Tag: science

Weeds are a major challenge in agriculture: competing for critical resources and killing crops. While herbicides have been pivotal in weed control in the last fifty years, their overuse has led to herbicide resistance in 273^[1] species globally, with annual losses to herbicide resistant blackgrass in the UK estimated to be $500million^[2] , and annual weed losses of $3.3billion in Australia and $11billion in the USA^[3].

As the world’s population grows and demand for food increases, so too does the need for new herbicide modes of action that can overcome these resistant “superweeds” and help farmers protect their crops.

At Moa, new mode of action herbicide candidates are discovered using high-throughput screening of chemical libraries in GALAXY followed by extensive profiling in glasshouse. Profiling of chemistries in the glasshouse is not only labour-intensive but can also be subjective, relying on the expertise and the judgment of the person analysing the growth and damage to plants.

Through the Knowledge Transfer Partnership (KTP) program sponsored by UKRI and in collaboration with the Department of Computer Science in Loughborough University, we have been harnessing deep learning to automate chemical library phytotoxicity testing process. Specifically, I  have developed an advanced AI-driven solution to analyse glasshouse imagery, improving the accuracy, efficiency and scalability of herbicide activity assessment.

To achieve this, our collaboration started by collecting image data from our glasshouse testing, capturing images of control and treated plant samples across multiple weed species. Each image was then manually annotated to specify the plant species and identify visible leaf matter, creating an adequate dataset for deep learning model training.

We then used these annotations to build object localisation models, enabling automated identification of plant species and precise detection of plant matter in both control and treated samples. By analysing these paired images contrastively, our approach quantifies relative growth differences, allowing a statistical estimation of each compound’s efficacy with improved speed and stability. This process was iteratively refined to optimise model performance by expanding the dataset, including more species, and analysing relationships between model predictions and manual human evaluations.

This new AI tool scored an impressive 95% accuracy when evaluated against real glasshouse data, demonstrating robustness across different levels of herbicide activity and species.

Our AI tool is now being integrated into Moa’s glasshouse pipelines, to give scientists in a standardised, scalable, reproducible process, reducing manual labour and improving the overall quality of assessment.

What next?

AI cannot find and develop new, safe and effective herbicides all on its own, out of nowhere.  It needs data: the higher quality the better.  Moa has a huge bank of data, from the 750,000+ compounds our platforms have already screened. So  now we are starting to use this new tool to help clean up and interrogate our data: re-analysing archived experiments, identifying potential misclassifications and recovering previously overlooked compounds.  This will help us – and AI – to make better decisions.

This project was made possible by a Knowledge Transfer Partnership between Moa Technology and the Department of Computer Science at Loughborough University. The author would like to thank Professor Baihua Li, Dr Haibin Cai, and Professor Qingqing Meng at Loughborough and Dr Laura Davies and Dr Nasir Rajabi at Moa Technology for their collaboration on the project. 

The results of the collaboration will be presented at VISAPP 2025 in Porto, Portugal and have been submitted for publication in Machine Learning with Applications journal.

^[1] Heap, I. (2024). The International Herbicide-Resistant Weed Database. Retrieved from https://www.weedscience.org.

^[2] Varah, A. et al. (2020). The costs of human-induced evolution in agriculture. Nat. Sustain., 3(1), 63–71.

^[3] Chauhan, B. S. (2020). Grand challenges in weed management. Frontiers in Agronomy, 1, 3.

GALAXY is our proprietary high-throughput screening platform at Moa: the first step in how we find new herbicides that will help farmers protect their harvests from herbicide-resistant weeds.  With GALAXY, we have been able to screen 750,000 compounds and identify over 70 new mode of action areas in the last three years: a massive industry breakthrough in speed, reliability and accuracy.

As one of the scientists working in the GALAXY team, I like to think I may be the first person ever to see the next big herbicide that will help to protect the world’s crops in a few years’ time.

But why me?

There are many parts to our process, but the team I work in is involved right at the start.  We treat special miniaturised whole plants (grown in our own labs) with thousands of different synthetic and naturally occurring compounds, and then record images of them. We compare how the miniaturised plants treated with our compounds appear in the images – a response we call a phenotype – to images of plants treated with commercial herbicides. If they look different, we know there’s a good chance we’ve found a compound that acts in a novel way and it will move on to our other proprietary platforms to be investigated further.

Our process is based on recognising known modes of action in images we take of the plants. To obtain these crucial images of our test plants, we use a technique called high-content fluorescence microscopy. When we shine red light on chlorophyll, it absorbs light then re-emits it, fluorescing at a different wavelength. We can filter out the wavelengths of light that we shone onto the plant, and only image the wavelengths that chlorophyll emits. Since chlorophyll is naturally fluorescent, we don’t need to genetically modify our plants or use dyes to see chlorophyll and therefore the chloroplasts. However, we still have the option of using dyes and reporter lines to study other biological processes, should we wish to.

How accurate is this method?  Our team has been tested in our capability to decipher mode of action compounds blindly.  In these tests, we were given compounds with known mode of actions to screen in GALAXY, without being told what they are or how they work.  Using only our phenotypic assay, we  scored about 96% accuracy in those blinded tests, and I would like to think we have since improved as we have become more and more experienced with this process!

Moa Technology is pleased to announce that Lee Sweetlove, Professor of Plant Sciences at the University of Oxford, has joined our team.  Professor Sweetlove will be working to further understanding of the novel herbicidal modes of action discovered by Moa, which we are currently developing into safe, effective and affordable solutions to protect harvests from crop-killing weeds.

Professor Sweetlove has over 30 years of experience working in plant science and is internationally recognised for his work in deciphering fluxes in plant metabolic networks.  His expertise and network will be invaluable to Moa in helping prioritise and deliver our innovative plant protection solutions.

Spun out of Oxford University in 2017, Moa has developed proprietary technology to discover a whole new generation of herbicidal modes of action to provide farmers with modern crop protection solutions which break weed resistance.  In the last three years, Moa’s GALAXY platform has already screened over 750,000 compounds and discovered over 70 promising novel herbicidal mode of action areas.  In July 2024, Moa signed a major partnership with Nufarm to co-develop and commercialise one of these new chemical series.

Comments Professor Sweetlove:

“Working with Moa is an ideal opportunity for me to put my many years of experience of plant biochemistry and metabolism research to direct practical use.  I am very excited to be working with this world-class scientific team on deepening our understanding of the dozens of novel herbicidal modes of action already discovered by the company, and helping progress these solutions as quickly as possible to the field.”

Dr Virginia Corless, Chief Executive of Moa Technology, adds:

“As we grow ever closer to bringing a new generation of modern, safe, effective herbicides to farmers’ fields, we are so pleased to welcome Professor Lee Sweetlove to our team.  From the outset, Moa has combined the most innovative academic minds in the world with the best commercial talent in the industry to address the urgent global problem of weed resistance.  Following our appointments in the last year of Hartmut van Lengerich, former SVP at Bayer Global Crop Protection, as Chair of the Moa Board, and Apostolos Klontzaris, formerly of Bayer Crop Science and Syngenta Crop Protection, as Chief Business Officer, Lee’s appointment will help ensure our expertise at the leading edge of weed biology remains second to none.”

With herbicide resistance a key threat to crop production across the world, we at Moa are working to develop herbicides with new modes of action that break this resistance. Worldwide, herbicide resistance has been reported in 273 different species, across 23 of the 26 known herbicidal modes of action [1].

To achieve our mission to break resistance we need to test our new herbicides on resistant weeds –  and one of the worst species is right on our doorstep.

Blackgrass (Alopecurus myosuroides) is one of the toughest weeds in the UK and is resistant to multiple modes of action [1]. Annually, it causes £400million in lost profit to UK farmers [2] and can be so bad that in some cases farmers resort to crop destruction to prevent greater weed problems in following years [3].

Different blackgrass populations can have diverse levels of herbicide resistance to different herbicide modes of action, so it was important that our team collects samples from multiple populations. Resistance can be at the target site, where a mutation in the plant leads to the target protein no longer being sensitive to the herbicide, or non-target site, where in the case of blackgrass enhanced metabolism leads to the breakdown of a herbicide before it reaches its target.

Blackgrass seeds are ready to collect in early-mid July, so our team of Moa scientists including myself, Alice Hannam, Elisabeth Bowles, Tanya Mathie, Tom Clarke, Harry Wheelwright, Mark Bartlett, and Silvia Rubio-Diaz was standing by to make collections from farms in Oxfordshire, Wiltshire, Cambridgeshire, Lincolnshire, and Yorkshire.

The process starts with contacting local farmers to get permission to access their land and find the most infested fields – the more blackgrass there is, the more likely it is to be resistant. Fortunately (or unfortunately), most farmers are motivated and engaged in the fight against blackgrass and are happy to help.  Some even come out to the field with us!

Blackgrass is easy to identify in the field, with the purple-green seed heads standing tall above cereal crops [4]. The seeds are ready to collect when around 25% of the seeds have naturally fallen off the seedhead and the remaining seeds come away easily. One blackgrass seed head can have as many as 100 seeds, and there can be as many as 500 seed heads per m², meaning our team had no problems collecting large amounts of seed [5].

In July, we collected over 20 different populations of blackgrass but did not stop there, as blackgrass is not the only resistant weed in the UK. While out collecting we also took the opportunity to collect samples of Italian ryegrass (Lolium multiflorum), wild oats (Avena fatua), rye brome (Bromus secalinus), sterile brome (Bromus sterilis), great brome (Bromus diandrus), and groundsel (Senecio vulgaris).

The next steps involve cleaning the seeds to remove debris, breaking dormancy to ensure we get good germination, and testing our collections against a wide range of commercial herbicides to determine their resistance status.   The collections also help us refine our own novel herbicides to make sure they can control the toughest weeds that farmers face.

And of course, herbicide resistance is not just a UK problem, but a global one.   While it’s relatively straightforward for us to collect British weeds, it’s also critically important to understand how herbicidal resistance is increasing in other parts of the world, so we are engaging with international scientists to expand our collections and testing capabilities.

[1] International Herbicide-Resistant Weed Database

[2] The costs of human-induced evolution in an agricultural system

[3] Get timing right for blackgrass destruction

[4] Arable weeds: Basic black-grass biology

[5] Black-grass (Alopecurus myosuroides)