Introduction

It is well established that crop nutrient deficiencies are a pressing problem in the agricultural landscape. Without elements crucial to synthesizing plant tissues, such as nitrogen, phosphorus and potassium, crop growth is stunted. In fact, plants’ uptake of one nutrient is interdependent on whether or not they have other nutrients - for example, calcium deficient plants are found to have a reduced uptake of elements like nitrogen and iron1. Additionally, nutrient deficiencies make plants more susceptible to disease2. Therefore it is evident that nutrient deficiencies limit crop yields, directly contributing to issues like world hunger. While some farmers experience reduced crop yields, many others choose to overapply fertilizers - according to a 2021 study, annually, only 35% of applied nitrogen is used by crops3; the rest runs off into natural environments, leading to problems like pollution and eutrophication.

With these considerations in mind, we saw the need for a diagnostic device for crop nutrient deficiencies. Not only would it be able to warn farmers if crops are nutrient deficient so that they can execute timely interventions, it would also allow farmers currently overapplying fertilizers to check if their crops are actually nutrient deficient before adding more, which is beneficial both financially and environmentally.

Before designing our product, we contacted various people with experience in agriculture to get a better picture of the current landscape of crop diagnostics.

Call with Georgina Bray

  • We spoke with Georgina Bray, manager of Cambridge-based Hope Farm.

Current diagnostic methods

  • Hope Farm usually monitors nutrient deficiencies judging from crop yields and visual symptoms; at the same time, they have sent samples off to labs for conducting more expensive tests, as described below.
  • Soil tests, where samples are taken across the farm, are done every five years, and total to a few thousand pounds each time. Even if these tests are refined and comprehensive, there is an unavoidable disconnect between soil and crop nutrient content. (For more on the limitations of soil testing, you can refer to our Stephanie Swabreck interview below)
  • They have sent plant tissue off to labs for testing, but only five samples at a time. They also started conducting grain nutrient tests recently, which cost 35 pounds a sample. She explained that this decision was made because they learnt that nutrient deficiencies increase crop susceptibility to diseases, as well as the interconnectedness of different nutrient deficiencies. In the interest of obtaining higher crop yields, she saw the need to obtain a more comprehensive picture of their nutritional states - not only testing if they are deficient in elements like nitrogen, phosphorus and potassium, but also trace elements like boron and manganese.

Key learnings for design

  • If a crop is deficient in a specific nutrient, other facets of its health are affected as well. This reinforces the importance of a crop nutrient deficiency diagnostic test.
  • Farmers are beginning to move beyond relying on crop yield and visual inspection alone for diagnosing nutrient deficiencies, instead valuing tests that can provide specific, holistic (i.e. able to test for different nutrient deficiencies) diagnoses of crop health.
  • There is a gap in the plant diagnostic market for a cheaper nutrient deficiency test, so that they could be conducted more frequently and on more samples.
  • Georgina mentioned that it would be useful for our diagnostic test to be applicable to different types of crops. Even if we cannot achieve this within the timeframe of our project, it would be good to come up with a modular system so that this can be accomplished easily.
  • For our diagnostic test to be comparable to other alternatives on the market, it would also be good for it to be at least semi-quantitative, so that farmers can identify their next steps (i.e. how much more fertilizer they should add).
  • While understanding that the results from our device would probably not be as refined as those coming out of a lab test, Georgina was intrigued by our proposal of a test that could be done on the field and provide same-day results. It appears that there is currently no product of this nature on the market.

A photo we took with Georgina at the end of our call.

Call with Stephanie Swarbreck

  • Stephanie is a group Leader in Crop Molecular Physiology at Cambridge.

Current diagnostic methods

  • Stephanie provided various examples of onsite tests for nutrient deficiencies that we had not previously heard of. She first talked about handheld devices for measuring chlorophyll fluorescence, where farmers could just clamp a leaf and use the readout as a proxy for crop health. Although their cost is in the hundreds (of pounds), it can be used to conduct an unlimited number of tests, so the cost per test is very low. In less developed countries, leaf colour charts are used instead, operating on the same principle (the greener the leaves, the higher the concentration of chlorophyll, the healthier the crops). While these tests are very convenient to conduct and cheap, they are not specific in their diagnoses - even if they indicate that there is a drop in crop productivity, farmers would not even know if it is really due to a nutrient deficiency, disease or other problems.

Potential influence on policy

  • Stephanie also told us about RB209, a nutrient management guide provided by the UK government to industrial farmers. It provides information on how to work out levels of different nutrients to apply to different crops, based on information like soil type and rainfall level. However, she acknowledges the unavoidable disconnect between nutrient content in soil and how much is actually accessible to crops, for reasons such as individual differences in plants’ ability to take up nutrients, mineralization of certain elements with changes in seasons, or nutrients being bound to soil particles. This leads to variation across fields in different regions, or even within the same field.
  • If we can create a crop nutrient deficiency test that can be done inexpensively and conveniently, we believe it would act as a valuable tool to complement this guide. Farmers can first obtain a rough idea of how much fertilizer to apply with RB209, and later fine-tune this amount based on the readout of our diagnostic test. Our device would help farmers account for differences that the guide cannot, taking a further step in maximizing crop productivity originally limited by nutrient deficiencies.

Key learnings for design

  • There is a demand for a crop nutrient deficiency test that is specific. This is lacking in current tests, but also resources provided by the government (at least within the UK).
  • It is important for our diagnostic test to be sufficiently cheap and simple to conduct, such that farmers are willing to use it instead of the inexpensive and convenient alternatives on the market. (Though it should be acceptable for the test to be slightly more complicated than the aforementioned methods if it has the advantage of specificity)
  • She points out however, with the example of fertilizer being applied to winter wheat three times a year only, that nutrient deficiency tests won’t be conducted too often. Therefore it would be acceptable for our protocol to take a bit of time and effort to conduct.
  • Stephanie also mentioned that semi-quantitativeness would be a good quality for our diagnostic test to have. Readouts do not have to be too precise - as long as farmers have a rough idea of the level of deficiency that their crops are experiencing, and hence know what next steps to take.

A photo we took with Stephanie at the end of our call.

Call with Paul Hill

  • Paul is a British smallholder farmer.

Current diagnostic methods

  • Paul mainly conducts visual checks of symptoms to check for nutrient deficiencies in his crops.

Potential influence on policy

  • Paul brought up incentives set up by the UK government for farmers to optimize nutrient management, for example grants CNUM14 and CSAM15. These require farmers to test their soil once per year and add fertilizers accordingly, such that soil nutrient content just matches crop demand.
  • If we can produce a test that is inexpensive and convenient enough to conduct, it could potentially be used to replace soil tests in these initiatives. This would allow the government to request more frequent testing, as well as providing a more direct readout of plant health. Therefore, as a whole, this would increase the effectiveness of these initiatives in matching nutrient demand with application.

Key learnings for design

  • Some smallholder farmers are still mainly relying on visual symptoms to diagnose crop nutrient deficiencies. This is likely to be because of the costs of lab tests - hence a cheap test would allow smallholder farmers to access precise, preemptive (as opposed to waiting for symptoms to appear) diagnostic methods as well, reducing inequalities in the agricultural industry.
  • Paul believes that faster feedback times compared to the other currently done tests would be an attractive feature for our product to have.
  • As costs for agricultural resources have “gone through the roof” in recent years, he believes that farmers would be willing to use nutrient deficiency tests to ensure that they are adding just enough fertilizer, especially if these tests could be done cheaply and conveniently.

Incorporating learnings into our design

With the suggestions collected from stakeholders and experts in mind, we were able to identify a few key features to definitely include in our product. It was clear that our product needed to be specific, cheap, user-friendly and quick to be a valuable and competitive alternative to current methods for testing for crop nutrient deficiencies. We adapted our design to ensure all these specifications were met.

Specificity

Due to the nature of the detection methods we chose, our diagnostic test is inherently specific. Our methods function on the principle of diagnosing using specific miRNAs as proxies for nutrient deficiencies. As a particular miRNA would only be upregulated to a detectable extent if a specific nutrient was deficient (e.g. miR399f is only upregulated during phosphate deficiency), if our device gives a positive readout, it would definitively indicate that the sample was deficient in the nutrient of interest. Therefore, our diagnostic test would go beyond telling users that their crops are not at optimal fitness, informing them of the root cause of this problem so that they can solve it more effectively.

Cost

Cost was one of our major considerations in choosing our detection methods - as we wanted to ensure that our kit would be accessible to smallholder farmers who may not have the funds to send soil or plant tissue samples off to labs for expensive tests, reducing inequalities in the agricultural industry. By maintaining a low cost, our test could potentially be a valuable tool to farmers with limited resources in developing countries as well.

  • In the extraction protocol, we tested reducing the amount of RNAse-OUT compared to manufacturers’ recommendations as it was one of the most expensive reagents in the procedure. We successfully verified that it is still possible to get a decent RNA yield if we do so, while cutting costs by nearly 50%.
  • We initially considered using the detection strategy, SHINE. It looked promising to us as the paper 6 that presented this method reported a limited of detection of 10 cp/μL, and 100% specificity using a fluorescent readout. However, we realized that the SHINE master mix is quite expensive, and would make our diagnostic test unaffordable to our target audience. Hence we decided to turn to other methods.
  • Users do not have to purchase expensive, specialized digital equipment to obtain the results for our assays. Instead, they can simply obtain a readout using an app on their smartphones, and a 3D-printed dark box that comes with our kit.
  • In the end, we estimated that our kit costs only 8 pounds (hybridization) / 22.6 pounds (RCA/G-quadruplex) per test. Upon presenting this figure to Georgina of Hope Farm, she expressed that it sounded like a reasonable figure, especially if the test could be done on the field and provide same-day results.

User-friendliness

User-friendliness was evidently also one of the major things we thought about while designing our diagnostic test. As we hoped that farmers could conduct the test by themselves so that they could get same-day results, instead of having to send results off to a lab and wait for days, we tried to develop a protocol that is as simple as possible.

  • All reagents would come in premeasured quantities to minimize the measurement and transfer steps that farmers have to perform.
  • Typical laboratory methods for RNA extraction involve pipettes, centrifuges and toxic reagents like Trizol - we replaced these with syringes, a hand-centrifuge and isopropanol, which is non-toxic.
  • To ensure full understanding of the steps they have to conduct in extraction, we filmed instructional videos for farmers and wrote up a detailed protocol for them to follow. The protocol was checked by a team member who had no prior involvement with extraction to ensure that wordings are easily understandable.
  • We purposefully chose to develop detection methods that could provide results quickly, and involve minimal steps for users to conduct. In our RCA/G-quadruplex method, users would simply have to add the solution obtained from extraction to a paper disc, followed by a 2 hour incubation step at room temperature. Meanwhile, in our hybridization method, the incubation steps are also at room temperature and shorter, totalling to under 1.5 hours, but some washing steps are involved. For details of steps users have to follow for each method, please refer to the project description page.

Speed

Lastly, we took the feasibility of transport and storage into account while designing our diagnostic test, to maximise its accessibility to potential users.

  • By lyophilizing components of the kit containing oligonucleotide probes, they are made stabler and suitable for transport and storage at room temperature. We were able to verify that our hybridization method, and the RCA component of our RCA/G-quadruplex method, are still viable after lyophilization.

Future directions

We could not finish developing our project due to time constraints. Therefore, we recognize that there is room for us to improve in meeting the needs of stakeholders. Key features that we hope to incorporate, if given the opportunity to develop our product further, are as follows -

  • It is extremely easy to modify our device so that it would detect other miRNAs, as this can be done just by swapping out the miRNA binding site on our probes (for both hybridization and RCA/G-quadruplex) to the sequence complementary to a new miRNA of interest. The modularity of our device means we can adapt it to detect miRNAs upregulated during deficiencies of different nutrients, or deficiencies in different crop species. This is further facilitated by our gene ontology database, which would allow us to identify the appropriate miRNA to target based on the nutrient and crop of interest with ease. It could even be used to detect viral RNA, and hence be expanded to crop disease diagnostics.
  • We want to develop our diagnostic test such that it can produce a semi-quantitative readout. We have produced preliminary proof that this is possible - for our RCA/G-quadruplex method, we were able to verify that the level of ThT fluorescence varies with the concentration of G-quadruplexes in solution. We initially considered simply displaying how deficient in the nutrient of interest the crop is on the app. But upon speaking with Georgina of Hope Farm, she suggested that it would be even better for us to explicitly outline next steps to farmers through our app, for example providing an estimate of how much fertilizer they should add. A semi-quantitative readout would also allow users to track improvements in the nutritional state of their crops, to see if the actions they take are in the right direction.

Footnotes

  1. Wang, Y., Yang, H., Wei, L., Liang, H., Huang, X., Tang, J., Feng, D., & Li, X. (2023). The Effects of Calcium and Magnesium Deficiency on the Growth, Physiology, and Nutrient Uptake Characteristics of Walnut Trees. Journal of Biobased Materials and Bioenergy, 17(5), 547–555. https://doi.org/10.1166/jbmb.2023.2305 ↩

  2. Marschner, P. (2012). Mineral nutrition of higher plants. Academic Press. ↩

  3. Ritchie, H. (2021, September 7). Excess fertilizer use: Which countries cause environmental damage by overapplying fertilizers? Our World in Data; Global Change Data Lab. https://ourworldindata.org/excess-fertilizer ↩

  4. Department for Environment, Food & Rural Affairs & Rural Payments Agency. (2024, May 21). CNUM1: Assess nutrient management and produce a review report. GOV.UK. https://www.gov.uk/find-funding-for-land-or-farms/cnum1-assess-nutrient-management-and-produce-a-review-report ↩

  5. Department for Environment, Food & Rural Affairs & Rural Payments Agency. (2024, May 21). CSAM1: Assess soil, produce a soil management plan and test soil organic matter. GOV.UK. https://www.gov.uk/find-funding-for-land-or-farms/csam1-assess-soil-produce-a-soil-management-plan-and-test-soil-organic-matter ↩

  6. Arizti-Sanz, J., Freije, C. A., Stanton, A. C., Petros, B. A., Boehm, C. K., Siddiqui, S., Shaw, B. M., Adams, G., Kosoko-Thoroddsen, T.-S. F., Kemball, M. E., Uwanibe, J. N., Ajogbasile, F. V., Eromon, P. E., Gross, R., Wronka, L., Caviness, K., Hensley, L. E., Bergman, N. H., MacInnis, B. L., & Happi, C. T. (2020). Streamlined inactivation, amplification, and Cas13-based detection of SARS-CoV-2. Nature Communications, 11(1), 5921. https://doi.org/10.1038/s41467-020-19097-x ↩