Skip to main content
Contact Us

Phyto-threats project team meeting April 2018

April 26th 2018 held at CEH, Wallingford, Oxfordshire

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 October 3rd  2017, and to outline and receive feedback on future research plans.


WP1 Phytophthora distribution, diversity and management in UK nursery systems – David Cooke (JHI), Leighton Prichard (JHI), Pete Thorpe (JHI) and Tim Pettitt (University of Worcester)

David Cooke introduced the entire WP1 team, including those involved in nursery sampling, processing in the lab, and analysing data, and he reminded everyone of the WP objectives and methods. He noted that processing of the filters from the water samples is very time consuming and that processing only the buffer solution that the filters are placed in might yield the same information with much less sample preparation time. They will be comparing buffer and filter samples from a subset of nursery samples to ensure that they are not missing any phytophthoras trapped on the filters.

For the fine scale sampling they have now obtained 49 sample sets from the 15 partner nurseries (6 in England, 1 in Wales and 8 in Scotland). Some results have now been reported back to all nurseries and this has promoted a more positive attitude among nursery managers towards the project. The WP1 team now have 2292 samples from 150 different host plant species, plus associated metadata. Of these samples, 1660 have been PCR tested for Phytophthora, including 617 from plant roots, 193 water filters and 850 buffer samples associated with water filters. They have also carried out isolation from some samples, yielding P. austrocedri, P. cambivora and P. lateralis. The team have been testing new methods for DNA extraction and have moved on to using kits rather than the phenol/chloroform method.

A question was asked about how sick plants were targeted – by eye?, or does the nursery manager ask the team to test specific plants? The answer was both, and that field notes and photos are taken of sampled plants so that results can be linked back to symptoms.

For the OPAL community sampling of local waterways, in cooperation with David Slawson and Vanessa Barber, the WP1 team have received 26 samples to date from three locations (north Wales, south Wales and Glasgow). These have been PCR tested for Phytophthora and about 50% are positive. This sparked some discussion on the usefulness of rolling out in the future the OPAL citizen science/public engagement element to a broader project looking at pathogens in the wider environment, particularly for early detection in sentinel plantings.

For the broad scale sampling, 64 nurseries have been sampled to date; 25 in Scotland and 39 in England/Wales. Samples usually involve 5-10 root samples per nursery. So far 408 root samples have been received. Root extractions have yet to be done on these samples. Sample packs for the 2018 broad scale sampling have been sent out to PHSI.

David then showed a range of photographs of plants sampled across nurseries and other general observations and reiterated that messages about management are being communicated to nursery managers. He presented a slide showing progress since the last project team meeting in October 2017, including data on the large number of samples processed during this time as lab testing has been accelerated. The first Illumina plate has been run and analysed, and results from the PCR testing and Phytophthora species data were reported back to nurseries in March 2018. David went on to show a number of graphs summarising PCR results by substrate and nursery and emphasised that data interpretation in relation to management practice is needed next.

There was a question on whether the sampler affects the number of positives, and that it might be good to see if certain people are getting higher numbers of PCR positive samples. The answer was that there is a fairly large team of people sampling most nurseries and that the team always includes at least one experienced plant pathologist. The sample team usually walks around the nursery first thing and decides together where possible which plants to focus on.

Another question was asked about whether the distribution of PCR positives and negatives could be plotted across nurseries to look at outliers, ie those nurseries with particularly high or low numbers of Phytophthora positives, and that could be related back to management practice. The discussion moved on to how the nursery sampling data can be analysed statistically, for example, how can we show that a particular practice results in a high Phytophthora load? It was suggested that the project establishes a small working group to decide how to analyse the data statistically.

David showed a slide summarising the Illumina run carried out with the first set of positive PCR samples. Samples containing synthetic control sequences at different concentrations were included on the plate to test the sequencing error rate, the indexing error to set an acceptable threshold read number (currently 50-100 reads) and the sensitivity range. David stressed that the species data obtained in this first sequencing run are preliminary and that full validation is required against a set of reference Illumina sequences obtained from 45 known Phytophthora cultures already set up on a control plate due to be run in May. This should provide important data on ITS1 sequence variation within species.

David then went on to show an example of the results files sent to the nursery managers back in March, including a cover letter explaining the methods and what the results are likely to mean. A brief comment was included on each Phytophthora species reported by the sequence data, ie its usual hosts and whether it is regarded as an aggressive species or not. David also discussed the Illumina findings and challenges – ie most species were Phytophthora, with some downy mildew or Nothophytophthora. There were low levels of Phytopythium in the sequence data. He ran through an overview of species found in some samples, for instance a high diversity of species particularly in river and puddle samples. Some species results were unexpected and we need a means of examining false positives. Reporting results back to managers is also challenging as the implications for management are not always clear. Summary findings will be reported back to the Plant Health Risk Group via Jane Barbrook of APHA. David also pointed out the dynamic nature of Phytophthora taxonomy, showing examples of the clade 6 phylogenies in 2003 (12 taxa) versus 2015 (30 taxa). He finished his presentation by outlining ongoing and future work; that the nursery sampling will be completed in 2018, many more Illumina plates need to be run and the need for data interpretation in relation to management practice, ie practices need to be related to Phytophthora findings.

A question was asked about relating DNA findings to actual viable propagules and that it would be useful to back up the DNA data by doing isolations/colony plate counts. This was addressed in part by a presentation from Tim Pettitt (below).

Tim Pettitt (University of Worcester) was able to demonstrate some results from his baiting and plating water samples conducted at a subset of nurseries sampled in this project. His data showed numbers of viable oomycete spores in different water samples that had also been PCR tested for Phytophthora. He recorded zero viability of oomycetes in a treated water sample (filtered and chemically treated) that was positive for Phytophthora in the PCR test. Therefore we do need to be careful in interpreting data because presence of DNA in a sample does not necessarily indicate viable propagules ie the PCR test is not a good indicator of viability. Tim also demonstrated his amendment to the water sampling methodology using a bike track pump used to push water through a bottle containing the water sample linked to several filters. This greatly speeds up the process of water sampling and reduces potential for cross contamination of samples. He uses bottles of fizzy water – they are sterile and the fizzy water can be used to clean the filter heads and tubes, then replaced with the sample water.

Leighton Pritchard gave a presentation on the bioinformatics element of the WP1 work, entitled ‘Classification Performance Evaluation’. Through the use of sound tracks illustrating distortion of spoken phrases to represent the sequence ‘noise’ that the bioinformatics pipeline has to deal with when attempting to assign a sequence to a species, he demonstrated the importance of removing the ‘noise’ from a sample containing sequences (ie ‘noise’ meaning the sequences and sequence fragments that we are not interested in) so that the sequences are assigned to the correct species. He reiterated the importance of training and test sample sets, for example the 45 known Phytophthora species that will be Illumina sequenced so that we are clearer about which ITS1 sequences belong to which species. This will help us determine true positives and true negatives from false positive and negatives in our nursery samples.

The question arose about how much importance we attribute to distinguishing between closely related species. This would depend upon the species e.g. distinguishing P. rubi or P. fragariae would be not important. Sometimes it’s necessary to take a look at the sequences and determine manually. Some difficulties encountered include reference sequences that appear to differ but are actually from the same isolate sequenced by different labs. Pete Thorpe discussed some of these conflicts in his presentation.

Pete Thorpe finished up the WP1 presentations by running through the metapy pipeline (to remind us of the methods and clustering tools). He also presented an update on the Phytophthora ITS1 reference database in which Sanger sequences have been obtained for 40 Phytopthora isolates and run through metapy to check the accuracy of the database. Some sequences did not cluster to the database but did match 100% to sequences in Genbank; these entries were added into the database.

The performance of each clustering tool in metapy was tested on the Illumina sequence data generated from four samples containing mixes of DNA from known Phytophthora species. For each clustering tool Pete tested the sensitivity, precision, false negative rate and false discovery rate. Some species with very similar ITS1 sequences cannot be separated by most of the clustering tools e.g. P. capsici. Bowtie can separate them, but only if the ITS1 sequence is a perfect match with the database sequence. Basically Swarm performed best in terms of the above criteria but none of the tools were perfect. Manual assessments are necessary when determining which species of a species cluster is most likely to be the one present.

Pete also talked about the error rates in the Illumina sequence output. Four random synthetic control sequences were synthesised with the same mean length and base composition as the Phytophthora database sequences but with no BLAST hit and processed in the same way as the nursery samples. Errors can occur during PCR, through Taq error and during Illumina sequencing. He showed the frequency distribution of errors in the control sequences and where on the sequences the mismatches occurred. Errors also include indels and chimeras. The majority of the error variation occurred within two mismatches. Looking at the sequences, when the clustering threshold was set at 3 mismatches then P. ramorum was mis-identified as P. lateralis; at 2 mismatches, P. ramorum was correctly identified. Since a large amount of the dataset was represented within one mismatch of the control sequences a strict 99% threshold was used to cluster sequences.

WP2 Feasibility analyses and development of ‘best practice’ criteria – Mariella Marzano (FR)

Mariella started the WP2 presentation with an overview of milestones and outputs, and outlined progress so far, including the public consumer survey and resulting publication – a ‘glossy’ summarising the results of the public consumer survey which was produced for the THAPBI dissemination event in February. The online consumer ‘smart survey’ aimed at the public, nurseries, garden centres and landscapers has been distributed via horticultural magazines and email, although participation to date has been very limited. The WP2 team have been carrying out interviews with managers of the partner nurseries, liaising with FERA/HTA over the pilot assurance scheme, and liaising with FERA on the cost and responsibility sharing project, in which Gregory Valatin (FR) accompanied FERA economists during a number of nursery interviews to gain an understanding of the costs of different management practices.

The next steps for WP2 are to increase participation in the online ‘smart survey’. This could be done by using some of the budget to pay a company to conduct the survey for the project team. The WP2 team also need to improve economic data gathering through interviews with nursery managers. Interviews to understand purchasing habits and attitudes towards accreditation are also going to be carried out with retailers and garden centres, local authorities and other managers of large parks and gardens, as well as the landscaping sector. A series of focus groups will also be held in response to the interview and survey findings on appetite for accreditation.

Mariella ran through the questions asked in the nursery interviews and presented some of the findings so far in terms of nursery manager perspectives on disease threats, what they think about management for best practice, and what influences their plant purchasing decisions. She found that appetite for accreditation tended to be based on the size of the nursery and business objective. Some of the perceived benefits are that accreditation will provide reassurance to the customer as well as a training/‘safety net’ for the nursery (in terms of compensation). Participating nurseries would also be ‘seen to be doing something’ and it would allow traceability. Some of the perceived challenges are that there is currently little consumer awareness of the need for accreditation, the benefits of accreditation need to outweigh the cost of membership, and there was scepticism as to whether accreditation would change behaviours. Additionally, there is the common misunderstanding that we are trying to impose yet another scheme, instead of providing scientific evidence that would feed into a scheme. Not every nursery has got the message about disease threats so engagement on this issue will be important, and how would accreditation be policed?

One comment was that there needs to be an update on progress of existing assurance schemes in the UK, looking to see how these could be run. Mariella confirmed that the WP2 team will resume engagement with the HTA and Defra over the pilot assurance scheme to ensure Phyto-threats project findings are used to influence the development of the scheme.

Mariella then held three discussion sessions. In the first session she asked the project team ‘what is ‘best practice’? She listed twelve nursery ‘best practices’ and asked if all of these should form the basis for accreditation and whether other best practices should also be included. This resulted in some discussion with one comment being there was a need for continuous monitoring of stock as part of accreditation criteria. This would require staff training. Another comment was that plant protection products can often ‘mask’ symptoms rather than solving the problem, and that this needed to be considered within accreditation.

In the second session project team members were asked to get into pairs to discuss which key questions the WP2 team should be asking of retailers, garden centres, local authorities and landscapers. Questions should include what are their key suppliers? what is their biosecurity knowledge and experience?, what are their purchasing practices and what influences them? and what are their perceptions of consumer demand?. One important point raised during this session was the need for biosecurity and plant health to be stipulated as part of the plant procurement process to allow purchasers to select what they consider to be the healthiest/least risky bid rather than the cheapest bid. This needs to be taken on board by the government for example in setting plant procurement policies for local councils. During this session Mariella also asked for local authority, landscaper, garden centre and retailer contacts for interviews. Several members of the project team responded that they would be able to supply contacts, and even help with the smart survey in this way.

The third session dealt with focus groups. Three focus groups are required this financial year. These will involve small groups of selected stakeholders discussing questions around feasibility of best practice and attitudes towards accreditation. This could be done via the HTA and the pilot Plant Health Assurance Scheme; for example would they be influenced by the consumer survey? The WP2 team could also meet around APHA and retailers to have a session on best practice and accreditation. More discussion is needed on this.

Finally, Mariella presented some slides on behalf of Gregory Valatin (FR) who is looking at cost-benefits of best nursery practice. The objectives are to undertake an appraisal of the costs and benefits of options for developing best practice in UK nurseries to mitigate risks of further Phytophthora introduction and spread, both from the UK nurseries’ perspective and from the perspective of society as a whole. Mariella asked the team to consider the different scenarios that Gregory could use in his analysis, including the most appropriate baseline of best practice, which nursery characteristics to assume in the analysis and how to estimate cost of benefits to nurseries as a result of introducing best practice.

There was some discussion around whether an accreditation scheme should imply no use of imported seed. Current schemes vary on this issue, with the Woodland Trust’s UK Sourced and Grown Scheme not allowing importation of seed from outside the UK. Seed can certainly be a source of some pathogens. A baseline scenario should involve nurseries having implemented some, but not all, best practices, assuming some importation of stock from outside the UK. More time would be needed to explore these economics questions fully and it was decided that Gregory should speak directly to Tim Pettitt, Jon Knight (AHDB), and Jane Barbrook and Kelvin Hughes of APHA as experts in the nursery sector willing to offer advice.


WP3 Global Phytophthora risks to the UK – Beth Purse (CEH) and Mike Dunn (FR)

Beth Purse presented the WP3 work, starting off by reminding the project team of the WP objectives. For Objective 1 (Risk of Introduction) the aims are to identify the most important trade and recreational pathways linking Phytophthora source regions to the UK, to model introduction risk based on transport networks, source and destination characteristics and to test links between introduction risk and traits. One milestone was to compile a global country-level database of records of occurrence/arrivals of Phytophthora. This milestone is now nearing completion with 17,371 country level Phytophthora records, and 1417 species x country combination records obtained from various sources. These state where possible the source/recipient country, year of first record (linked to arrival?) and invasion status. There are of course large differences among countries in national recording and biosecurity effort and this is reflected in the number of records per country. In the analyses the WP3 team are looking at pre-2000 records as the potential ‘source’ distribution of Phytophthoras and post-2000 records as Phytophthora ‘arrivals’.

Preliminary analyses have been carried out using post-2000 Phytophthora records per country as the response factor and a set of predictors including trade connectivity to pest source countries (based on total imports of live plants), biosecurity effort (amount of invasive species legislation per country) and surveillance effort (number of official pest records from IPPC). Live plant imports and surveillance effort together explained 59% of the deviance in the number of new Phytophthora species recorded in a country since 2000.

UK Phytophthora records have also been compiled from a range of sources to enable models to be developed that relate Phytophthora species frequency of interception and extent of onward spread to biological traits. To assist with these analyses Beth held a short focus session in which she asked the project team to list what they felt were the most important factors influencing Phytophthora establishment in a new location, and rate of onward spread. These were then passed back to the WP3 team.

Beth then moved on to talk about the WP3.2 work (risk of establishment and spread) in which the team have continued to build the database of global Phytophthora records. They now have 11407 records for 82 species from 38 countries. In collating these data they have been prioritising countries that are climatically similar to the UK. Their niche models will predict potential global impact of Phytophthora species and the risk of their establishment and spread in the UK, excluding those species only known to occur in soil or water (ie no known plant host), species with no known woody hosts and species with non-relevant woody hosts (ie hosts not important in the UK). Their analyses will have to account for biases in reporting effort as developed countries have much higher species reporting than non-developed countries. This will be done by mining the scientific literature, oomycete and fungal databases (including Genbank) and adjusting weighting of records for species in highly recorded areas.

The Phytophthora traits database was merged in June 2017 with a similar database compiled by researchers in Australia and New Zealand. It contains data on 179 species and will likely be maintained in the longer term by Scion in New Zealand. A phylogeny will also be included in the database at some point. A publication is planned which will use the traits database as a conceptual framework for linking biological traits with invasion success of Phytophthora. Questions to be asked include whether closely related species share similar values for traits or groups of traits, or have traits linked to invasion evolved independently in several places in the phylogeny? They will also consider strength of phylogenetic signal in traits. Beth then ran through some traits which they have hypothesised to affect invasion success (ie survival structures, thermal tolerance, sporangial features). To do this they are using an ITS-based phylogeny (from Treena Burgess in Australia) for 179 species as well as two recent multi-gene phylogenies in order to resolve deeper nodes. Beth showed a series of slides with results from analyses so far illustrating strength of phylogenetic signal from sporangial features, reproductive traits and temperature traits, as well as rate of trait diversification over time – the latter showing high within-clade disparity suggesting rapid diversification and independent evolution to share common traits among clades.

Other questions being asked in the analyses are ‘do thermal traits especially cold tolerance modulate invasion of Phytophthora into temperate regions? (ie are emerging infections at higher latitudes linked to cold tolerance?) and how traits co-vary with each other (trait syndromes) and how much of this is driven by phylogeny?  In terms of the global impact of Phytophthora the main question being asked is ‘can species traits explain the global impact of Phytophthora?. This analysis uses impact metrics including geographical extent (number of countries in which a species has been reported) and host range (number of known host families). These data are being compiled from various international databases. The team have also looked at whether trait syndromes outperform individual traits as predictors of global impact. Their findings so far suggest that root (and foliar) disease symptoms predict a broad host range, that trait syndromes are more ecologically informative about a species’ global impact than individual traits and that it might be possible to develop a traits based ‘early warning’ system for pathogens which have similar traits but no impact yet.

Beth finished up by running through the plans for this coming year which are to submit papers on (i) the traits database and phylogenetic analyses and (ii) linking traits and global impact. They will fine-tune their species-specific trade models, develop UK and European spread models and finalise the niche models. They plan to finalise the model outputs with policymakers and practitioners in order to develop tools that people want, ie interactive source maps and lists of Phytophthora spp. associated with key forestry species.

Mike Dunn presented on work done so far looking at tourism and recreation as a pest and disease pathway, using methods such as literature review, visitor data from susceptible tourist attractions and by conducting a survey of internationally based plant pathologists. The literature review has found numerous studies linking spread of a range of invasive organisms to tourism, including P. ramorum. A recent VisitBritain survey revealed 31 million people visited Britain over a 12 month period with a third of these visiting parks/gardens as a stated objective. Visitor data have also been collated from seven of the most popular parks and gardens in Britain, including weekly visitor data from Kew Gardens. The team are now going to look at tourism data to see if there are links between tourism and Phytophthora introductions in the UK. Sixty-one plant pathologists worldwide have responded to a survey which included among the questions whether they considered international tourism to be a potential pathway for invasive pests and diseases. Fifty-six of the respondents said yes to this question, which is backed up by data from several studies. When asked to rank levels of perceived threat of bringing in pests and disease, imported plants and trees was ranked much higher than incoming international tourists.


WP4 Predicting risk via analysis of Phytophthora genome evolution – Ewan Mollison (University of Edinburgh)

Ewan presented on the WP4 work done so far (since the WP started in August 2017) and began by outlining what can drive the evolution of a pathogen through intrinsic factors (ie duplication, rearrangement, insertion, deletion of DNA regions) and extrinsic factors such as hybridisation between species and transfer of genes between species. He ran through the aims of the work which are (i) to compare genes from available sequenced Phytophthora genomes in order to identify a core set of Phytophthora genes common to all species as well as species-specific genes, (ii) sequence the genomes of three less damaging species which are closely related to highly damaging species so that genes involved in virulence might be identified and (iii) study target genes and gene families known to be important for virulence to identify how variations in these genes change pathogen behaviour such as host range and pathogenicity.

Ewan illustrated the sequencing and assembly strategy for P. austrocedri as an example of how a genome assembly can be improved. He found that 49% of the P. austrocedri genome consists of repetitive DNA. At this point Ewan was asked whether it would be possible to correlate species with high levels of repeat content with host range of the species – the answer was yes, it might be worth looking for associations with genome size and biological traits although generally for other species genome size correlates with nothing! 

Genome assemblies are now available for 26 Phytophthora species, all in varying stages of ‘finished-ness’. Most genomes are released along with predicted genes and protein sequences. Ten genomes have been released purely as scaffolded assemblies and gene prediction will need to be carried out on these. Ewan presented a graph showing large variation in the number of predicted proteins (over 30 amino acids in length) among currently available genomes, with the caveat that the protein data are likely to be an over-prediction and that the gene-prediction tools will need to be refined. A graph showing assembled genome size versus repeat content also illustrated the high levels of repeat content in certain genomes such as P. alni (hybrid), P. cambivora (putative hybrid) and P. infestans. Larger genomes are often a result of expansion of repeat regions, with these repeat regions often evolving rapidly which is very useful for overcoming host resistance.

Ewan assessed the completeness of coverage of each of the 26 Phytophthora genomes by looking for the presence of 234 ubiquitous genes expected to be present in all species. His findings showed that many genes may be missing from an assembly. For example 22/26 genomes were estimated to be 90% ‘complete’, 3/26 over 70% complete and the P. alni genome only 37% complete. Ewan then looked for orthologous genes present in all 26 genomes and found 2,107 genes or gene clusters common to all genomes. A phylogenetic tree inferred from these ‘core’ genes split some of the species from the same clade (ie species in clades 1, 3 and 8) however much more work is needed to refine this analysis.

Ewan then described some analyses done using an example gene family; the xylanases (xyn), which are cell wall degrading enzymes which specifically target hemicellulose, an important constituent of plant cell walls. There are four major xylanases which have been identified in Phytophthora; xyn1, xyn2, xyn3 and xyn4. Sequences from all xyn genes found in the 26 Phytophthora genomes were aligned and phylogenetic trees constructed for each gene. The xyn1 and xyn2 genes grouped into two distinct clades whereas the clades were less clearly defined for xyn3 and xyn4. Looking at the presence/absence of each gene among the 26 Phytophthora genomes revealed that not all species contain all four xyn genes which may be due in part to incomplete assemblies, and also that xyn sequence differences occur in some species which results in their falling outside the expected clade groupings for that particular gene.

Ewan finished his presentation with a cautionary tale; strangely anomalous xyn gene sequences were downloaded for P. taxon totara because the source database had been mistakenly linked to a P. kernoviae genome download. Always check the source!

Further WP4 work will involve sequencing P. obscura, P. foliorum and P. europaea to add into the analyses, together with any newly available Phytophthora genome assemblies. The xylanase gene family analysis will be expanded and other gene families of interest will be investigated ie RXLR effector proteins.

WP5 Synthesis and integration – Sarah Green (FR)

Sarah Green rounded off the meeting with a short overview of WP5 activities since October 2017, including Board meetings and the reports/research summaries recently posted on the project website as well as the successful uploads of all project outputs and outcomes to ‘Researchfish’. There was some discussion on stakeholder engagement activities planned for the coming year. It was decided not to attend the National Plant Show this year but rather to wait until next year to present project results. There was agreement for the need to engage more formally with Defra and the HTA over the pilot assurance scheme to ensure that the Phyto-threats project data can help to shape this scheme. The next stakeholder workshop will have the theme of ‘securing resilient outcomes – scoping the potential of an accreditation scheme and building a framework for its continued development’. It was thought that delaying the workshop until spring next year in order to have a more complete set of results to present would not fit in well with the nursery timetable – spring is very busy. So a decision will be made soon on whether to stick to the same format of holding the project team meeting followed by stakeholder workshop over two days in October. This will be decided at the next Board meeting.