Phyto-threats project team meeting

October 3rd 2017, APHA, Sand Hutton, York

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 May 4th 2017, and to outline and receive feedback on future research plans.

The meeting started with a welcome by Sarah Green (Forest Research, FR) and brief introductions from everyone present including their affiliations.

09.30-11.00: WP1 Phytophthora distribution, diversity and management in UK nursery systems - David Cooke (JHI) and Pete Thorpe (JHI)

David Cooke introduced the WP1 team and reminded everyone of the WP objectives and methods used for the fine-scale sampling of the 15 partner nurseries, including some of the practical issues. A simple questionnaire is applied to each nursery to collect basic data and to get to know the manager, asking for example how material comes onto site and where it goes off site, where material comes from, where it goes to, what water sources are used (ie borehole/river/pond) and how it is treated, and whether the manager has any particular concerns. David explained that sampling involves collecting material from known and unknown Phytophthora hosts, some common to all sites and a mix of symptomatic and asymptomatic. Plant tissue collected is mostly roots. For water sampling, water is passed through batches of potted plants in trays which are left to sit for 30 mins or so. This allows for any Phytophthora propagules on the roots or in the potting soil to be flushed out. The flow-through water is then collected and filtered to trap Phytophthora propagules on the filter. Water supplies and water collection areas (ie puddles, irrigation ponds, drainage ditches) are also tested and filtered. David reminded everyone, and planned to reiterate to stakeholders at the workshop the following day, the importance of water in Phytophthora spread and that water management in nurseries is a critical area to get right in order to control disease. On each nursery site the team evaluates potential disease control points and contamination hazards but it’s always a balance between time available and the need for detail.

In terms of a sampling update for the fine scale sampling, the WP1 team have carried out 34 sampling visits to the 15 partner nurseries (6 in England, 1 in Wales and 8 in Scotland). Thus each nursery has been sampled at least twice, and a third round of sampling is currently underway, with a total of 1700 nursery samples plus associated meta-data collected so far. Just over 400 samples have been PCR tested for Phytophthora: this includes 93 plant root samples from 35 different host types, 132 water filter samples and 170 samples of buffer associated with the filters. Isolations have also been attempted from selected samples, resulting in three confirmed P. austrocedri findings and a finding of P. cambivora on shelterbelt trees.

For the OPAL community sampling, carried out through co-operation with David Slawson and Vanessa Barber of the OPAL project, the sample kits were sent to the volunteers in June 2017 and a training course/skype video made by David to inform volunteers of the sampling protocol. Ten samples from OPAL have been received to date from various sites in North Wales, all collected by one of the OPAL volunteers. David also showed a leaflet that is being used to explain the project and to aid understanding of Phytophthora diseases.

The broad scale nursery sampling is carried out as part of the national statutory sampling programme in liaison with Alexandra Schlenzig (SASA) and Jane Barbrook (APHA). This element of the project targets 50 nurseries/garden centres in England and Wales and 25 in Scotland (to be sampled in 2017/18). Sample packs were sent out to Plant Health Inspectors earlier in the summer and 27 sample packs have been returned so far (ie representing 27 nurseries; 11 in Scotland and 16 in England and Wales). Each sample pack contains 5-10 different root samples per nursery as well as a limited amount of information on each nursery.

For Phytophthora detection using metabarcoding, the Phytophthora-specific PCR assay is the first key to understanding how many samples are Phytophthora +ve. Currently JHI has a large backlog of samples, and they are splitting the lab processing with Forest Research, who are dealing with the root samples. There are no data yet on Phytophthora species findings as the first Illumina plate failed quality control, probably due to sample loss during one of the clean-up stages in the library preparation. The plate is being redone with the aim of completing the Illumina runs this month. The sequencing runs will also include synthetic control sequences generated as a test of; a) sequencing error, b) indexing error and c) sensitivity range.

The team at JHI have been putting together a database of known, verified Phytophthora ITS sequences, comparing databases from Santi Català, Treena Burgess and a database developed by David Cooke. Phylogenetic trees have been produced for each database so that duplications and variations among sequences can be observed, including sequence errors, for sample one problem has been the truncation of even some type-strain sequences. A new database is being constructed through manual editing of the phylogenetic trees for the existing databases, for example if three sequences for a given species are identical in each database then any of the three sequences can be used, and any sequences with errors are discarded.

David presented some of the slides he plans to show at the stakeholder meeting the following day, including Phytophthora findings by nursery (anonymised), findings by sample type, and observations while sampling. In terms of risk of Phytophthora coming onto site there is the need to keep water sources clean, to be aware of the health and source of plant material coming in (this presents high risk especially if from EU or third countries), and to consider biosecurity for staff and visitors as mud is a problem in some nurseries. He recommended a concrete pad for delivery/despatch areas. In terms of Phytophthora dissemination on site, plant to plant spread tends to be least problematic in cells raised above ground, and since puddles are oftenPhytophthora +ve then drainage is important. Infection by Phytophthora has been picked up in shelterbelt trees and having nursery ‘hospital/recovery’ areas is not a good idea. Rapid disposal of sick plants is optimal. David also recommended quarantining new plant material if possible – this material is often put at the back of the nursery in unkempt areas. David then went on to show photographs of some of the issues encountered during the nursery sampling, ie puddles, muddy and/or flooded ground, soggy possibly Phytophthora-infected roots. Photographs of good management included covered water-holding tanks, collection ponds that are lined and free of vegetation, well-built drainage ditches, graded surfaces that minimise puddling in key parts of the nursery, for example where vehicles move in and out, plants sitting on raised benches, on well drained gravel or clean mypex with good spacing between plants.

The next steps for the WP team are to complete this autumn round of sampling, accelerate the lab testing, run the metabarcoding analyses to identify the species present, complete the computational biology platform, report species findings to nurseries and begin data interpretation.

Questions and comments:

Q: What about the implications of finding Phytophthora in the nursery – what do they do about it ?.

A: Being aware of symptoms is important and we make management recommendations when reporting on where samples have proven to be positive for Phytophthora. For example if roots from a specific supplier are consistently Phytophthora +ve then avoid that supplier. Also, not all Phytophthoras are pathogens – although we need to be careful when making this statement, for example P. gonapodyides is also a pathogen as well as being fairly ubiquitous in water.

Comment: It is early days yet in the project and as results come in they will help to indicate how nurseries can reduce Phytophthora through making changes to certain practices. In the longer term the project will be able to offer very good advice, very targeted, to guide management.

Q: What about findings of statutory importance ?.

A: Findings based on DNA data alone cannot be a basis for statutory action. The nursery manager however needs to be aware and Plant Health kept informed.

Q: Is there competition between so called pathogenic and non-pathogen Phytophthora species, and if you get rid of the non-pathogens are the pathogens more likely to take hold ?.

A: Interesting point. A study in Austria is looking at the question of whether some Phytophthora species are outcompeting others, and investigating the interactions among species.

Q: What is the best way of killing Phytophthora in plant disposal areas ?.

A: By composting to a standard that ensures the required temperatures are reached. There are numerous published studies on this. For example Fera protocol on P. ramorum suggests that composting will be effective in killing this species. Any waste disposal system will need to be built into the accreditation system.

Bioinformatics analyses

Pete Thorpe (JHI) presented on the metapy bioinformatics pipeline developed by himself and Leighton Pritchard at JHI. This pipeline is freely available on the github open source site. Pete reminded everyone of what the pipeline does including the five clustering tools. A new clustering tool has been recently published (ZOTU: Exact Sequence Variants: Callahan et al. 2017). ZOTU explicitly tries to correct PCR and sequencing errors and has now been incorporated into metapy. There was some discussion on ZOTUs versus OTUs. Leighton Pritchard explained that ZOTU tries to account for systematic sequencing errors before it clusters sequences. This is different from biovariation, which we are interested in. Pete then showed the mathematical model that ZOTU uses for this correction.

Pete also emphasised the importance of the database as the critical determinant of classification accuracy and he showed the differences in outputs from the five clustering tools, and explained the need to trim the ITS sequences to include the ITS1 region only. Blastclust is not specific enough and so will be removed from the pipeline. The next steps are to verify the pipeline and database with the control samples from the sequencing plate and to write a Bayesian based clustering/probabilistic model.

Questions and comments:

Comment: In the POnTE project, which compares metabarcoding detection of Phytophthora with a traditional baiting method, sometimes metabarcoding has not found a species when it has been baited out of the same sample!. Something is not right if this happens.

A: Metabarcoding is never going to be 100% accurate, however we are striving to get it better. If a species is not being picked up then this may be because the sequence is not present in the database. The database team are meeting to discuss this.

Comment: One issue has been with false positives, due in some cases to ITS sequence variability within species. Most ITS sequences for species in Genbank are Sanger-generated so will be the most abundant/easily amplified sequence in that species that is deposited as the Genbank reference sequence. Illumina sequencing has such great read depth that it will also generate reads for less abundant ITS sequences in a species. For example an ITS sequence present in P. gonapodyides also seems to match P. mississippiae. In these cases the less abundant sequence will occur at low read numbers in the presence of higher read numbers of the most abundant sequence. So it can be picked up, though this emphasises the need for data verification/interpretation by those who know the species and their sequences.

A: Yes there will be sequence variants within species – these can be pulled out and identified.

11.00-12.30: WP2 Feasibility analyses and development of ‘best practice’ criteria – Mariella Marzano (FR), Glyn Jones (FERA) and Colin Price (free-lance academic consultant)

Mariella reminded everyone of the WP2 team members, objectives and methods, including the consumer survey (1500 respondents) which explored the plant buying habits of the public (reported on at the last project team meeting on May 4th) and the interviews with nursery managers, of which six have so far been conducted. An on-line smart survey has also been produced targeting a broader range of consumers including nursery owners, garden centres, landscapers and plant-buying members of the general public. This survey is being circulated via a number of on-line avenues.

For the nursery interviews, carried out by Mariella, Mike Dunn and Tim Pettitt, a range of questions are asked on what influences their decisions, where nurseries are least and most able to change, and their perspectives on accreditation. The aim is to interview all fifteen partner nurseries in the project this financial year. Mariella then ran through a number of slides illustrating some of the comments/perspectives received so far from nursery managers on issues such as plant health, consumers, biosecurity practice/challenges and accreditation. Some points to consider on best practice are that even nurseries who don’t import might unknowingly buy plants from another nursery that does import. Some nurseries aim to be ‘green’ by re-using plastic pots etc, however reducing plastic waste in this way also increases disease risk. In terms of accreditation there is cynicism. Source is an issue, for example a plant that came from Holland and arrived in Scotland – is it fair to say it’s locally sourced ?. Also, landscapers are asking for plants based on design rather than asking nurseries what’s possible and suited. This pressurises nurseries to arrange risky imports. If a nursery cannot import then the customer will just get it elsewhere. Some nursery managers feel that accreditation is just a tick-box exercise; customers don’t ask for it, and there is little support for accreditation at present. The most popular place to buy plants now is at the big retailers and garden centres, and these are viewed as being not so concerned with biosecurity. There also seems to be a perception that accreditation would not change the behaviours of fellow nurseries. However, some nursery managers might consider accreditation if the costs were not prohibitive and the required actions not unreasonable, if there was a safety net and a demand from the consumer.

Mariella then posed a suggested list of questions for the focus session at the stakeholder workshop to be held the following day and the ensuing discussion largely centred around those questions.

Questions and comments:

Comment: Other projects led by Fera are asking similar questions and we’re finding that retailers have less interest in talking about pests and diseases than growers.

A: It’s about figuring out how to talk about pests and diseases. Garden centres and superstores don’t want negative messages.

Comment: At the stakeholder meeting tomorrow we should talk about the consumer survey message on willingness to travel further to support accreditation. There is willingness, and it is important to show this.

Comment: Are we using the wrong language to talk about accreditation ?. It’s not about enforcement or accreditation having ‘teeth’ but rather ‘what’s in it for me ?’. For example, if we get accredited and there is a big government planting scheme, will you buy our plants in preference?. Think of the positive benefits, not what will happen if we don’t become accredited. Perhaps the question tomorrow needs to be phrased ‘what would you like to see from government?’. Could the government make the climate more effective for accreditation?', ie if the government has a planting scheme would the government only use accredited plants ?.

A: We shouldn’t mention the government specifically, but rather ask the question ‘what support is needed for accreditation ?’. Then this doesn’t focus on support from a particular sector.

Comment: Use the term ‘credibility’ when talking about accreditation. Don’t underestimate the influence of landscape architects and materials coming straight through Dutch orchards. How do you deal with that ?. Government involvement is a very good point, it is needed for credibility.

Comment: How is the Phyto-threats project addressing existing initiatives for accreditation schemes, for example UK Sourced and Grown and the HTA-led assurance scheme ?. Discussions have been going on with these schemes, you need to know where they have got to.

Comment: We should ask growers how they think they can persuade consumers to buy accredited plants and focus on the positives, ie more healthy, beautiful plants that improve quality for the producer and greater profits too !. What would they need for this to happen ?., what are the economics?. There should also be a clear logo/badge for such plants.

Q: What about a list of possible incentives ? and what are the scenarios ?. Best case, draconian to almost utopian. For example ‘Landscapers like disease-free plants’. Or is it more likely ‘you’re not allowed to sell unless you have accreditation’ ?. (ie accreditation is mandatory).

A: Accreditation needs to be down the whole line, from landscapers right back to source.

Economic feasibility of best practice and accreditation

Glyn Jones presented on Fera parallel activities including the Forestry Commission decision tool project (a generic decision tool for assessing response options to tree pests in the UK) which is relevant to the Phyto-threats project as it might be applied to new Phytophthora outbreaks in the future. Glyn outlined the tool including why it was needed, what it does, how it was developed and its limitations. A draft final version of the tool is now in review. Essentially, the tool was developed because of the need to provide support on the economic impact of new or unknown species in a new area within a very short timeframe. The idea was to produce a single model for all pests and diseases for all end users using a standardised framework for scenario assessment. Development of the tool required input from a wide range of end users (economists, scientists, policymakers) who formed a steering group.

Glyn ran through the model and its development in his presentation. It is a combination of a prevalence model (based on pest/disease surveillance and a rule of thumb on how hard you are looking versus the actual infested areas found plus an epidemiological model) linked to an economic impact model. Glyn listed at the end some of the model’s limitations including the spread model which assumes a constant spread of the pest/disease over time. Also there are very few control options for new pests and diseases so not being able to input various control options into the model is a weakness. There are also uncertainties as to which environmental values to incorporate into the model. Glyn finished this part of his talk with the questions; will the model be used ?, should it be used ?. It is a highly generic model that can be abused. How should it be used ?.

Finally, Glyn outlined plans for a workshop to be held in November 2017 on cost and responsibility sharing by industry when pest and disease outbreaks occur. Points to be discussed at the workshop include current biosecurity activities, how an industry scheme might work, and incentives to join. There will be cross-reference with the Phyto-threats project via Mariella and Gregory.

Questions and comments:

Q: Can you put case studies in, for example P. ramorum, horse chestnut bleeding canker ?.

A: Yes, and they are there.

Q: Is this decision support only for outbreaks ?. Is there decision support for interceptions ?.

A: Probably this tool could be adapted for interceptions if it were represented as the minimum area for an outbreak.

Colin Price then gave a presentation demonstrating how an adapted version of the CARBBROD model, developed for Dothistroma, could be used in a scenario analysis to estimate the cost of potential timber and carbon impacts when Phytophthora outbreaks occur. This analysis will be done in year three of the project to explore the wider costs associated with no change to UK nursery practice (i.e. a continuation of the status quo in which new Phytophthora impacts can be expected) compared with a scenario where changes to nursery practice will reduce the likelihood of future outbreaks by new Phytophthora species.

The model is based on Forestry Commission yield models. It uses DECC (Department of Energy and Climate Change) carbon prices based on international agreed limits, or the social cost of carbon (i.e. the effect of increasing CO2 concentration), or any other carbon price schedule. Discounting is done using the Treasury’s schedule of discount rates, or any other schedule, or a single discount rate. When there isn’t a unique carbon price, and when discount rates vary, every year has to be evaluated on its own, and all individual values summed. Thus tracking carbon pools into the indefinite future is a challenge. Colin showed how the model predicts the optimal rotation value for Japanese larch. Sometimes CARBBROD gives better carbon values when there is infection in trees, using the DECC values. This isn’t the answer we would expect. Other carbon prices give more expected results.

Questions and comments:

Q: Has CARBBROD been tested against real-world data? i.e. documented case studies? Do the predictions emulate what really happened?

A: All data come from the real world!

Q: Can you add more than just carbon values to give you a more intuitively correct answer?

A: Various impacts of trees have been looked at, for example temperature reduction, pollution absorption, flood alleviation. Carbon values are so big that they will always dominate models.

Comment: More talks with pathologists are needed since it is not clear what the primary inputs would be for the Phyto-threats project. Assuming best nursery practice, what scenarios can we predict? This is a two-way discussion between economists and pathologists on the team and we need to identify over the next six months specific scenarios of use/interest to the project that Colin can apply to CARBBROD. For example WP3 might identify specific high-risk Phytophthoras not yet present in the UK, but which might enter the country if no change in existing practice occurs, that could be applied to CARBBROD.

A: Gregory/Colin will initiate these discussions by email.

Read about the afternoon session