“I was devastated when I realised the disease had arrived in the UK because I had heard about how ash dieback had affected parts of Europe,” says Dr Anne Edwards of the John Innes Centre (JIC), an institute which receives strategic funding from BBSRC, based on the Norwich Research Campus.
From the moment the ash dieback pathogen arrived in the UK, Edwards and colleagues have been first over the top in attempts to characterise and develop a solution to the problem. Her volunteer work in a local forest led to Edwards being one of the first people to identify Chalara fraxinea, the fungal pathogen that infects ash trees, in the wild in the UK.
Realising the scale of the problem and the need to act quickly, JIC and others utilised quick-response BBSRC funding and collaborated with a wide variety of scientists from around the world through the OpenAshDieBack project – an ‘open source’ platform for scientists to share their data and publish results more quickly than publishing first in traditional journals (ref 1).
Funding on this project has led to work on the fundamental bioscience of the host-pathogen relationship and DNA sequencing of the ash tree and its nemesis Chalara, focusing on a Danish ash tree that shows low susceptibility to infection. Novel research technologies have also been developed, such as a puzzle game hosted on Facebook called Fraxinus that utilises the game-playing skills of the public to help analyse some of the huge amounts of genetic data.
The game aims to identify useful regions of the ash tree genome and the fungus that vary from individual to individual and sample to sample. These valuable tiny inputs then will feed into research trying to identify such things as the tree genes responsible for conferring reduced susceptibility to the fungus that could be used in future breeding programmes.
A walk in the woods
The arrival of ash dieback in the UK in 2012 appeared sudden, but was not a surprise to everyone. The disease had long been known in Europe where it had devastated ash tree populations from 1992 (ref 2). Edwards has worked as a volunteer in the Norfolk Wildlife Trust reserve Ashwellthorpe Lower Wood, near Norwich, for 20 years and says she was on the alert for the disease since 2011.
“We even did a survey of the ash trees in our village, Hethersett, that year as part of a larger biodiversity project,” she says. “I had heard about the disease because I am Tree Warden for our village and we get regular tree-related bulletins from the Tree Council.”
So when news filtered through of ash dieback’s arrival – probably imported from a tree nursery in the Netherlands – Edwards and her forest group friends were on the lookout. Finding ash trees in her own familiar grounds showing signs of infection was the sight she’d been dreading. After the Norfolk Wildlife Trust had contacted Defra and the Forestry Commission, she took some samples back to her lab for DNA analysis.
“I was probably the first to identify it from the wild using DNA,” she says, adding modestly that the Wildlife Trust and a number of other volunteers were aware very early on that there was a problem with the ash trees in Ashwellthorpe.
The positive results for the infectious agent Chalara from Edwards’ samples posed the question: what could be done about it?
“We deduced that the John Innes Centre was probably in the centre of an epidemic and we decided to try to see what we could possibly do,” says Professor Allan Downie, also from JIC who heads up the research group that Edwards works in. “Ultimately that led to joint discussions with colleagues and a research proposal.” He also identified and isolated the pathogen in his local village wood, and C. fraxinea has now been confirmed at 500 UK sites and is now treated as a quarantine pest under national emergency measures.
Those discussions led to the JIC team co-founding the Nornex consortium to tackle the problem. However, JIC usually focuses on food-producing plants, rather than trees, but Downie says they came at it from the perspective of genetics, and that long-term control of the disease had to involve selecting trees that are tolerant. “The key things we thought we could bring were genetics and genomics techniques,” he says. “So we set out to generate appropriate tools that could help others select for tolerance.”
Downie’s lab has been working for many years on plant-microbe interactions (rhizobial symbiosis with legumes) and so he was familiar with how the diversity of natural variation could influence plant-fungus interactions. “It seemed likely that the same could be true of the interaction between the fungal pathogen and the ash trees.”
The lab now plans to collaborate with Erik Kjaer’s group from the University of Copenhagen’s Forest & Landscape department, Denmark, where ash dieback has infected 60-90% of ash woodland. But there are good grounds to be optimistic: a tree has been identified, known as ‘Tree 35’, which is one of two trees identified on the Danish island of Zealand with high levels of resistance.
“What was particularly critical was to establish the collaboration with Professor Erik Kjaer’s team who were willing to collaborate using their panel of tolerant trees and their segregating population generated from a single tolerant tree,” says Downie.
The tree from Denmark that displays resistance is a useful lead, but there is no guarantee that it can be crossed successfully with a UK tree to make a resistant variety. It also makes sense to scour UK ash tree genomes for similar regions of DNA that might confer resistance, or other useful attributes. However, a problem with handling vast amounts of genomics data is the processing time it takes to annotate the genome – that is identifying where the genes are and what they do – and that huge amounts of data go unused.
Dr Dan MacLean grapples with these problems as a bioinformatician at The Sainsbury Laboratory, which is also based on the Norwich Research Campus with JIC. “When it became clear we had data that would likely not get used in some of the downstream applications, because we lacked the accurate tools to properly validate that data, we felt we should try and develop something that would allow lots of people to give us a hand,” he explains. “And the best vehicle for that seemed to be a game.”
In recent years, various games and applications have appeared to use computer and human power to solve scientific problems. An early example was the SETI screensaver that used a computer’s idle time (when you’re away making a cup of tea) to crunch information from the search for extra-terrestrials. The next-generation applications, sometimes called ‘serious games’ use real data and turn it into a game, and the results feedback to the scientists who can use it in further analyses (ref 3). Examples include Fold.it, a puzzle game that advances knowledge about protein structures and how to target them with drugs (ref 4). Zooniverse, too, contains myriad such ‘crowdsourcing’ scientific projects where many small efforts from a sufficiently large crowd can outpace the efforts of a single computer (ref 5).
“The concepts of Fraxinus emerged in conversations and brainstorming, says MacLean, adding that he’d love to play more serious games if he had the chance. “I have dabbled with Phylo and am very interested in a new and upcoming one called Dizeez that’s very prototypical and is about identifying gene function.”
Fraxinus presents players with reference DNA sequences from the ash tree genome (ref 6). Players are challenged to match up multiple DNA sequence reads from other samples against the references, making as close a match as possible, by shifting base pairs of DNA. The differences between the references and the reads are what scientists are interested in – ultimately these differences might prove to be in regions of the genome that confer useful characteristics such as resistance.
Published research has demonstrated crowd computing techniques can be successfully used to help improve the accuracy of this process, which is known as multiple sequence alignment (MSA). This method of comparing stretches of DNA (or amino acids) lies at the heart of comparative genomics – examining how evolutionary similarities and differences in genomes scale up to comparable characteristics in organisms. In 350,000 solutions submitted from more than 12,000 registered users of the Phylo game, solutions from game-players helped improve the accuracy of up to 70% of the alignment blocks considered (ref 7). Because these games can be played with no biological training, some consider crowdsourcing or ‘citizen science’ approaches will exploit billions of “human-brain peta-flops” to solve problems in the future (among other things).
And MacLean says that while computers can still do the easy but laborious tasks, with Fraxinus it’s the skill and guile of the human eye that’s being used. “Humans can really add value on top of the computer programs,” says MacLean. “The really hard stuff that would benefit from insight and inspiration can be done by humans if we can find a way to wrap them up in interesting and useable game-like interfaces.”
It’s an approach that, so far, appears popular. More than 10,000 people played the game in the first week of its release in August 2013, and over a two-week period there were 32,267 total visits from 18,100 unique visitors – meaning over 43 % of visits are from repeat visitors. The average visit duration was 18 minutes 16 seconds, a number that continues to rise showing a core of dedicated, repeat players.
“Each of 10,000 initial datasets have had at least some sort of play and some have been ‘saturated’ so that no more improvements seem possible, so were retired from the game,” MacLean explains. “An immense response really. We didn’t expect to have this sort of reaction.”
Fraxinus and the future
MacLean is now undertaking analyses and looking at incorporating the results of the game into the studies on genetic variation, and looking at new bioinformatics approaches for identifying key genetic variants in plants without draft genomes, but in other information is available from genetic crosses.
Edwards now splits her time between her original work with Downie and new work on legumes, as well as continuing to help with the ash dieback research by collecting material for the Nornex consortium (see below) to work on.
“I will continue to monitor the trees in the wood and will be looking intently for any that seem to be holding the disease at bay” says Edwards. “The infection is so bad now that any trees that are still looking healthy next year might have a degree of tolerance. We can only hope.”