The document summarizes research from Work Package 3 of the Phytothreats project. The research analyzed pathogen traits and global distribution data to identify and rank threats from Phytophthora species to the UK. Key findings include: (1) Species that are cold-adapted are more likely to be introduced through the live plant trade; (2) Thermal tolerance traits influence a species' ability to establish at higher latitudes; (3) Models using global occurrence data and environmental factors accurately predict UK habitat suitability. Traits like cold tolerance and symptoms modulated a species' global spread and host range. The research aims to develop tools to inform policies and practices around Phytophthora importation, establishment, and
1. Beth Purse beth@ceh.ac.uk, Louise Barwell, Dan Chapman,
Ana Perez-Sierra, David Cooke, Sarah Green, Mariella
Marzano, Michael Dunn,
Phytothreats: WP3 highlights
Linking global spread and impact of
Phytophthoras to biological traits, trade
and travel, suitable habitat and climate
2. Pathogens as Invaders: Invasion Stages & Barriers
Which Phytophthora species can overcome
barriers to cause impact? Does this depend on
their biological characteristics or traits?
3. Phytophthoras spread by trade and travel, sensitive
to climatic and habitat conditions
Agricultural
trade flows
into Europe
(source)
Growth rate at different
temperatures
4. WP3: Analyse pathogen behaviour and traits to
identify and rank global Phytophthora threats to UK
What range of hosts and sectors could be
impacted?
How are Phytophthoras introduced into
new areas, from which key source areas,
by which trade pathways?
Which Phytophthoras establish in the
wider environment, does this depend on
traits? How much of UK is at risk?
Which Phytophthoras arrive? Does this
depend on traits?
Could tourism provide a potential
mechanism for spread and how could this
be investigated?***
5. Culture collections Citizen Science Projects
Governmental bodies
Global distribution databases
Global Phytophthora distribution: data sources
Published records
Diagnostic laboratories Sequence data repositoriesConsultants
Researchers
Agricultural
Forest / forestry
Nursery / ornamental
6. Global Phytophthora distribution database
39 888 country level records
1585 species x country new detections
27 706 site level records
179 countries
173 species
107 expert contributors
7. • 179 species (8 hybrids, 13
provisionally named)
• Phylogeny included
Phytophthora trait database
Peter Scott Nari Williams Giles Hardy Treena Burgess
Ana Perez-Sierra
Sarah Green
Beatrice Henricot
David Cooke (JHI)
8. Phytophthora traits included in the database
• Arrival / establishment
o Oospores or not
o Oospore wall width
o chlamydospores
o hyphal swellings
o proliferating sporangia
• Establishment
o thermal tolerance for growth
o growth rate at optimum temperature
o minimum temperature for growth
• Spread
o caducous sporangia (aerial spread)
• Impact
o root disease (below-ground)
o foliar disease (above ground)
9. Risk of introduction: countries where arrived
75 countries with ≥ 1 documented Phytophthora arrival since 2000
How are Phytophthoras introduced into new areas, from which key
source areas, by which trade pathways?
Which Phytophthoras arrive? Does this depend on traits?
species X country arrival
12. National recording and biosecurity effort
Early, R., Bradley, B., Dukes, J. et al. Nat Commun 7, 12485 (2016)
13. Introductions more likely where countries are
connected to source regions by the live plant trade
Estimate z P
Total imports -0.188 -0.952 0.341
Network connectivity 1.330 7.516 <0.001
Network connectivity explains
21% of the variation in arrival
14. Some species better able to exploit trade pathways
than others
Graph of key interactions
between trade and traits please!
15. Risk of introduction: key findings
How are Phytophthoras introduced into new areas, from which key
source areas, by which trade pathways?
Which Phytophthoras arrive? Does this depend on traits?
Arrival
All commodities Agricultural Horticultural Live plants
Species that are cold-adapted (lower min. temp for growth) more
likely to arrive through live plant trade
Logtradevolumes
16. Risk of Phytophthora importation to UK: Key source
countries and pathogens
Excerpt from your interactive
map for the UK please!
17. Risk of establishment: escape from nurseries UK
Key data sources
THDAS (268 records)
SASA (820 records)
eDOMERO P type (12,596)
SOD inspection (28,515)
RHS (1595)
Premise types:
sites of interception: ports, nurseries,
garden centres, shops, warehouses
sites of onward spread: private and
public gardens, amenity planting,
woodland and heathland, landscaping
Which Phytophthoras establish in the wider environment, does this
depend on traits?
18. Risk of establishment: escape from nurseries UK
Which Phytophthoras establish in the wider environment?
S
Species in forests:
subset of those in
nurseries and
gardens
Species
Numberofrecords
19. Risk of establishment: escape from nurseries UK
Species thermal traits seem to affect the latitudinal range of spread
Cold-tolerance – linked
to establishment at
higher latitudes
elsewhere in Europe
(Redondo et al. 2018)
20. Risk of establishment: global niche models
Global niche models
How much of UK is at risk of establishment?
21. Risk of establishment: global niche models
How much of UK is at risk of establishment?
Non-UK
records
(training)
UK records
(validation)
First UK
record
Phytophthora ramorum 3401 12037 2002
Phytophthora cinnamomi 3921 221 1937
Phytophthora lacustris 962 6 1972
Phytophthora plurivora 259 334 1999
Phytophthora cryptogea 144 302 1915
Phytophthora gonapodyides 375 62 1900
Phytophthora x alni 382 16 1993
Phytophthora cactorum 127 190 1959
Phytophthora cambivora 219 55 1938
• All 9 species for
which we have
sufficient records
to make a niche
model are already
present in the UK
• Very few records
for focal species
not already in the
UK
22. Risk of establishment: global niche models
Models based on environmental factors predicted global
distributions well
Species
Modelaccuracy
23. Risk of establishment: global niche models
Models with
environmental
factors and global
occurrence data
give accurate
predictions of UK
distribution in the
wider
environment
(UK data excluded
from model
training)
25. Risk of establishment: global niche models
Varying role of environmental drivers between species
0%
10%
20%
30%
40%
50%
60%
P. cactorum P. cambivora P. cinnamomi P. cryptogea P.
gonapodyides
P. lacustris P. plurivora P. ramorum P. x alni
Variableimportance
Summer temperature (Bio10) Winter temperature (Bio6) Precipitation seasonality (Bio15)
Moisture index Forest cover Urban cover
Agricultural cover
26. Risk of establishment: global climate limits linked to
traits
y = 0.4889x + 10.18
R² = 0.272
8
10
12
14
16
1 3 5 7 9
Minwintertemperature
atoccurrences(C)
Minimum growing temperature (C)
y = 0.3112x + 16.354
R² = 0.362
20
22
24
26
28
30
24 29 34 39
Maxsummer
temperatureat
occurrences(C)
Max growing temperature (C)
Thermal traits >>>
Climaticlimits>>
27. Risk of establishment: global niche models
• Global niche models: valid methodology for identifying suitable
habitat for Phytophthoras in the UK
• But lots of occurrence data needed to develop them, not available
for many species that are yet to arrive in the UK
• Many Phytophthoras unknown to science when 1st emerge in
invaded range
• Centralised occurrence and trait databases, integrating across
sectors and enhanced sampling in source regions needed to be
able to develop models and understand species behaviour when
invade temperate areas
28. Risk of spread: do traits explain variation in global
spread and host range of Phytophthora species
range = 1 to 132 countries
median = 2.5 countries
range = 0 to 90 host plant families
median = 2
Geographical extent:
Number of countries
“occupied” by species
Host range:
Number of host families
recorded for species
29. Risk of spread: do traits explain variation in global
spread and host range of Phytophthora species
Fitness
Trait A
Trait B
Trait C
Growth rate
Survival
Reproduction
Global impact
SpreadTrait D
Assumed indirect effects
30. Do individual traits explain geographic extent?
• Cold-tolerance
• Root symptoms – soil-
borne transport with
ornamentals, better
buffered from dessication,
temp extremes
• Foliar symptoms –
sporulation and local
spread
31. Do individual traits explain host range?
• Optimum growth rate
• Thicker oospore walls
• Root symptoms – soil-
borne transport with
ornamentals, better
buffered from dessication,
temp extremes
• Foliar symptoms –
sporulation and local
spread
32. Do thermal traits especially cold-tolerance modulate
invasion of Phytophthora into temperate regions?
• Biological basis of cold-tolerance in Phytophthoras?
• Thermal traits and oospore wall width that we have linked to invasion success
also very variable across the phylogeny, suggesting recent adaptation
Min temp for growth
values across family
tree of Phytophthoras
33. Risk of spread and impact : findings and horizon scanning
• Phylogenetic relatedness: a strong a
predictor of impact than biological traits
• Years known to science
• Cold-tolerance and disease symptoms
predict geographic extent
• Oospore wall width, disease symptoms
and optimal growth rates predict host
range
!
• Phylogeny + traits accounted for 50-60% of variation we see in host
range and global extent, potential for horizon scanning
• Traits measured routinely for species descriptions, are we missing
key traits linked to invasion?
34. Conclusions and future work
Invasion success of Phytophthoras modulated by biological traits at
each invasion stage, live plant trade from source areas is key pathway
Transport/Introduction
• Connectivity to source
countries through live
plant trade (which
products/hosts?)
• Some species, that are
cold-adapted, better
able to exploit live
plant trade pathways
Establishment
• UK: cold-adapted
species established
further north in wider
environment
• Predictable from
global niche models
• Global climate limits
linked to traits
Global spread and impact
• Cold-tolerant species,
cause root and foliar
symtoms - wider global
extent
• Sp. with higher growth
rates, more resistant
oospores, cause root
and foliar symtoms -
wider host range
• Closely-related species
are similar in impact
• horizon-scanning
35. Global Phytophthora databases
• Data pooled across sectors is valuable for understanding and
predicting pathogen behaviour - traits and environment
• Available to end-users and researchers, updated in future
36. Model outputs into policy & practice:
How are Phytophthoras introduced into
new areas, from which key source areas,
by which trade pathways?
Which Phytophthoras arrive? Does this
depend on traits?
Phytophthora
importation
tool
38. Model outputs into policy & practice:
Hosts and sectors impacted
Demonstrate host-phytophthora relationship part of tool
39. Model outputs into policy & practice: your feedback
Would you use any of these tools to inform your practices?
If yes, how could your practices change as a result of using
these tools?
What modifications or additional information would you like to
see in the tools or models to improve their usefulness?
How would you like to access them? Website, phone app
40. Louise and I would like to thank:
Daniel Chapman, University of Stirling
> 100 global contributors to the trait, host and occurrence
data-bases
Ana Perez-Sierra, Treena Burgess
Beatrice Henricot, Anna Harris, Giles Hardy, Peter Scott,
Nari Williams, David Cooke, Sarah Green, Paul Sharp
Jane Barbrook, Alexandra Schlenzig, Royal Horticultural
Society for sharing UK data
Mike Dunn and Mariella Marzano
Funders of Phytothreats
Acknowledgements
Editor's Notes
We’ve targeted These organisations are a selection of those we have contacted representing the different
Beth works a lot with zoonotic diseases where there is a movement towards a One Health approach. This is where emerging diseases are best understood when human health, wildlife health and livestock health are dealt with together. And there is a similar argument for mapping the global distribution of Phytophthora species in all of these sectors, so we can properly understand the factors driving their distributions. Because we know that horticultural hosts are a potential reservoir for Phytophthora and there’s some agricultural pests that are shared with forestry species too. So I think it’s important that we try to get as much data as possible from all of these sectors, even if it’s mainly forest and forestry disease that is our focus.
We’ve targeted These organisations are a selection of those we have contacted representing the different
Beth works a lot with zoonotic diseases where there is a movement towards a One Health approach. This is where emerging diseases are best understood when human health, wildlife health and livestock health are dealt with together. And there is a similar argument for mapping the global distribution of Phytophthora species in all of these sectors, so we can properly understand the factors driving their distributions. Because we know that horticultural hosts are a potential reservoir for Phytophthora and there’s some agricultural pests that are shared with forestry species too. So I think it’s important that we try to get as much data as possible from all of these sectors, even if it’s mainly forest and forestry disease that is our focus.
Reproductive strategies can be more flexible than homo-heterothallism distinction would suggest, so Treena suggested modifying to just consider oospores versus no oospores, minimum temperature included because most labile.
Look up biological significance of proliferating sporangia and check terminology
We hope that the trait-based framework can be used as a predictive tool, when a pathogen is first described, when only a few morphological or biological traits are known or measured. The root and foliar disease traits may only be fully described after a species has been known for quite some time, so we compared the performance of models with and without these impact traits. Because information on other trait or impact values like host range may also take time to accumulate, we included time known to science as a potential explanatory variable in our analysis. This should eventually let us predict for recently described species, the eventual host range for example.
This is a quick overview of the prevalence of different Phytophthora species in nurseries, gardens and forest sites. It’s difficult to unpick these patterns from the sampling effects as we clearly have lots more records from gardens because of the huge amount of data in the RHS dataset. But it does seem that the species in forests are subset of those found in nurseries and gardens. So it be that we
Next is the results from the global niche models. Dan Chapman has been working on these very recently and he’s made loads of progress really quickly. So I apologise if I don’t do these results justice as I haven’t really digested them yet, but the results look really exciting.
Methods:
Available species data
How models account unsuitable habitat, dispersal distances and recording effort
Environmental drivers considered (expert-elicitation)
This summarises the performance of the niche models for all nine species. It’s based on a 80:20 split of the non-UK data.
The model performance is based on the area under the receiver operating characteristic, which is a measure of how well the model classifies the presences and absences at different thresholds of probability of occurrence predicted by the model. A model that predicts perfectly would get a value of one, which is represented by the blue dotted line at the top. So it seems like within the non-UK data the model is performing pretty well at discriminating presences and absences.
Looking at the traits which are significantly linked to host range in blue here, Phytophthora species are recorded as using more host families if they have been known to science for longer, if they cause foliar or root disease, and again if they have a low minimum temperature for growth. In addition, closely related species are likely to be similar in the number of host species they are found on.
Looking at the traits which are significantly linked to geographical extent in blue here, Phytophthora species are recorded as using more host families if they have been known to science for longer, if they cause foliar or root disease, and again if they have a low minimum temperature for growth. In addition, closely related species are likely to be similar in the number of host species they are found on.
cold-tolerance may be labile than heat-tolerance? There’s evidence for this among other taxa (e.g. Araujo et al. 2013). Implications for colonising novel regions? Are higher latitudes more prone to new arrivals? Are recent emerging infectious diseases at higher latitudes associated with adaptation to cold-tolerance? Do we know (have a feeling about) the origin of any of these (e.g. P. ramorum, cinnamomi, austrocedri) that might corroborate this?
Treena’s pictures
Pr versus Pk
Look at with spread models in Europe and Global
Biological basis of thermal tolerance, how thermal traits co-vary with others?
Root disease more likely to kill the plant than foliar disease – related to disease severity?
soil / root disease pathogens more buffered from external environmental fluctuations (protected from dessication, temperature extremes) compared to foliar pathogens?
Soil-borne /root disease pathogens more likely to arrive with live plants?
This summarises the performance of the niche models for all nine species. It’s based on a 80:20 split of the non-UK data.
The model performance is based on the area under the receiver operating characteristic, which is a measure of how well the model classifies the presences and absences at different thresholds of probability of occurrence predicted by the model. A model that predicts perfectly would get a value of one, which is represented by the blue dotted line at the top. So it seems like within the non-UK data the model is performing pretty well at discriminating presences and absences.