. Disease severity declines over time after a wild boar population has
531
been affected by classical swine feverlegend or actual epidemiological process? -
532
Preventive veterinary medicine 106: 185195.
533
Lange, M. et al. 2012b. Efficiency of spatio-temporal vaccination regimes in wildlife
534
populations under different viral constraints. - Veterinary Research 43: 37.
535
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
Lebarbenchon, C. et al. 2008. Evolution of pathogens in a man-made world. - Molecular
536
Ecology 17: 475484.
537
Lenski, R. E. and May, R. M. 1994. The Evolution of Virulence in Parasites and Pathogens:
538
Reconciliation Between Two Competing Hypotheses. - Journal of Theoretical Biology
539
169: 253265.
540
Marescot, L. et al. 2021. ‘Keeping the kids at home’ can limit the persistence of contagious
541
pathogens in social animals. - J Anim Ecol 90: 25232535.
542
Melis, C. et al. 2006. Biogeographical variation in the population density of wild boar (Sus
543
scrofa) in western Eurasia. - J Biogeography 33: 803811.
544
Messinger, S. M. and Ostling, A. 2009. The consequences of spatial structure for the
545
evolution of pathogen transmission rate and virulence. - The American naturalist 174:
546
441454.
547
Olsen, B. et al. 2006. Global Patterns of Influenza A Virus in Wild Birds. - Science 312: 384
548
388.
549
Osnas, E. E. et al. 2015. Evolution of Pathogen Virulence across Space during an Epidemic.
550
- The American Naturalist 185: 332342.
551
Ostfeld, R. et al. 2005. Spatial epidemiology: an emerging (or re-emerging) discipline. -
552
Trends in Ecology & Evolution 20: 328336.
553
Patz, J. A. et al. 2004. Unhealthy Landscapes: Policy Recommendations on Land Use
554
Change and Infectious Disease Emergence. - Environmental Health Perspectives
555
112: 10921098.
556
Portugal, R. et al. 2015. Related strains of African swine fever virus with different virulence:
557
genome comparison and analysis. - Journal of General Virology 96: 408419.
558
R Core Team 2020. R: A Language and Environment for Statistical Computing. - R
559
Foundation for Statistical Computing.
560
Real, L. A. and Biek, R. 2007. Spatial dynamics and genetics of infectious diseases on
561
heterogeneous landscapes. - J. R. Soc. Interface. 4: 935948.
562
Scherer, C. et al. 2020. Moving infections: individual movement decisions drive disease
563
persistence in spatially structured landscapes. - Oikos: oik.07002.
564
Sciaini, M. et al. 2018. NLMR and landscapetools : An integrated environment for simulating
565
and modifying neutral landscape models in R (N Golding, Ed.). - Methods Ecol Evol
566
9: 22402248.
567
Sodeikat, G. and Pohlmeyer, K. 2003. Escape movements of family groups of wild boar Sus
568
scrofa influenced by drive hunts in Lower Saxony, Germany. - Wildlife Biology 9: 43
569
49.
570
Tasakis, R. N. et al. 2021. SARS-CoV-2 variant evolution in the United States: High
571
accumulation of viral mutations over time likely through serial Founder Events and
572
mutational bursts (YE Khudyakov, Ed.). - PLoS ONE 16: e0255169.
573
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
Visher, E. et al. 2021. The three Ts of virulence evolution during zoonotic emergence. - Proc.
574
R. Soc. B. 288: 20210900.
575
Wilcox, B. A. and Gubler, D. J. 2005. Disease ecology and the global emergence of zoonotic
576
pathogens. - Environ Health Prev Med 10: 263.
577
578
.CC-BY 4.0 International licensemade available under a
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
The copyright holder for this preprintthis version posted August 5, 2022. ; https://doi.org/10.1101/2022.08.03.502590doi: bioRxiv preprint
ResearchGate has not been able to resolve any citations for this publication.
Full-text available
Article
Since the first case of COVID-19 in December 2019 in Wuhan, China, SARS-CoV-2 has spread worldwide and within a year and a half has caused 3.56 million deaths globally. With dramatically increasing infection numbers, and the arrival of new variants with increased infectivity, tracking the evolution of its genome is crucial for effectively controlling the pandemic and informing vaccine platform development. Our study explores evolution of SARS-CoV-2 in a representative cohort of sequences covering the entire genome in the United States, through all of 2020 and early 2021. Strikingly, we detected many accumulating Single Nucleotide Variations (SNVs) encoding amino acid changes in the SARS-CoV-2 genome, with a pattern indicative of RNA editing enzymes as major mutators of SARS-CoV-2 genomes. We report three major variants through October of 2020. These revealed 14 key mutations that were found in various combinations among 14 distinct predominant signatures. These signatures likely represent evolutionary lineages of SARS-CoV-2 in the U.S. and reveal clues to its evolution such as a mutational burst in the summer of 2020 likely leading to a homegrown new variant, and a trend towards higher mutational load among viral isolates, but with occasional mutation loss. The last quartile of 2020 revealed a concerning accumulation of mostly novel low frequency replacement mutations in the Spike protein, and a hypermutable glutamine residue near the putative furin cleavage site. Finally, end of the year data and 2021 revealed the gradual increase to prevalence of known variants of concern, particularly B.1.1.7, that have acquired additional Spike mutations. Overall, our results suggest that predominant viral genomes are dynamically evolving over time, with periods of mutational bursts and unabated mutation accumulation. This high level of existing variation, even at low frequencies and especially in the Spike-encoding region may become problematic when super-spreader events, akin to serial Founder Events in evolution, drive these rare mutations to prominence.
Full-text available
Article
Social networks are considered to be ‘highly modular’ when individuals within one module are more connected to each other than they are to individuals in other modules. It is currently unclear how highly modular social networks influence the persistence of contagious pathogens that generate lifelong immunity in their hosts when between‐group interactions are age‐dependent. This trait occurs in social species with communal nurseries, where juveniles are reared together for a substantial period in burrows or similar forms of containment and are thus in isolation from contact with individuals in other social groups. Our main objective was to determine whether, and to what extent, such age‐dependent patterns of between‐group interactions consistently increased the fade‐out probability of contagious pathogens that generate lifelong immunity in their hosts. We hypothesised that in populations of species where juveniles are raised in communal nurseries, a high proportion of recovered adults in a group would form a ‘protective barrier’ around susceptible juveniles against pathogen transmission, thereby increasing the probability of epidemic fade‐out in the population. To test this idea we implemented a spatially implicit individual‐based Susceptible‐Infected‐Recovered (SIR) model for a large range of generic host and pathogen traits. Our results indicate that (i) the probability of epidemic fade‐out was consistently higher in populations with communal nurseries, especially for highly contagious pathogens (high basic reproduction number, R0) and (ii) communal nurseries can counteract the cost of group‐living in terms of infection risk to a greater extent than variation in other traits. We discuss our findings in relation to herd immunity and outline the importance of considering the network structure of a given host population before implementing management measures such as vaccinations, since interventions focused on individuals with high between‐group contact should be particularly effective for controlling pathogen spread in hosts with communal nurseries.
Full-text available
Article
Global change is shifting the timing of biological events, leading to temporal mis-matches between biological events and resource availability. These temporal mis-matches can threaten species' populations. Importantly, temporal mismatches not only exert strong pressures on the population dynamics of the focal species, but can also lead to substantial changes in pairwise species interactions such as host-pathogen systems. We adapted an established individual-based model of host-pathogen dynamics. The model describes a viral agent in a social host, while accounting for the host's explicit movement decisions. We aimed to investigate how temporal mismatches between seasonal resource availability and host life-history events affect host-pathogen coexistence, that is, disease persistence. Seasonal resource fluctuations only increased coexistence probability when in synchrony with the hosts' biological events. However, a temporal mismatch reduced host-pathogen coexistence, but only marginally. In tandem with an increasing temporal mismatch, our model showed a shift in the spatial distribution of infected hosts. It shifted from an even distribution under synchronous conditions toward the formation of disease hotspots, when host life history and resource availability mismatched completely. The spatial restriction of infected hosts to small hotspots in the landscape initially suggested a lower coexistence probability due to the critical loss of susceptible host individuals within those hotspots. However, the surrounding landscape facilitated demographic rescue through habitat-dependent movement. Our work demonstrates that the negative effects of temporal mismatches between host resource availability and host life history on host-pathogen coexistence can be reduced through the formation of temporary disease hotspots and host movement decisions, with implications for disease management under disturbances and global change.
Full-text available
Article
: We provide a temporal overview (from 2012 to 2018) of the outcomes of tuberculosis (TB) in the cattle and badger populations in a hot-spot in Asturias (Atlantic Spain). We also study the badger’s spatial ecology from an epidemiological perspective in order to describe hazardous behavior in relation to TB transmission between cattle and badgers. Culture and single intradermal tuberculin test (SITT) were available for cattle as part of the National Program for the Eradication of TB. A field survey was also carried out in order to determine the paddocks and buildings used by each farm, and the information obtained was stored by using geographic information systems. Moreover, eighty-three badgers were submitted for necropsy and subsequent bacteriological studies. Ten badgers were also tracked, using global positioning system (GPS) collars. The prevalence of TB in cattle herds in the hot-spot increased from 2.2% in 2012 to 20% in 2016; it then declined to 0.0% in 2018. In contrast, the TB prevalence in badgers increased notably (from 5.55% in 2012–2015 to 10.64% in 2016–2018). Both cattle and badgers shared the same strain of Mycobacterium bovis. The collared badgers preferred paddocks used by TB-positive herds in spring and summer (when they were more active). The males occupied larger home ranges than the females (Khr95: males 149.78 ± 25.84 ha and females 73.37 ± 22.91 ha; Kcr50: males 29.83 ± 5.69 ha and females 13.59 ± 5.00 ha), and the home ranges were smaller in autumn and winter than in summer. The averages of the index of daily and maximum distances traveled by badgers were 1.88 ± (SD) 1.20 km and 1.99 ± 0.71 km, respectively. One of them presented a dispersive behavior with a maximum range of 18.3 km. The most preferred habitat was apple orchards in all seasons, with the exception of winter, in which they preferred pastures. Land uses and landscape structure, which have been linked with certain livestock-management practices, provide a scenario of great potential for badger–cattle interactions, thus enhancing the importance of the badgers’ ecology, which could potentially transmit TB back to cattle in the future.
Full-text available
Article
1. Neutral landscape models (NLMs) simulate landscape patterns based on theoretical distributions and can be used to systematically study the effect of landscape structure on ecological processes. NLMs are commonly used in landscape ecology to enhance the findings of field studies as well as in simulation studies to provide an underlying landscape. However, their creation so far has been limited to software that is platform dependent, does not allow a reproducible workflow or is not embedded in R, the prevailing programming language used by ecologists. 2. Here, we present two complementary R packages NLMR and landscapetools, that allow users to generate, manipulate and analyze NLMs in a single environment. They grant the simulation of the widest collection of NLMs found in any single piece of software thus far while allowing for easy manipulation in a self-contained and reproducible workflow. The combination of both packages should stimulate a wider usage of NLMs in landscape ecology. NLMR is a comprehensive collection of algorithms with which to simulate NLMs. landscapetools provides a utility toolbox which facilitates an easy workflow with simulated neutral landscapes and other raster data. 3. We show two example applications that illustrate potential use cases for NLMR and landscapetools: First, an agent-based simulation study in which the effect of spatial structure on disease persistence was studied. Here, spatial heterogeneity resulted in more variable disease outcomes compared to the common well-mixed host assumption. The second example shows how increases in spatial scaling can introduce biases in calculated landscape metrics. 4. Simplifying the workflow around handling NLMs should encourage an uptake in the usage of NLMs. NLMR and landscapetools are both generic frameworks that can be used in a variety of applications and are a further step to having a unified simulation environment in R for answering spatial research questions.
Full-text available
Article
What drives the evolution of parasite life-history traits? Recent studies suggest that linking within- and between-host processes can provide key insight into both disease dynamics and parasite evolution. Still, it remains difficult to understand how to pinpoint the critical factors connecting these cross-scale feedbacks, particularly under non-equilibrium conditions; many natural host populations inherently fluctuate and parasites themselves can strongly alter the stability of host populations. Here, we develop a general model framework that mechanistically links resources to parasite evolution across a gradient of stable and unstable conditions. First, we dynamically link resources and between-host processes (host density, stability, transmission) to virulence evolution, using a ‘non-nested’ model. Then, we consider a ‘nested’ model where population-level processes (transmission and virulence) depend on resource-driven changes to individual-level (within-host) processes (energetics, immune function, parasite production). Contrary to ‘non-nested’ model predictions, the ‘nested’ model reveals complex effects of host population dynamics on parasite evolution, including regions of evolutionary bistability; evolution can push parasites towards strongly or weakly stabilizing strategies. This bistability results from dynamic feedbacks between resource-driven changes to host density, host immune function and parasite production. Together, these results highlight how cross-scale feedbacks can provide key insights into the structuring role of parasites and parasite evolution. This article is part of the theme issue ‘Anthropogenic resource subsidies and host–parasite dynamics in wildlife’.
Article
Understanding host–pathogen dynamics requires realistic consideration of transmission events that, in the case of directly transmitted pathogens, result from contacts between susceptible and infected individuals. The corresponding contact rates are usually heterogeneous due to variation in individual movement patterns and the underlying landscape structure. However, in epidemiological models, the roles that explicit host movements and landscape structure play in shaping contact rates are often overlooked. We adapted an established agent‐based model of classical swine fever (CSF) in wild boar Sus scrofa to investigate how explicit representation of landscape heterogeneity and host movement between social groups affects invasion and persistence probabilities. We simulated individual movement both phenomenologically as a correlated random walk (CRW) and mechanistically by representing interactions of the moving individuals with the landscape and host population structure. The effect of landscape structure on the probability of invasion success and disease persistence depended remarkably on the way host movement is simulated and the case fatality ratio associated with the pathogen strain. The persistence probabilities were generally low with CRW which ignores feedbacks to external factors. Although the basic reproduction number R0, a measure of the contagiousness of an infectious disease, was kept constant, these probabilities were up to eight times higher under mechanistic movement rules, especially in heterogeneous landscapes. The increased persistence emerged due to important feedbacks of the directed movement on the spatial variation of host density, contact rates, and transmission events to distant areas. Our findings underscore the importance of accounting for spatial context and group size structures in eco‐epidemiological models. Our study highlights that the simulation of explicit, mechanistic movement behaviour can reverse predictions of disease persistence in comparison to phenomenological rules such as random walk approaches. This can have severe consequences when predicting the probability of disease persistence and assessing control measures to prevent outbreaks. This article is protected by copyright. All rights reserved.
Article
Viruses use the host machinery to replicate, and their performance thus depends on the host's physiological state. For bacteriophages, this link between host and viral performance has been characterized empirically and with intracellular theories. Such theories are too detailed to be included in models that study host-phage interactions in the long term, which hinders our understanding of systems that range from pathogens infecting gut bacteria to marine phage shaping the oceans. Here, we combined data and models to study the short- and long-term consequences that host physiology has on bacteriophage performance. We compiled data showing the dependence of lytic-phage traits on host growth rate (referred to as viral phenotypic plasticity) to deduce simple expressions that represent such plasticity. Including these expressions in a standard host-phage model allowed us to understand mechanistically how viral plasticity affects emergent evolutionary strategies and the population dynamics associated with different environmental scenarios including, for example, nutrient pulses or host starvation. Moreover, we show that plasticity on the offspring number drives the phage ecological and evolutionary dynamics by reinforcing feedbacks between host, virus, and environment. Standard models neglect viral plasticity, which therefore handicaps their predictive ability in realistic scenarios. Our results highlight the importance of viral plasticity to unravel host-phage interactions and the need of laboratory and field experiments to characterize viral plastic responses across systems.
Article
The virulence of viruses is a major determinant of the health burden of viral infections in humans and other species. In this article, Geoghegan and Holmes discuss how largely disparate research fields — theoretical modelling of virulence evolution and experimental dissection of genetic virulence determinants in laboratory model systems — can be bridged by considering real genomic data of viral evolution in a phylogenetic context. They describe the underlying principles of virulence evolution and how they apply to real-world viral infections and outbreaks of global importance.
Article
Understanding and predicting the spatial spread of emerging pathogens is a major challenge for the public health management of infectious diseases. Theoretical epidemiology shows that the speed of an epidemic is governed by the life history characteristics of the pathogen and its ability to disperse. Rapid evolution of these traits during the invasion may thus affect the speed of epidemics. Here we study the influence of virulence evolution on the spatial spread of an epidemic. At the edge of the invasion front, we show that more virulent and transmissible genotypes are expected to win the competition with other pathogens. Behind the front line, however, more prudent exploitation strategies outcompete virulent pathogens. Crucially, even when the presence of the virulent mutant is limited to the edge of the front, the invasion speed can be dramatically altered by pathogen evolution. We support our analysis with individual-based simulations and we discuss the additional effects of demographic stochasticity taking place at the front line on virulence evolution. We confirm that an increase of virulence can occur at the front, but only if the carrying capacity of the invading pathogen is large enough. These results are discussed in the light of recent empirical studies examining virulence evolution at the edge of spreading epidemics. This article is protected by copyright. All rights reserved.