Preparing to search
The aim of this meeting was to bring the entire project team and members of the Expert Advisory Panel together to share and discuss research progress since the last all-project team meeting on April 23rd 2018, and to outline and receive feedback on research plans including proposed funding for continuing strands of the project beyond this financial year.
WP1 Phytophthora distribution, diversity and management in UK nursery systems – David Cooke and Leighton Prichard (JHI)
David Cooke started off his presentation by introducing a new member of the WP team; JHI bioinformaticist Peter Cock replaces Pete Thorpe who has taken up a position at St Andrews University.
To date the team have completed 59 sampling missions across the 15 partner nurseries, collecting 4016 samples of which 2165 have been PCR tested for Phytophthora. The lab team have determined that processing the buffer for DNA, rather than the filter on which the DNA was originally collected, gives lots of DNA yield, so they are no longer processing filters. The team are still working their way through the fine-scale nursery root samples.
The broad-scale sampling has resulted in root samples from 101 nurseries; 36 in Scotland and 65 in England/Wales. With 5-10 root samples per nursery a total of 653 samples have been returned. Top hosts returned from the broad scale sampling are: Rhododendron, Viburnum, Pieris, Juniperus, Chamaecyparis, Hebe, Fagus, Olea, Taxus, Prunus. It will be interesting to see which hosts turn out to be Phytophthora positive. These root extractions are under way. For the OPAL wider environment sampling of water courses (now finished) a total of 26 samples have been received.
Much effort has gone into root extraction optimisation. JHI tried automation with a Qiagen kit but the result was not good. They are automating with PowerPlant Pro kit, which is currently being used for the fine-scale root samples.
PCR test results to date show that ~ 50% of samples are positive overall. The frequency of positives varies from nursery to nursery (ranges from 30-70%). These results will be presented by management practice, not by nursery, in public fora. Positives for roots are slightly higher than 50%. There are 1600 samples still to process.
Since April this year more samples have been sequenced using Ilumina on plates containing nursery samples as well as samples from other Phytophthora metabarcoding projects – which all help to validate the results. David gave an overview of the Illumina rationale which Leighton will cover in his talk later, reminding everyone that sequencing outputs are 250 bp reads, 15M barcode reads, 156K reads per sample. Synthetic controls are included on each plate to give an estimate of technical error rate and sensitivity range. Also included on one of the plates were 45 samples from cultures of known Phytophthora spp. David went on to present some of the nursery sample findings from two of the nurseries:
Plus some (closely related) downy mildew species
Samples from another nursery yielded DNA reads matching:
Data on Phytophthora species detected in nurseries will feed into the community modelling (linking with WP3) which will assess factors such as nursery size, management practices, suppliers and geographical location on Phytophthora species assemblages detected. Nursery Phytophthora data will also be linked to Phytophthora data from natural ecosystems in concurrent projects (i.e. Ponte) and community eDNA /metabarcoding.
Next steps are to complete sample processing and barcoding, re-run samples where contamination is suspected, validate the data with the new bioinformatics classifier (involving Leighton and Peter Cock), and feedback data to nursery managers and the WP3 modelling team.
During the discussion a point was raised that some PCR positives initially reported back to nurseries as Phytophthora positive have subsequently been identified by sequencing as downy mildew species i.e. other oomycetes. How we are reporting back to nurseries in light of this?
It was agreed that downy mildews are also important pathogens and should be reported to nursery managers, but not included in the data as Phytophthora +ves.
Another comment related to the urgent need to report back to nurseries soon as they are overdue sample results, also, that it would be interesting to see if there are any changed behaviours according to results. David agreed, and said we would be able to relate photos of incidence/symptoms to results in the feedback.
It was asked whether peat or peat alternatives had been tested? The answer was no, potting mix was not tested as part of this project.
A final question: can the species data be related back to circumstances/ best practices? The answer was yes, this is the plan for the analyses.
Leighton Pritchard (JHI)
Leighton’s presentation focused on the question of what was really being measured by Illumina sequencing. He reminded us that there are real stakes from our results (i.e. the consequences of a regulated pathogen turning up in a nursery might mean closure) so we need to be held to a higher standard of accuracy than other types of studies.
ITS1 marker sequences are used in the metabarcoding analyses. We assume one species means one ITS1 sequence, but this is not the case!
Leighton ran through the metabarcoding method and described three ‘pinch points’. These are:
Synthetic control sequences are used to help set an error rate for each sequenced plate and to check for levels of cross contamination in test samples. Synthetic control mixes are comprised of 4 unique sequences with a base composition similar to Phytophthora ITS1. In the tests that Leighton conducted (on one of the plates) the four synthetic sequences were used at six different combinations and at three different dilutions i.e. 18 synthetic control samples in total on the plate.
Analyses of the Illumina output showed that some synthetic control samples contain 1000s of sequence variants, differing from the actual control sequences by up to 10-15 bp. How much is real variation? Leighton presented a plot showing the variants and how different they are from the original sequence that was put in. Leighton also spoke about the dilution effect: at the most dilute, worryingly, an artefact of 1 snp difference is seen. Some sequences can’t be matched to the input sequences and are likely to be contamination. These occur at a threshold of 100. So, setting a threshold above 100 gets rid of a lot of variants. If a DADA2 clean-up is performed, the threshold is reduced to 20 variants. Also seen are PCR artefacts in low abundance (i.e. when PCR amplification introduces variation) and what appear to be contaminating sequences from environmental samples. So what is in the contamination? Many look like ITS1 sequences; many don’t. Most cross-contamination comes from soil bacteria, some are oomycete sequences. The synthetic sequences pop up too at low abundance in some of the environmental samples. These are easily recognised as cross-contamination and their abundance (in terms of read counts) in samples across the plate can be used to set thresholds.
Leighton then went on to explore the ‘one species, one ITS1’ concept. Many species have more than one ITS1 sequence since the ITS region can occur in a Phytophthora genome in anything from 40-170 copies. If a single isolate is Illumina sequenced then one major cluster should result. One of the plates contained ~40 single isolate samples to test for ITS1 sequence variation across a range of species. However, from a single isolate up to 2000 sequences resulted! Thus a single isolate does not give 1 single sequence. The resulting clustering of sequences by Swarm produced up to 150 OTUs (operational taxonomic units) (mostly 40-60) from a single isolate. However, many of these sequences are variants occurring at low abundance. Removing all amplicons occurring at abundance of less than 20 before clustering reduces the number of OTUs per isolate to around 6.
The next question is ‘how are the OTUs present within each isolate ‘individualising’? By looking at the Jaccard distance between OTU members among isolates it can be seen that several OTUs are unique to individual Phytophthora species and some are common to several sequenced species. If one species has a specific ‘classifier’ (i.e. unique OTU) as well we can ignore the ambiguous ones.
Leighton asked ‘what is a classifier?’ Associating input sequences with species classes, this is the classifier. So if you change anything in the method, the classifier changes. Leighton and bioinformaticist Peter Cock are currently developing the automation of the ‘classifier’ process. This includes building the reference database framework based on trusted sources. The next step is to create a training dataset. This will be combinations of reads from single isolate controls, individualising sets for a species, applying and evaluating the current pipeline and applying validated classifiers to old and new samples.
A question was asked about the timeframe by which we can have confidence in the latest plates to present the results back to nursery managers. Leighton said he couldn’t give a definite date, but hopefully fast! He thought early next year was achievable.
Another question related to whether there were data on isolations to go with these sequence outputs? It was confirmed that there were P. lateralis/P. cambivora isolations with DNA samples from the same material progressing down the pipeline. However isolations were never intended to be done as part of this project. It was mentioned that a Scottish Government-funded project is generating companion data from live (baited) organisms as well as metabarcode data from the same soil and water samples. We’re assuming that what is in a sample represents a threat but this may not be so. Detectable live presence may be a better marker.
Another comment related to the sequencing process: the synthetic control in most abundance gave the same reads when replicated and those sequences that were spiked appeared in the correct order. The lowest dilution of synthetic sequences was not found or was equal to background noise, which is good. It shows this as a robust and accurate tool. In this respect, testing with control sequences, we seem to be at the forefront of metabarcoding.
It was pointed out that the nested PCR itself could potentially introduce error and there was general concurrence on this point given that the technique is so sensitive. It is recommended when starting up to use blanks and do a control plate before anything else so that an estimate of cross-contamination can be made. However, overall, when considering what species have been found in nursery samples to date, the results make sense. Pathogens appear on expected hosts (i.e. P. austrocedri on juniper, P. lateralis on Chamaecyparis, P. pseudostugae on Douglas fir, P. occultans on Buxus). If a result looks erroneous, this has usually been present at very low read number and is most likely to be cross-contamination. Hence the need to set a read threshold below which results are not considered as true. On the question of doing some qPCR validation of results, the answer was yes, this had been done, but as part of other metabarcoding projects.
WP2 Feasibility analyses and development of ‘best practice’ criteria – Mariella Marzano (FR), Mike Dunn (FR), Gregory Valatin (FR), Glyn Jones (Fera) and Colin Price (external consultant)
Mariella opened her presentation with a brief reminder of the objectives of WP2. She then went on to update the team as to where they are with the different sectors:
Nursery interviews – the team have conducted 19 nursery interviews so far and are aiming to have interviews completed by Jan/Feb 2019. One of the problems has been that there are so many questionnaires in circulation because of the competing accreditation schemes that nurseries have been deluged by questions and become fatigued! The team have, however, managed to interview several on-line retailers and garden centres, where previously they were short of inputs from these sectors. A survey company has been recruited to take up the task from here.
Landscapers – Mariella explained that after engaging with landscapers e.g. LI (Landscape Institute) and BALI (British Association of Landscape Industries) she was made aware that her team’s questions were too generic and not specific enough. This will need further discussion in the focus groups. ‘Landscapers’ are more accurately landscape architects, contractors or garden designers, each with different roles and implications for biosecurity. Furthermore, individual contracts determine how much impact landscapers have in choosing plants: they may be told what to plant; may be able to offer choices or alternatives; and depending on the client base they may or may not have biosecurity awareness. They may have access to a reference guide (for example the LI is releasing a ‘biosecurity toolkit’ document for their members). A good question to investigate is at what stage in the process is biosecurity important, if at any?
Glyn Jones and Barbara Agstner of Fera have done a lot of groundwork on cost-sharing. This will be presented by Glyn later.
Retailers/Garden Centres – Mariella attended a recent workshop that involved large retailers. Discussions indicated the feeling that customers trust that this sector is doing things correctly and expect quality, and retailers are wary of negative messaging. However, customers are now starting to ask more questions e.g. on plant origins. Through interviews one issue raised was that customers can take their diseased plants to the garden centre to ask experts for advice on ailing plants!, thus, potentially disseminating pathogens into the garden centres. There seems to be an opportunity here for a plant health message to the public.
Mariella then posed a question to the floor – the WP2 team had proposed to do a series of focus groups in the later stages of the project, so what should the focus groups be about? This will be discussed later during the WP5 presentation.
Mike Dunn – Mike has been exploring biosecurity issues with public parks/gardens. Both sectors have an interest in plant health standards, which is driven by the obligation they feel to meet visitor requirements, not driven by pest/disease concerns. Generally, visitors want and expect to see exotic species. The National Trust has bronze/ silver /gold plant health standard checklists of what factors to consider when procuring. Mike and team have also developed a Local Authority question framework and plan to interview 15-20 Local Authorities.
The consumer survey, including gathering of more economic data, is now going to be the responsibility of a survey company and telephone surveys will be conducted on 50 each of nurseries, garden centres and landscapers, to be completed by February 2019. Questions will elucidate location within the supply chain and key biosecurity factors.
Mariella asked for input on how to integrate all the data from the surveys of the three main sectors and how to disseminate advice. It was agreed that packaging up information and advice from the project and making project data accessible to stakeholders would be key.
A comment was made that the team’s choice of target groups was correct, particularly the landscapers. The ‘landscaper’ sector could have some biosecurity weaknesses, because following the planning and approval stages there are apparently few subsequent checks and balances to ensure that plans are followed.
Glyn Jones presented on the Defra Future-Proofing Plant Health (FPPH) work of relevance to the Phyto-threats project, outlining three projects as follows;
Early warning system/pathways analysis – Glyn presented a slide outlining the structure of the project, incorporating figures on global imports /exports, global tariffs, gathered from governmental data and the World Bank, used to look for trade anomalies using algorithms/machine-learning. This highlighted non-usual datasets.
Industry data – this project is revealing the difficulties of actually obtaining species import data. For example one company provided 240 combinations of species and specifications. The work has highlighted the very complex network of business connections that changes from quarter to quarter across the year. They are looking at certain species and asking; how much is brought in and what is the UK demand? For example, lavender, for which the demand is high, so could there be a good case to increase production internally?
Costs and responsibility – much of the work has been done by Barbara Agstner (Fera) as part of her PhD. Barbara has been on a road trip around Britain visiting various nurseries and getting estimates as to the costs of implementing various biosecurity measures. This work has again highlighted the many gaps where costs are unknown. One company was able to provide costing for a range of activities – this is being shared with Gregory.
In the second part of his presentation, Glyn presented on the emergence of various industry accreditation initiatives over time, i.e. BOPP, Grown in Britain, UK Sourced and Grown, the Plant Health Alliance. Currently the HTA, Grown in Britain and others have been working on a Plant Health Management standard and assurance scheme which is ready for launch in early 2019.
Any scheme will be about applying a plant health management standard(s) that will be owned by the governing body. The standard would then be adopted by the various assurance schemes. Individuals apply to the relevant certification body to join their scheme – thus currently there are several competing schemes all of which have to adhere to the agreed standard.
A scheme will have self-assessment questions, a full audit checklist, will require an audit. Everyone would sign up to the scheme and all would have to meet the criteria. However, certification schemes could compete – they would just apply standards, and would set their charge.
The question was asked as to what best practice recommendations are needed to feed in to the standard from Phyto-threats? i.e. step-wise introductions of recommendations into a common industry standard since certain infrastructural changes need to be applied, such as water storage and treatment, raised benches, drainage, etc to make a scheme work at preventing disease. The idea of competing schemes did not seem very efficient either and stakeholders had preferred a single, overarching UK scheme at last year’s stakeholder workshop.
Gregory Valatin spoke about the cost-benefit analysis of introducing best practice in nurseries from a nursery perspective. Following an initial low response from nursery managers to the economics questions included in the WP2 survey questionnaire, Gregory joined Barbara Agstner on some nursery visits as part of the costs and responsibility sharing Fera project. Over the summer 2018 a FR intern also helped gather economic data.
Gregory listed 12 ‘best practices’ which formed the basis of the questions. Nursery managers were asked for estimates of the cost of implementing these practices. The best practices included water testing, water treatment, storage of water in fully enclosed tanks, clean/covered storage of growing media, use of raised benches, disinfestation stations for tools/containers etc, boot washing station, vehicle washing station, quarantine holding area for imported plants, installation of drainage systems, composting/incineration system for disposal of waste plants and buying only from trusted/accredited suppliers.
In addition to questions about the costs of implementing the best practices, growers were also asked for an estimate of the cost to their nursery were it to be affected by a future Phytophthora outbreak. The cost of implementing the best practices can then be compared with the expected avoided cost of the outbreaks expected to be prevented by implementing the best practice measures. The benefit of preventing outbreaks partly depends on the frequency of the outbreaks expected to be avoided through introducing the best practice measures. It is anticipated that other members of the Phyto-threats team – especially those working on WP1, may be able to advise on the frequency of outbreaks prevented (with this information also needed for wider exploratory cost-benefit analysis from a societal perspective).
Gregory presented preliminary conclusions for a number of scenarios based on economic data gathered so far (for which there are many gaps). These included where a nursery implements all 12 best practices (i.e. the baseline is that none of the measures have been implemented to date), as well as a scenario where a nursery implements just 8 of the best practices (assuming a ‘common practice’ baseline of 4 measures implemented to date). In each case it was assumed that at most one Phytophthora outbreak per year would be prevented. For the scenario of implementing all 12 measures, the initial survey responses indicated that the mean cost to nursery managers of implementing best practice, whether start up or where just the annual cost is considered, is much higher than the expected benefit of avoiding the costs of dealing with an actual outbreak. For the scenario of implementing the 8 best practice measures, the initial estimates similarly indicated that the mean set-up cost of implementing the best practice is much higher than the expected benefit of avoiding the costs of dealing with an actual outbreak, with the annual costs being of a similar magnitude to the expected benefit of avoiding an outbreak. For a ‘typical’ (1 ha) nursery, estimates from a sector expert also indicate that set-up costs for the best practice measures exceed the benefit of avoiding an outbreak (whether introducing all 12 best practices or the 8 best practices that initial responses to the survey indicate are not already common practice). Whether costs exceed the benefits to the nursery depends not only on the frequency of outbreaks that the nursery expects to avoid (which the responses to date suggest are currently far fewer than one per year – but potentially might increase above one per year if detection rates increase due to technological improvements), but also on any sales price premium nurseries are able to obtain once they introduce the best practices. Based on the preliminary data, Gregory questioned the level of take-up of a voluntary accreditation scheme if the perceived costs to the nurseries of introducing the best practices outweigh the perceived benefits.
Some of the qualitative responses from nursery managers support the conclusion that implementing best practice to reduce the risks of Phytophthora outbreaks is not seen to be cost-effective from a nursery perspective. However, as one of the nurseries contacted had noted, the cost of an outbreak will depend on which Phytophthora species is causing the damage. Gregory also asked the question; what happens if we factor in the benefits of avoiding other pathogens such as Xylella? Although implementing nursery best practices to reduce the risk of spread of phytophthoras may not seem a priority for many nurseries, current biosecurity measures across the plant trade sector are viewed as inadequate by some nurseries. Indeed, some nurseries advocate urgent action by the government in shaping which practices are permitted – especially in the context of risks of introduction of Xylella, with mandatory restrictions needed that apply more widely to the plant trade than just the nursery sector. Information on costs is still being incorporated from personal interviews and telephone interviews carried out this summer. More economic data will also be gathered by the survey company over the next few months.
It was asked whether outbreak costs were based on cost of loss of stock alone as it was pointed out that outbreak costs should include Plant Health inspector time, litigation costs, loss to suppliers etc. Gregory noted that the survey asked both about the anticipated loss of nursery stock, as well as other costs to the nursery, but that nurseries are frequently very uncertain about the level of costs likely to be incurred – as illustrated by the wide range of estimates and proportion of respondents unable to provide estimates for the costs of implementing specific measures. Phytophthora risks are often perceived by nurseries as low compared to other pathogens, so there appears relatively limited interest currently in in implementing measures specifically aimed at reducing risks of Phytophthora outbreaks.
The comment was made that since improved biosecurity will be good for preventing all diseases, we need to emphasise this, in addition to talking about Phytophthora. Xylella should be included since quarantine holding areas and careful plant sourcing reduces the risks of a Xylella outbreak. It was reiterated by others present that core best practices work for all, and that best practice reduced the weeding/husbandry bill. Gregory noted that a question had recently been added to the questionnaire to enquire about the costs anticipated were there to be a future Xylella outbreak at the nursery, although initial responses indicated that it was also proving very difficult for nurseries to answer this.
Colin Price presented ‘some economic costs of Phytophthora: nightmare scenarios’ using the CARBBROD model– a model being used as the core element for the exploratory cost–benefit analysis of introducing best practices from a wider societal perspective. Thinking in a forestry context, disease outbreaks matter economically because of curtailing rotations and timber yields, wasting forest expenditures, taking land out of production, and by reducing carbon sequestration and storage. From a societal perspective, carbon impacts are key, owing to the high social values placed on carbon in relation to climate change mitigation. One of the scenarios presented using the model was of a new and unknown Phytophthora species entering the UK (e.g. via the nursery trade) and spreading to infect Sitka spruce plantations (based on a crop of yield class 14, age 30) along similar lines to the recent impact of P. ramorum on larch, with bare land following (if no other species replaced the Sitka on upland, acid sites); and another if Sitka spruce is replaced by noble fir. Owing to the trajectory of the social value of carbon used by the UK government for policy appraisal, infection of the Sitka followed by replacement with noble fir was found to give a better return to society than not having an outbreak! This was because of the increase in the discounted social value of carbon over time (with a lower price of carbon applying to the emissions associated with losses of the Sitka than to subsequent carbon sequestration by the noble fir). Everything depends on prices and timing! In prior discussions, Gregory had suggested potential for adopting other approaches to valuing carbon (e.g. assuming a fixed carbon value in real terms) to avoid the perverse conclusion that introducing new pests and diseases to the UK could be beneficial to society: Colin noted that implementing other approaches to valuing carbon was easy to do within the model, yielding the expected conclusion, that outbreaks modelled as above did indeed have adverse economic consequences. In worst cases, the cost through infection could be many tens of thousands of pounds per hectare.
Colin also looked at a scenario involving the speculative consequences of infection of oak stands with a new Phytophthora to the UK, assuming 30% mortality, 30% unaffected and 40% having reduced growth increment. In this case the stand increment recovers and regeneration comes from surviving trees. Forest managers can thin and choose subsequently to replant either with oak or something else (e.g. sycamore). In contrast to the scenarios for Sitka, a key difference here is that continuing to grow the same species on the same site is an option.
The CARBBROD model can also be applied to a scenario involving a new disease coming in and infecting urban trees, for example a broad host-range pathogen that can take out 10% of all urban trees. The resource baseline proposed for this scenario was the urban forest from the London i-Tree Eco project survey. Colin compared the costs of trees dying due to a new disease versus those of dying due to age. Costs in terms of the associated reduction in ecosystem services were evaluated, including loss of pollution abatement, loss of aesthetic services, and loss of carbon sequestration, as well as replacement costs’ being brought forwards. The costs need to be evaluated by species, age and life expectancy.
Colin closed his session by asking the floor for likely scenarios for a new Phytophthora coming to the UK and infecting Sitka, oak etc. For example how much of the crop is likely to be affected, what is it likely to be replaced with? Discussions with Colin over lunch and tea did not prove as helpful as hoped in trying to answer some of these questions. Wider enquiries will be instituted.
Attention could be drawn to some reminders on CARBBROD, compiled by Colin and presented by Mariella to the mid-year project meeting. CARBBROD does not deal specifically with the economics of nursery practices, but with the effects of these practices on forestry plantations and woodland health based upon the likelihood of the outbreak and spread of a new tree pathogen (e.g. entering the country via the plant trade). It evaluates the consequences for timber production and carbon fixing of disease arising under different scenarios of tree age, consequences and management responses. In this project the model will be applied to important threatened species and genera (e.g. Sitka, oak, urban trees). It is flexible to allow adoption of different forms of carbon pricing, discount schedules and interactions.
In terms of future developments of CARBBROD, as the carbon effects are dominant, incorporating the effects of the new disease on litter and soil carbon (if these are known) would be useful, as well as adapting urban tree carbon for effects of disease, including landscape scale effects of disease if feasible and incorporating a simple application to scenarios of disease spread (as has been done previously for Dothistroma).
WP3 Global Phytophthora risks to the UK – Beth Purse (CEH), Dan Chapman (CEH) and Mike Dunn (FR)
Pathways of risk of introductions
Traits modulating introduction risk
Social factors and environmental/ecological traits influencing spread once here
Environmental niche models to map areas at most risk
Dan Chapman presented on risk of introduction of pathogens which they are addressing by modelling the risks based on transport networks and pathogen traits. Louise Barwell has accumulated a database of country level occurrences which includes 17,371 country level records with dates and 1,417 species x country combinations (the source of this information is the recipient country/ year of first arrival/invasion status).
A preliminary analysis of this database asked the question: ‘Can live plant trade network explain arrivals?’ There are data for total import volumes, network connectivity (imports of focal species from source countries), arrivals-trade-species-country. Arrivals detected post 2000 were from 94 countries and 56 Phytophthora species with ≥ 1 documented arrival. The result was that connectivity (total live plant imports from countries in which Phytophthora species occur) explains about 21% of the variation in new arrivals better than total imports (i.e. from anywhere).
Moving on to risk of establishment and spread, Beth ran through the original aim of the global niche models: to predict the area of extent of impact on UK tree species of global Phytophthora species that have not yet arrived in the UK (30) and 24 species that have arrived. Of the global Phytophthora occurrence data gathered, regions climatically similar to the UK give 11,497 records comprising 82 Phytophthora species in 38 countries. Data are still to come from countries such as New Zealand, Australia, South Africa and Canada.
Looking at the habitats in which Phytophthora species are recorded, most records are drawn from closed forest types, also urban areas and cropland. Many of the species for which sufficient information is available to make niche models on risk of establishment and spread are already in the UK. Focal species not yet present in the UK have very few occurrence records worldwide. So the aims of the niche models need to be revised. For those species already present in the UK, niche models will look at their potential distributions here. For high impact invaders such as P. cinnamomi and P. ramorum new global niche models will be developed with a range of environmental factors.
Records of pathogen interceptions/escapes in the UK have been gathered from key sources (THDAS, RHS, SASA, eDOMERO). Some data e.g. THDAS have related data on interceptions to subsequent spread in the different settings, linking back to traits and trade/recreation proxies. Geographical occurrence will be linked to environmental characteristics in order to predict potential distribution.
Beth then updated everyone on the Phytophthora traits database which was merged last year with a database held by Scion (New Zealand) and which will be maintained there. Beth finished up by outlining future work which will be to submit papers on the trait database and phylogenetic analyses and traits and global impact (by January 2019), trade models, finalise global niche models, policy briefs and co-development of final model outputs with policy makers and practitioners. An example of the latter might be the production of interactive source maps from trade and tourism models.
A comment was made that host ranges did not seem to be included in the traits database, possibly because information only comes in when disease outbreaks occur. Yet host range data are very important. Generally, when a new species is found it is tested for pathogenicity on hosts closely related to the affected host, or on hosts that closely related Phytophthora species infect. It was agreed however that phytophthoras can surprise, for example host range tests were performed on trees expected to be P. ramorum hosts based on analogy with the US, but no one expected larch! It was asked if closely related Phytophthora species have similar impact? If you look at host range at the Phytophthora species level, rather than related group level, some species have a wide host range, others narrow. It was cautioned however that wide host range doesn’t necessarily mean they are more devastating on a host.
Mike Dunn presented on his work looking at tourism and recreation as a pathway for pathogens, carried out through literature search, survey of Plant Health researchers around the world, and visitor data to UK parks and gardens. Generally, it is considered true that recreation and tourism have acted as a pathway for spread of plant diseases previously, but that this pathway is perceived to be of lower risk in terms of spread of pathogens than other pathways such as traded plants and plant material.
In terms of visitor data, the 2011 Visit Britain survey on passengers visiting Britain has been useful in providing number of visitors, origin and timing of visits and, based on a questionnaire, the estimated number that are planning to visit a public park/garden while in the UK. Mike has also approached 23 parks and gardens asking for data on visitors, with a mean of 471,000 visitors pa. Very few of these sites collect data on visitor origin. Moving on to discuss domestic spread within the UK, a survey monitoring engagement with the natural environment conducted in 2017-2018 found that 62% of adults living in England reported making visits to the natural environment at least once a week.
The next steps for this piece of work are to write up the data and think about whether incoming tourists are a viable means of spread of pathogens.
It was suggested that people travelling between (for example) National Trust properties in a single day could potentially be transferring pathogens from site to site. Bus tours, train tours etc go from site to site and it might be interesting to look at the tours and what their schedules are. Another suggestion was to check international plant propagators conference proceedings as examples of people going overseas and bringing back plant samples independently.
WP4 Predicting risk via analysis of Phytophthora genome evolution – Paul Sharp and Ewan Mollison (University of Edinburgh)
Paul Sharp explained that this work package focuses on understanding what can drive the evolution of a pathogen and allow pathogens to adapt to evolving host defences, expand host range and increase virulence. Comparative genomics may enable us to identify the genetic basis for some of the Phytophthora traits, for example evolving to infect woody hosts. Previously his group successfully completed similar work with 64 strains (genomes) of P. syringae: 38 from woody hosts and the rest from non-woody hosts. They were able to associate specific genes with woody hosts and they will try to accomplish this with Phytophthora.
The aims of this work package are to compare genes from available sequenced Phytophthora genomes, identify a core set of Phytophthora genes common to all species, identify species-specific genes or variation, sequence the genomes of three less damaging species which are closely related to highly damaging species (the topic of this talk) and study target genes/gene families known to be important for virulence.
Currently, they have Phytophthora genome information available from various sources, covering 27 species and 10 clades, but most genomes are from species in clades 7 and 8. Paul showed a phylogeny of all Phytophthora species with available genome data.
The project has recently sequenced the genomes of three less damaging species. These are;
Carolyn Riddell (FR) successfully extracted high quality genomic DNA from all three species for PacBio long-read sequencing and Ewan has been assembling the genomes.
Ewan Mollison ran through the results from the initial genome assemblies. The reason for choosing PacBio over Illumina is because it gives much longer read lengths (10s of Kbps), giving better resolution of repetitive regions and greater overall contiguity. Error is random rather than systematic, so if there is very high read coverage across the genome the errors can be corrected rather than amplifying bias.
Ewan ran through an earlier assembly of the P. austrocedri genome. This species was sequenced using a ‘hybrid’ method combining both PacBio and Illumina reads. The hybrid assembly was hampered by not having sufficient read depth of either sequence type for optimal assembly. Pete Thorpe (formerly of JHI) re-did the assembly based on the PacBio reads only, using the Illumina reads only for error correction. The outcome was a much improved assembly (862 scaffolds compared with 43,700 with the hybrid assembly). Therefore a PacBio only assembly was pursued with the three Phytophthora species targeted here using two SMRT cells to achieve plenty of reads across the genomes.
The data for the three new genomes (P. europaea, P. foliorum, P. obscura) were presented. A good overall read length was achieved for all three species across both SMRT cells and Ewan explained how he made his scaffold assemblies and gene model predictions. In terms of genome size the estimates for each species are; P. europaea 95Mbp (has more repetitive genome content); P. foliorum 70Mbp; P. obscura 63Mbp; therefore it is safe to use 100Mbp as an estimate of genome size for all three species. The number of contigs for the three species is low (103 to 127) indicating a very high degree of contiguity in all three assemblies.
Scaffolding links contigs together with gaps of known length padded out with ‘N’ characters. The number of scaffolds across the three genomes is also low, ranging from 67-77, again indicating a high quality of assembly. All three species sequenced here have a very low number of scaffolds compared with other sequenced Phytophthora species (except for P. sojae which has a very complete assembly), for example most species have over 2000 scaffolds, with P. cambivora having 120,000 scaffolds indicating a highly fragmented assembly.
The three assemblies reported here have a high completeness of assembly (98%), a low level of genome duplication (~1%) and good resolution of haplotypes. The likelihood of polyploidy is low. Looking across the other genomes the trend is for larger genomes to have larger repeat content. Looking at predicted proteins the trend is less clear. The number of predicted genes is generally similar for many of the Phytophthora genomes. P. cambivora came out as having a particularly high number of predicted proteins but rather than being extra genes this is likely to be gene fragments incorrectly identified as separate genes/proteins.
Ewan showed an example of xylanases as a sample gene family which may be of interest in this study. Xylanases are an important class of plant-cell-wall-degrading enzymes.
There are four major xylanases which have been identified in Phytophthora; xyn1, xyn2, xyn3 and xyn4. Sequences from all xyn genes found in 30 Phytophthora genomes were aligned and phylogenetic trees constructed for each gene. Clades 9/10 only have 2 xyn genes, clade 8 species have 3 xyn genes and clades 1-7 have all 4 xyn genes. xyn sequence differences occur in some species which results in their falling outside the expected clade groupings for that particular gene.
A question was asked about the use of the tool BUSCO for assessing genome assembly. Ewan explained that BUSCO tends to evaluate core genes, for example those in DNA replication and protein metabolism that are more ubiquitous. Checking that they are there would certainly provide a measure of confidence.
A non-biologist within the group asked for a digest on progress for a non-biologist! Much of the language used in this presentation was incomprehensible to those unfamiliar with the field. It was explained by Paul that, put simply, the three sequenced genomes have good sequences so they can now look for genes to see if there’s an association between having a gene and having a certain trait. Also, for genes with target activity e.g. genes known to be involved in pathogenesis like xyn genes, they are now in a position to compare amongst the pathogens and non-pathogens.
WP5 Synthesis and integration – Sarah Green (FR)
Sarah Green presented an overview of coordination/communication events since the last meeting. This included two board meetings (June and August), the minutes of which were posted on Huddle. The April team meeting report and research updates for 2017/18 were also posted on the project website.
The date and location of the next all project team meeting were discussed. It was agreed that a meeting would be held in York in October 2019, together with a stakeholder workshop. It was pointed out that this should not clash with the Phytophthora IUFRO meeting in Sardina over 17-25th October.
Sarah, Tim Pettitt and Jane Barbrook also represented the Phyto-threats project at the HTA National Plant Show in June 2018 following the offer of a stand for free! Sarah took the WP2 consumer survey leaflet to hand out, as well as the two-sided flyer on best practice, based on outcomes from nursery sampling to date. Both were well received. Sarah conducted interviews for, and completed, 7-8 garden centre questionnaires, gaining an appreciation for the difficulties of social science research! She also presented a talk at the show, following which the ‘Gardening Which’ magazine approached to say they would stop recommending purchasing ‘bargain’ (i.e. unhealthy-looking) discount plants to their readers!
David and Sarah went to the Oomycete workshop in Boston in July 2018. Their presentations covered this project, in particular how we engaged nursery stakeholders. David said their approach with validation got praise. David, Leighton and Sarah will all present talks at the ‘DNA Working group meeting’ in Derby, 26-27 November 2018.
Discussion followed on the nature of the focus groups that need to be held. What should they be about? How would it link to outcomes on accreditation? We need to hold them before the end of December 2019. Sarah suggested a focus group aimed at building in more with the HTA initiative and associated groups. Mariella clarified that a focus group meant an intense discussion for a couple of hours with just 8-10 people and she said that Fera should be included.
Beth suggested a group to discuss risk model outputs, perhaps toward the autumn, either within the October stakeholder meeting or as a stand-alone. At the stakeholder workshop next year Sarah would like stakeholders to receive a clear message from each WP including practical outcomes.
Mariella returned to the questions that might be asked at a focus group with nurseries. For example how do nurseries react to what their consumers think? She wondered if there were any outstanding questions not covered or not covered adequately so far. If so, Mariella would welcome feedback.
A comment was made that we all use specialised vocabulary in our presentations. What words are washing over everyone else? It would be helpful not to use too much jargon for the public audience. It was agreed that we should always try to deliver the key messages to stakeholders in easy language to follow (maybe a glossary would help too).
Discussion of future funding options – Sarah Green (FR)
The meeting finished up with a discussion on further funding opportunities given that current project funding ends in March. There has been a no-cost extension until end of December 2019 but no further funds for staff time.
Sarah is putting together a proposal to the Defra future-proofing plant health (FPPH) programme 2019-2022. Within this she is looking to build a collaborative framework to support the continued development of accreditation. The feedback was that the timing was good for collaboration and the Phyto-threats project has lot of information to feed into scheme.
Another part to the FPPH proposal is aimed at developing a standardised nursery testing protocol for metabarcoding to be incorporated into an accreditation scheme. This would use methods already developed as part of the project but broaden the scope to include other types of pathogens such as bacteria. It was suggested that nematodes could be included too. It was agreed that community modelling for areas within the nursery prone to accumulation of pathogens would be useful.
On the question of the ability of metabarcoding to assist in nursery surveillance during statutory plant health inspections, the comment was made that, crucially, you need to isolate an organism at some stage. Also, results need to be fast so staff can go back to the nursery almost immediately.
Another suggestion was that it might be possible to tie in nursery sampling with something else of service to the nursery, for example testing peat/water/ cuttings from Kenya! Even to ants or leaf-hoppers! Technology moves on too, the processing gets faster, better.
It was asked if Sarah had seen the HTA’s paper on how technology is being managed because this might give further guidance. Sarah confirmed she has had discussions with the HTA plant health lead and that he is supportive of the Phyto-threats project feeding into the Assurance Scheme.
Other thoughts on funding included developing something on the macroecology of Phytophthora and related groups. It might be possible to relate success in environmental/hosts/traits to host/pathogen relationships and latitudinal or elevational gradients i.e. what makes species successful? There could be added value of joining up research on trees with ornamental and agricultural research. The BBSRC standard grant round might be a good starting point, i.e. better predication on landscape-scale spread.
A comment was made in regard to the proposed accreditation scheme that the UK garden centre association has an annual meeting. They might be interested.
Two of the WP4 team have submitted a NERC DTP PhD proposal for continuing some Phytophthora genome work.
Another area for future consideration might be to identify where the unusual species have come from? What is the point of entry into environment? Traditional paperwork trail has been used as a tool. If you think about the nursery data, how does that Phytophthora come to be there? You can monitor sewage, fish farms, eDNA. NERC might provide support? It was also asked if there would be value in mapping Phytophthora species data for the UK? One of the bioinformaticians also works on a bacterium and is trying to identify through sequence analyses whether bacteria are transmitted vertically through seed or through the environment. They need to find the right bit of the genome, then they have to train the tool.
Following the end of this discussion the meeting was closed and Sarah thanked everyone for their contributions.