Lateral-flow-test
The global hunger index shows 155 million people are affected by hunger in the world in 2021. That are 20 million more compared to the year 20191. One aspect of this growing issue is the harvest losses that occur due to plant stressors. Plant stressors can be roughly divided into biotic and abiotic ones. While abiotic stress is provoked through non-living factors, like drought or soil composition, biotic stress is caused by living factors, such as animals, other plants or pathogens.
While there are hardly any treatments or ways to eliminate abiotic stressors, there are promising methods to cope with biotic stress. There are preventive and curative plant protection products, like fungicides or bacteriocides, against groups of pathogens that can keep crop losses low.
To be able to use these pesticides and preventative methods, one needs to know which pathogen is causing the infection.
To determine if a biotic or abiotic stressor is at hand, special drones can be used for large fields. With the help of multi-spectral cameras, stress can be identified before it causes visible symptoms on the plant2. Special algorithms can then distinguish between abiotic and biotic stressors. However, this technology is still far from being able to specifically identify the pathogen, and this information is needed for the induction of specific countermeasures.
With our test, we want to provide this type of information to possible customers like farmers.
In order to get aptamers for the Lateral-flow-assay, either proteins with pathogen-induced expression for protein-SELEX or cells of the pathogens themselves for cell-SELEX are needed. In the beginning we focused on protein-SELEX. To be able to show the whole process of making a LFA , we expressed our own proteins, although we were not able to actually use them in our SELEX process, due to time constraints.
Read more about our proof of concept here
Later on we changed from protein-SELEX to cell-SELEX as we got more information from our outreach work.
Our test could either be applied by non-experts (A & C) or by professionals that know the equipment (B) and we suggest it to be used in the following scenarios: Plant diseases are often brought into the field by seedlings obtained by plant breeders. Here, our test should offer the possibility to screen the young plants for the most frequently introduced pathogens. If these are plants that have high costs of cultivation and are expensive to purchase, it would make sense to pay a little more and do our test. In the long term, these tests could save the user time, effort, and money.
A. The test can be simply used by grinding up a small part of the plant and mixing it with a suitable buffer for the pathogen being tested for. A few drops of the mixture are then dropped onto the test and after a few minutes, the result is visible. If a whole field or a lot of plants need to be tested, either a sample pooling or a spot check with multiple different tests and samples can be done.
If a pathogen is detected and the test shows a positive result as two red lines, the farmer can immediately initiate plant protection measures and contain the pathogen.
B. Together with multispectral camera-assisted early detection, our test provides a second possible application. This is particularly suitable for large fields or large greenhouses. Drones or handheld devices equipped with multispectral cameras would be used to detect biotic stresses in the population even before visible symptoms occur. To identify the pathogen causing this biotic stress, our test would only be used on the stressed plants. Reducing the number of plants that need to be tested helps to reduce the costs. Stressed plants are the ones that show a reduction in their chlorophyll fluorescence. The chlorophyll fluorescence is an indicator for the photosynthetic rate and therefore an indicator for the plant’s health3. The photosynthetic rate drops before visible symptoms occur, hence why it is a great early detect system24.
We could purchase the equipment and offer the screening as a service. In our opinion a subscription service would be most sensible as the farmers could get regular testing, enabling us to detect pathogenic infections as early as possible, reducing overall crop losses.
The test is also applicable when the disease is already visible to the naked eye, but the symptoms could be caused by different pathogens. Some pathogens cause similar symptoms that can’t always be distinguished visually. In this case, the test would only target the specific pathogens in question. Our test would give clarity if the pathogen is for example a fungus or a virus and therefore would provide information on what treatment is appropriate. A single infected leaf or root as a sample would be enough for this scenario.
After identification and differentiation, a curative plant protection product specific for the detected pathogen can be considered.
C. The test is primarily intended for special crops where it is financially more important to prevent a loss of harvest. An example would be strawberries, which need a lot of time and money as an investment to get a good and profitable harvest. Investing some more money into the detection and differentiation of pathogens is worthwhile in these plants because their profit margin is bigger. The situation would be different with for example lettuce or wheat, where the investment is much lower and a lost harvest is not as detrimental. Because of that, the budget for treating these plants is lower, and in case of infection, these plants are often destroyed and replanted instead of treated. Safety aspects and challenges that we have to take into account are that our test method itself is very specific and therefore even small mutations that affect the structure and surface of the pathogen can lead to different tests being needed for different strains. But this is easily solvable. We can take the new pathogen and create a new test, which then could be on the market, the year after.
Overall our test is meant to be an adaptable system for early pathogen detection in a wide range of crops.
Statement of costs
To be able to talk about the demand and usability of our test we made a rough estimate for the costs associated with one test. This was essential so we could get a more realistic view on the uses of our test from stakeholders.
Here we listed the products that are needed for the preparation of the test itself. Not included are products needed for SELEX and lab equipment, as these are one-time purchases and cannot be easily split per test.
Ampelpflanze
As we already stated, the need for new solutions and improvement of agriculture is urgent and present. Since our initial idea is to detect different kinds of stress, we thought of a side project that could visibly demonstrate the need for our test. This project aimed to create a plant that changes color according to the stress it’s influenced by. The name "Ampelpflanze" is German and translates to "traffic light plant". The plant was meant to serve as a biological pathogen reporter system that turns yellow when it is facing drought stress, red if a pathogen is infecting it, and stays green if none of the two stressors are present.
This plant reporter has a lot of possible applications! We want it to serve as a control plant for daily laboratory usage where it can warn the user of stressors, e.g. indicating positive transformation events.
It could serve farmers as an indicator that the plants need treatment and, owing to the set color palette, even show them what kind of treatment is necessary.
In the current legislative state in Germany, the usage of GMOs is strictly regulated, so it could not be used in-field here. Still, in safe laboratory environments, "Ampelpflanze" can be used to study pathogen interactions even further and optimize the use of pesticides and water.
Learn more about our Proof of Concept
Pathogen Forecast
Proplant.de is a website which occasionally publishes pathogen reports. Those reports are helpful for farmers, because they can prepare for pathogens before they arrive.
We would like to build upon this idea and push it even further. The following information are helpful to build such a system.
Visualization
For modelling, we tried two different approaches to support our project. One of them was visualization of pathogen spread in a field - in order to help further research and also provide data for scientists to understand pathogens better.
First thing we tried was setting up KeplerGL.
However, after setting it up, we decided to temporarily switch to a custom engine and test simultaneous computation and visualization.
Initially, we just intended to use this prototype to calculate some possible scenarios and export them for keplerGL, which is way more advanced and beautiful. We approve of KeplerGL for a final app.
Visualization of in field data is crucial for research and user friendly apps/websites.
In-Silico SELEX
For modelling, we also tried to perform in-silico SELEX. In order to help progress our SELEX Team as fast as possible and make future findings of aptamer proteins easier and quicker, we decided to research and develop a machine learning model.
Before starting our own project, we first researched whether there were already existing, and promising, solutions we could build upon.
There were only two options for as far as we know.
Option 1: AptaNet
It sounds good on paper, even though some of our experts quickly spotted possible flaws. Ultimately, we couldn't reproduce their results in a timely manner.
Our GitHub fork with some fixes: Link
Option 2: Machine learning guided aptamer refinement and discovery
Paper: Link
Machine learning guided aptamer refinement and discovery was way too expensive and not feasible for us, even though it is probably the best paper we found in this regard.
Usually, complex problems require huge sets of data for an accurate model and unfortunately, there is close to no sufficient data for such a model to support SELEX.
Current data sources for SELEX
The only reliable source we found has very little useful data.
This means, that we probably need to start collection such data ourselves.
Data collection
In order to support future pathogen researches, we need to start collecting in field data as soon as possible. We suggest to develop an app, which makes it easy for laymen and scientist to sent us data of plants and pathogens. To make this app useful for it's users, we can provide them a "pathogen forecast light".
Pathogen Forecast "beta" App
Mainfunction
Prediction of pathogen spread across countries/fields.
Input
- Pathogentypes: Insects/moulds/virus/bacteria...
- Weather: dry/humid, hot/cold, windy...
- Field: Number of plants/crops and which kinds
- Initial spreads: where and when the pathogen spread started
- GPS data of fields and plants/crops
Output
- Map with pathogen spread
Calculation
- As soon as we have enough data: with a model
- Until then: with guesswork and formulas based on papers