Project team meeting

James Hutton Institute (JHI), Invergowrie, Dundee
May 4th 2017

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 over the first year of the project for the three work packages that have started, and to outline and receive feedback on future research plans. Following the talks and discussion sessions participants were given a short tour of the JHI pathology and sequencing labs.

The meeting started with a welcome by Sarah Green (Forest Research, FR) and brief introductions of 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, which consists of David Cooke, Leighton Pritchard Peter Thorpe, Eva Randall, Beatrix Clark (all JHI), Sarah Green (FR), Debbie Frederickson-Matika (FR), Tim Pettitt (Uni of Worcs), Alexander Schlenzig (SASA) and Jane Barbrook (APHA).

Nursery sampling

For context, David provided a snapshot of the distribution of P. infestans clones in Europe from 2013-2016 based on 5000 isolates from the Euroblight project, showing how an airborne pathogen can rapidly change its distribution. David then gave an overview of WP1 objectives and methods for the nursery sampling, including some practical issues such as range of plant species sampled, which varies among nurseries according to what stock they hold. Generally, the sampled plants are a mix of known and unknown hosts for Phytophthora, and may be symptomatic or asymptomatic at the time of sampling. Mostly root samples are taken, except in a few instances where foliage/stem samples have been taken from symptomatic plants. Water supply is sampled at source and run-off. In general, sampling effort at each nursery is a balance of time available and the need for detail, focusing on working through the hazard points – ie incoming plants/water/pots/ground.

So far the team have carried out 18 sampling missions at 15 nurseries (6 in England, 1 in Wales, 8 in Scotland). This has resulted in 1009 samples (everything in triplicate) including plant roots (93) from a range of 35 hosts, 132 water filter samples and 170 samples of buffer associated with each filter. Water samples came from water washed through plants, boreholes, ponds/ditches, equipment washing (eg trolleys) and water control blanks. To date, 395 samples have been PCR tested for Phytophthora. David described some of the necessary changes to lab protocols to improve the efficiency and accuracy of processing.

Data were presented showing the number of Phytophthora-+ve samples for 10 nurseries (labelled 1-10 and not named in the presentation). Sample types were broken down into water filter, water buffer and roots. There were clear differences among nurseries in relation to the number of Phytophthora +ve and –ve samples, and this could be related to practices observed during the sampling missions. David did outline the issues of working with each sample in triplicate, ie when 2 samples are negative for Phytophthora and 1 is positive, and how they deal with such results.

There was discussion on the different nursery sizes and practices and how that relates to the number of findings per nursery as this should be taken into account when presenting data. Hard evidence is required linking practice to Phytophthora findings. Mariella Marzano (FR) made the point that a survey of management practices accompanies each nursery sampling visit so that a basis of best practice/worse practice can be developed in relation to Phytophthora findings. Participating nurseries are willing to know their results in order to improve their practice. Nursery data include turnover/plant species sold to help interpret findings. Detailed (fine-scale) sampling is only done for nurseries that volunteer. The question was asked whether sampling was designed to avoid bias, ie if you focus on a sick area of plants within a very large nursery you will bias the data. This is more a problem for plant samples than water samples. David Cooke confirmed that detailed notes accompany each sample collected, including whether samples were collected on a random basis or due to presence of disease symptoms.

A broad-scale survey is due to start soon in which plant inspectors will be collecting root samples for the project from nurseries during their routine Plant Health inspections. In particular it will be important to clarify with inspectors where and what they should be sampling. This element needed further discussion between APHA and JHI. David made the point that the aim of the broad scale survey is not to repeat the statutory testing but rather to look more widely across nurseries for a different insight that will help inform the goals of the Phyto-threats project.

It was suggested that nursery samples are collected from areas not expected to have disease (for example where plants are apparently healthy), just to check what Phytophthoras might be ‘hiding’. In such cases the foliage could look healthy but perhaps the pathogens are present in the roots or soil. Some Phytophthoras such as P. gonapodyides could just be ‘root nibblers’ and not viewed as pathogenic. The consensus was that it is still useful to know the distribution of non-aggressive Phytophthora species to see how they are distributed. The presence of these species might indicate a route in for other more pathogenic Phytophthoras.

It was asked whether information is available on where nurseries with Phytophthora-positive samples have sourced their plants, ie from plant passport numbers, and whether such information could be used to know what pathogens are in those areas where plants are imported/bought in from ?. David Cooke responded that data on plant sources are with each nursery and could be obtained from some of them, however, the true origin of the plants might not be known by their plant passport number. The length of time that plants have been in a nursery is a factor included on the nursery questionnaires.

David Cooke listed the Phytophthora findings to date by sample type, showing a higher proportion of positive samples from roots than water, although this would reflect the fact that sickly plants are targeted for the root sampling. Different production methods included;

  • Water sources; mostly borehole, some mains plus rainwater and river water (‘filtered’).

  • Plants in cells/pots above ground or on the ground, or bare root grown in the ground (ie forestry nursery).

  • Production systems included wholesalers/holding/importing to production on-site.

  • Premises included garden centres/nurseries/mixed/horticulture/hedgerow/forestry tree growers.

  • Generally good plant health awareness was observed but fungicide use was widespread.

David then showed examples to illustrate sampling points, such as water sources and types of plants sampled.

In terms of the next steps for the nursery sampling; David Cooke is to discuss the broad scale sampling with Jane Barbrook (APHA) and Alexandra Schlenzig (SASA), as well as co-ordinating the OPAL sampling with David Slawson. For OPAL, water samples from streams and other waterways will be collected by OPAL community volunteers in Plymouth, Cardiff, North Wales and Glasgow. David Slawson emphasised the need for good photographs and descriptions of sampling methods for the volunteers, and also expressed the need for rapid feedback on OPAL results to enthuse and engage OPAL volunteer interest. The fine scale sampling of the 15 partner nurseries will also be repeated in June/July and again at the end of the year. A question was asked whether root samples are also taken from plants sampled by water flow-through (ie plants are placed on a tray and watered to capacity, left to stand for 30 mins or so, and the water flow-through in the trays then sampled). If so, this would provide an interesting comparison of the two methods in terms of Phytophthora detection. David Cooke said that where possible roots were sampled from plants also subject to flow-through sampling, although sometimes time constraints prevented this. At the moment there is also a backlog of samples requiring DNA extraction from the previous round of sampling and getting this done will be a priority.

The first 176 Phytophthora-positive samples will be run through Illumina sequencing in May 2017. This will produce 15M barcode reads=156K reads per sample; each read approximately 250bp in length. Synthetic control samples will also be included in each sequencing run as a test of error rate.

Bioinformatics analyses

Pete Thorpe (JHI) presented on the bioinformatics pipeline that he and Leighton Pritchard (JHI) have developed called METAPY – this pipeline is a key output from the project and will soon be publically available on GitHub for use by the wider scientific community. Pete began by describing the system for Phytophthora identification from sample collection, to DNA extraction, nested PCR of the ITS1 region with Phytophthora- specific primers, DNA library preparation, Illumina sequencing and analysis of sequencing reads by the bioinformatics pipeline, METAPY. Pete explained that other pipelines tend to use one clustering tool (ie to align sequence reads to a reference sequence in a database), however, each clustering method can vary and so METAPY produces individual results for five different clustering methods (Blastclust, Swarm, V-search, CD-hit and Bowtie) so that results can be validated by comparing the different methods.

Pete ran through the different clustering methods and showed results from a control sample containing a DNA mix of ten known Phytophthora species. He explained reasons for variation and false-positive results for each clustering method. Blastclust was the least discriminating method and reported 27 species; many of these false-positives were species in the same clade. Bowtie, on the other hand, only detects species with a 100% match to the reference sequence in the database and reported only 7 species in the test. CD-hit found 20 species, Swarm 16 and Vsearch 16. The reference database is yet to be adjusted for species which have highly similar or matching ITS sequences and this might sort out some of the false-positives. METAPY can also be set to error-correct reads before identification (Illumina can be prone to error – these sources of error are known and can be automatically corrected). METAPY is currently used for the ITS1 region but can be edited for other genes.

Pete then described the main issue with using the ITS region which can exist in 50-500 copies in the Phytophthora genome based on assessment of ITS copy number using qPCR. The short reads produced by the sequencing method can’t resolve repeat/repetitive regions and so these get collapsed to consensus sequences. Perhaps a different sequencing method such as PacBio, which produces longer reads, will help to resolve this. Pete is re-assembling ribosomal DNA regions for all Phytophthora genomes to determine copy number and within-genome variation.

The question was asked how Phytophthora hybrids might be identified using the bioinformatics method. David Cooke agreed that chimera detection in the pipeline might throw them out and thus identifying hybrids will be tricky. Ana Pérez-Sierra (FR) mentioned the POnTE project which is assessing both ITS1 and the Cox region for Phytophthora detection to see how resolution compares. For other fungi, for example powdery mildew, the single copy B-tubulin gene is used for barcoding.

It was also asked whether the different ITS sequences present within an individual would show up in the metabarcoding data given the very high number of ITS reads produced with Illumina, so that within-species variation would be identified. Or, is it only the predominant and more abundant ITS sequence that gets resolved?. Sarah Green (FR) described the scenario with P. austrocedri that had arisen in a previous metabarcoding study in which some samples yielded high numbers of reads of an ITS sequence typical of the UK lineage plus a few reads of an ITS sequence typical of the Argentinian lineage. Initially it was thought that the Argentinian lineage might be present in Scotland – however, subsequent analysis of the P. austrocedri genome revealed that the UK lineage also has the Argentinian-type ITS sequence in the repeats, probably at lower copy number. David Cooke responded that some species do have multiple ITS types and he wants to build them into the reference database.

Beth Purse (CEH) commented that the WP3 team are receiving data on global Phytophthora distribution based, in many cases, only on ITS sequence detection. Ca we be confident in these data?. David Cooke replied that whereas some species have identical ITS sequences and cannot be resolved, the majority of Phytophthoras can be distinguished based on ITS. Taxonomic irregularities can also cause difficulties. The WP3 team will need to have clarification over which species can be confused so that they can mine the data that they have compiled. Can they trust the Blast analyses used to identify species?, and is there a timeline for species identification?, ie before a certain date the identifications would have been based on morphology alone and thereafter more likely to have resulted from mainly molecular tools or a combination of both. David Cooke replied that from about 1998 onwards, Phytophthora researchers were using molecular tools and thus data are more likely to be reliably discriminated. The trick is to avoid over-interpretation. We also need to take care not to interpret previously uncharacterised sequences as new species just because they are not on the reference database. There was then a discussion/comment that ideally we should know the range of ITS ambiguity in each Phytophthora genome so it can be incorporated in the analyses. The existing reference database of Phytophthora ITS genomes is still being worked on and accuracy of the database is key to the outcome of analyses. Currently it is believed by the team that METAPY gives the best likelihood of a given Phytophthora species being present compared with other pipelines.

11.00-12.30: WP2 Feasibility analyses and development of ‘best practice’ criteria – Mariella Marzano (FR) and Mike Dunn (FR)

Mariella began with an overview of the social and economic research being carried out by herself, Mike Dunn, Gregory Valatin (all FR), Colin Price (contractor economist) and Tim Pettitt (Uni of Worcs, nursery engagement) and she reminded the project team of the WP2 objectives, listing milestones and outputs. Essentially this WP has three key parts; i) social analysis to assess applicability of nursery best practice criteria, ii) cost-benefit analysis of implementing best practice and iii) which elements of best practice that should underpin an accreditation scheme.

So far the WP2 team has mapped the stakeholder networks and created a stakeholder database. The consumer survey was supposed to be done in year 2 of the project, but since there was early interest in this work the survey was conducted in year 1 (more detail to follow). The team have done some context building; interviewing science team members and members of the Expert Advisory Panel to get a sense of what is needed, and Mike Dunn (FR) has also joined the WP1 team on nursery sampling visits. They are exploring existing values within the sector, experiences, and practices on disease and management. They are planning to conduct interviews and undertake participant observation at nurseries starting this summer. One of the key factors to assess will be potential attitudes and willingness to join an accreditation scheme. The team will also do wider industry focus groups but they have yet to finalise the methods. They will be led by the data as it comes in. More surveys are being developed for landscapers, nurseries and garden centres, including supermarkets and superstores. They will prepare questions on the supply chain, disease threat perspectives, current management, policy tools etc and obtain information on decision management, where can and can’t nurseries change, perspectives for future management and willingness for an accreditation scheme.

Consumer survey

Mike Dunn (FR) gave a descriptive overview of results obtained very recently from the consumer survey which involved 1500 people. The data have yet to be analysed statistically. Mike ran through the questions addressed by the survey and the demographic data for the respondents who were typically members of the public who buy plants for their own home. The 19 survey questions were compiled and sent for approval by the project’s ethics committee before the survey was sent out. The survey was undertaken by a survey company (Toluna) and it targeted buying habits, perceptions of the seller’s biosecurity, and attitudes towards accreditation. The 1500 respondents were generally representative of the nation’s demographics although 57% were aged 55 or over.

Survey results

Most people buy plants 1-2 times per year, most frequently sourcing plants from garden centres/DIY stores/supermarkets. Mariella noted that the project has not engaged these sectors yet and doing so will be important. Plant quality, cost and range were the top three factors in a consumer’s decision about where to buy plants from. Evidence of good biosecurity practice by the business came out as a fairly low priority. When it comes to selecting plants to buy, plant appearance, suitability for the intended site and cost were most important. Provenance and origin of stock were very low priority for consumers.

In terms of general awareness of pest and disease threats to plants; 10% of respondents had never heard of the problem, 62% had heard of the problem but didn’t know much about it, 24% had heard of the problem and felt reasonably well informed and 3% felt very well informed. Respondents were assessed for their perception of the level of risk of different potential pathways for pest and disease introduction; online and mail orders were perceived as higher risk whereas consumer’s self-grown plants, specialist plant suppliers and nurseries gave most confidence as being of low risk. In terms of willingness to support an accreditation scheme; 9% of respondents chose not to buy accredited products because of the increased cost, 32% gave accredited sources little thought when making a purchase and 38% bought accredited products in some cases because they believed in the goals of the scheme. In terms of a potential accreditation scheme for plant sellers, 25% of respondents agreed with the principles of such a scheme but had concerns about the cost being passed on to consumers.

The consumers surveyed spent on average £100 per year on plants, and 45% were happy to travel further (up to 160 miles, with an average increased willingness-to-travel distance of 26 miles) and 39% were willing to pay more (on average 18% more) for accredited plants.

Overall, the results of the consumer survey indicated a general willingness to support accreditation (within reason) but it also highlighted a need for consumer engagement about the risks of spreading pests and diseases.

Discussion of consumer survey

Mike’s presentation generated a number of questions such as ‘are attitudes to accreditation associated with i) a greater knowledge of pests, ii) experience with existing accreditation, iii) number of plants purchased, iv) types of plant sources, v) demographic factors ?.’ For example, it was asked whether the previous experiences of the consumer affected their attitude towards accreditation, ie if trees had been felled in their local area then is this likely to have increased their awareness and concern of pests and diseases?. The response was that, based on previous surveys, people generally don’t notice damage to trees in their area. The point of this survey was to find out where people buy their plants and what influences their behaviour. The OPAL project, involving community volunteers including schoolchildren, might help with raising awareness of younger generations. Another question was raised as to whether the behaviour of online respondents who completed the survey could be expected to differ from the public at large. The answer from Mike Dunn was that, overall, he is happy with the survey sample as it targeted a representative population in relation to location across the UK, gender and age bracket, bearing in mind that older people tend to be more likely to buy plants than younger people.

Next steps

A survey will be designed for the forestry industry and landscapers, and the project team and Expert Advisory Panel members will be asked for their inputs in shaping the survey questions. For the nursery and garden centre survey it is recognised that these groups can be both sellers and consumers. Mariella will also link in with the Horticultural Trades Association (HTA) who are working in partnership with nurseries to develop an assurance scheme.

The comment was made that growers need clear plant health advice on which to base best practice and that any assurance scheme should include management practice guidelines. The question was asked whether the HTA-piloted scheme was just for UK growers or international. Mariella responded that the scheme was for the UK market but the HTA would like it to extend internationally. The scheme is being piloted now and could possibly be launched next year.

Economic analyses

Mariella presented a number of slides on behalf of Gregory Valatin (FR economist researcher) who could not attend this meeting. Gregory will undertake an economic analysis of consumers’ willingness to pay and travel for accredited stock. Gregory will look at cost-benefits of other schemes, on potential impacts of new Phytophthoras entering the UK and potential costs of not having best practice thus increasing chances of new diseases establishing. Gregory had a number of basic questions for the project team and advisory panel including;

  • What kind of accreditation scheme shall we focus on?

  • What is the source of additional costs-holding plants/use of fungicides or pesticides?

  • Timescale for introduction of scheme?

  • To what extent are nurseries expected to take up scheme?

  • Timescale for uptake?

The team will consider these questions and report back to Gregory.

Sarah Green (FR) commented on a study from Californian native plant nurseries supplying restoration plantings showing how Phytophthora infections were greatly reduced after the nurseries implemented a number of key management changes. Although there were costs to implementing these changes, the resulting benefits were clear, ie less Phytophthora and less risk of passing disease to native landscapes, so reducing costs associated with outbreak eradication.

The question was asked ‘if we don’t have a scheme, what would happen?’. The answers were that there must be data on the costs of remediation following outbreaks. Obtaining such data could be tricky with some outbreaks being public and some private. Essentially, we don’t really know the baseline costs of disease eradication efforts. It may be possible to look at past trends and establish how action could level off the need for intervention.

The discussion returned to the topic of uptake of accreditation by nurseries; if a scheme involves too much paperwork it may restrict uptake. How do we bring growers on board?. Smaller nurseries will find it costly and not participate.

Gregory’s work will assess the anticipated effects of an accreditation scheme on introduction of Phytophthoras, their spread in the UK, effects on introduction of other pests and diseases as well as spread of pests and diseases already present in the country.

A last question posed: ‘What are the differences likely to be, compared to baseline risks, and how do we quantify them ?’ eg. % reduction in time of species x in year y.

Glyn Jones (FERA) presentation on Future Proofing for Plant Health (FPPH) project

Glyn presented on a couple of economics-related projects he is working on. One is funded through the DEFRA FPPH programme and is looking at cost and responsibility sharing, modelling nursery networks and assessing markets for plant health and biosecurity. He is considering the public good characteristics of biosecurity, ie market failure if nurseries won’t provide the biosecurity that the public wants. Glyn presented a figure showing nursery sector complexity and the impact of poor biosecurity, and the difference between a pest staying put or moving into wider environment.

Glyn commented that we are living in changing times and ran through traditional government practices and instruments for intervention, including inspection, monitoring and surveillance, quarantine and pest exclusion. Research funding allocations reflect budgetary pressure from increased and changing risks. He asked if the industry is making a move now due to concerns over Xylella and the impact of Chalara. How should an accreditation scheme be designed?.The situation is currently dynamic, with various Woodland Trust, Grown in Britain and HTA assurance schemes. The Woodland Trust want clean sourced trees, but there is a potential problem with the source, for example if the order is large then trees might have to be sourced elsewhere (ie outside the UK). Glyn also ran through a number of possible economic mechanisms to support an assurance scheme, such as i) cost sharing for outbreak control measures, ii) insurance (easier if the pathogen is defined and involves a small defined group of participants), iii) environmental bond (ie growers pay money, and if no disease they get reimbursed – this may have a positive impact on behaviour, iv) retrospective levy, v) cost sharing for risk reduction measures.

Interviews have been undertaken by Glyn and his team with some of the key players in the industry. There were some criticisms of accreditation/assurance schemes, in that if a scheme were voluntary you would likely only get growers joining who already operate best practice. If a grower were to be part of a scheme would they be subject to fewer Statutory inspections?. Generally, the idea of reduced inspections was not supported by industry, they actually like being inspected!. Questions that arose included why would growers join a scheme ?, what are the benefits to them?, what do you do about those growers who don’t join?. The feedback from growers was that any scheme would need to be government-endorsed and clear, with a strategic direction. There needs to be leadership from DEFRA and interaction with policy. The responses to malpractice need to ‘have teeth’ to incentivise adoption of good practice.

In terms of UK biosecurity policy, proactive temporary import bans of high risk plant material are hard to achieve. It was suggested that there should be expansion of nursery licenses and inspections to other importers, for example landscapers and landscape architects. One possibility would be to base border inspections on the biosecurity rating of exporting nurseries.

In general, growers supported the notion of fewer rather than more assurance schemes and the idea of earned recognition, in that there would be benefits to members of such a scheme. There can be market distortion and there is often a mis-match between demand and time to supply. Growers said that there needs to be increased distinction of UK-grown material and a transition to more home-grown, and perhaps more co-operative working of growers in order to achieve required supply. There needs to be confidence in the domestic market.

In terms of future FPPH work, Glyn said that he would be integrating with the LWEC Phyto-threats project and would liaise with Mariella and Gregory. He will be conducting more interviews of those involved in trade and undertake a demand-pull analysis, modelling design and required uptake of an assurance scheme. Glyn also mentioned the RAPID trade project, funded by BBSRC and US National Science Foundation. As part of this project they have been modelling high biosecurity trade networks following pest/disease outbreak scenarios.

It was asked whether a UK-wide accreditation scheme should be compulsory. There followed some general discussion of this along with suggestions of a licensing scheme (which might just benefit the largest operators) and whether public benefit could be used to influence what public funds could go into the scheme.

Afternoon session >