Human-Practices
Driven by the pervasive concerns over pesticide residues in food, Hong Kong Joint School aims to develop a pesticide biosensor, PestiGuard, for quick and accurate detection of specific pesticide molecules. Our product was designed with consideration to user accessibility, consumer needs, ecological sustainability, and scientific rigor and reproducibility. As such, we integrated human practice work throughout our study to establish the scientific, regulatory and practical foundations to guide our project.
Our constructive interactions with the general public, organic farmers, food safety regulators, and biotech experts not only played a decisive role in navigating our product design, but also enabled us to characterize the critical need for affordable and effective screening testing with minimal laboratory equipment, leading us to prioritize small-scale organic farms and residue testing laboratories as our primary end-users. In other words, our biosensor intends to serve as a quick screening test that identifies food samples requiring further testing.
Our reflection on and response to each stakeholder interaction were the most powerful driving force for bringing PestiGuard forward. From our interview with the general public and experienced agricultural practitioners, we identified the public need for a quick, accurate pesticide testing method and the key pesticides for testing. Having determined our project topic, we designed our plasmid with suggestions from experts in biotechnology and modelling, then carried out our experiments following the safety guidelines as advised by experienced teachers. Learning from professors in plant genomic studies, we implemented crucial controlled experiments to validate the core mechanisms of our biosensor, and, engaging with biotechnology experts, we explored new systems to standardize our biosensor. Finally, by communicating with private testing companies and other potential users, we identified strengths and weak points in our design to further improve upon our product.
We believe that scientific innovation only proves its true value by centering on the needs of those it serves. By being adaptive, integrative, and responsible, these meaningful conversations have provided us with the assurance to progress at each step, and we will continue to strive for the betterment of the world with our PestiGuard.
This map from Landgeist (2022) visually highlights regions of intense agricultural pesticide application, showing the extent of pesticide application around the world and revealing our home country as one of the regions with the heaviest pesticide use. [1] This visualization of the problem’s scale directly inspired our project, driving us to develop an affordable and user-friendly biosensor. PestiGuard aims to empower farmers and inspectors to make immediate and clear detection of pesticide on food samples, promoting safer agricultural practices and enhancing food safety.
Team members were selected from six different schools, bringing together different experience, expertise and perspectives. After thorough discussion, we came up with three key areas which our project shall be built upon:
With this approach, we brainstormed the following stakeholder map to identify and categorize the groups and individuals who may impact and be impacted by our project. Using the map, we formulated different communication plans that may engage our stakeholders in different steps of our study to inform our project direction.
Our project's development is guided by a continuous, 9-step Human Practice cycle, which transforms insights from the above stakeholders into concrete project improvements.
Our journey began by the recruitment of members from six different high schools, each of which brings with them their own talents and perspectives. As we carried on with our research, we further invited advisors and tutors from different areas of expertise to guide our project. By building a diverse team, our end product reflects the concerns and opinions of different communities from different angles, with different experiences, and of different expertises.
We identified pesticide use as a real-world issue to address and explored the stakeholders involved. By extensive literature review, agricultural practices and expert interviews, as well as engaging with the general public through street interviews, we explored the topic of food safety. This initial exploration identified pesticide residue in food as a major concern and inspired our project’s core concept.
Having pinned down the core concept of our project, we brainstormed the potential mechanisms to develop a biosensor aimed at detecting pesticide residues and ideated important features of the biosensors with inputs from the general public as well as multiple experts. We hope that, ultimately, our biosensor could encourage healthier and more environmentally friendly agricultural practices, so as to increase consumer trust in food safety.
To refine our idea, we documented technical and practical insights gained from expert interviews, such as Dr. Kwong’s advice of using aptamer as a biological switch and the change of end users from the general public to specialized labs. These important records marked the changes in our project direction and guided the design of our initial genetic construct as well as the practical application of our biosensor.
After conducting preliminary experiments, we were faced with the challenge of leakage expression. Thus, we turned to multiple synthetic biologists to troubleshoot. In addition, we gathered feedback and concerns regarding our project that we had missed and integrated these insights into our project.
As we developed our project, we strived to close the gap between our biosensor design and the real solution to stakeholders’ needs by seeking further expert advice on validating our biosensor mechanism and ensuring the biosensor accuracy and sensitivity. This helped formulate our validation strategy with additional control experiments and refine our biosensor design with a standardized sample preparation method.
As we developed a functional prototype, we presented the evidence of our progress to a wider range of stakeholders. Engagements with potential users such as biotechnology and agricultural experts provided valuable feedback on our overall biosensor design and market potential.
Crucially, we extended our engagement through 8. connect and share activities, such as producing educational posters and videos for raising public awareness about food safety to explore potential users beyond our initial target group.
The knowledge gained from these engagements was immediately reintegrated into the process, inspiring further brainstorming and investigation, and guaranteeing that our project developed in close connection with its surrounding environment.
To successfully create an end product that suits our users’ needs, we must thoroughly understand our potential users. The graph below categorizes our potential users based on Technical Literacy, Purchasing Power, and Data Usage Sophistication, which directs our product’s development and market focus.
This graph indicates that our product is primarily targeted at specialized laboratories and organic farms, as these parties have high technical capability, sufficient purchasing power, and require advanced data integration for compliance and quality control.
Even though organic farmers generally have less purchasing power and technical knowledge, PestiGuard offers them a quick, on-site method for daily screening of pesticide contamination in their crops to facilitate compliance. Therefore, organic farmers are also considered one of our target user groups.
While the general public was initially considered, our analysis shows a critical misalignment that they generally have low technical knowledge and limited requirement for data sophistication, so they neither have the compelling need nor the capability to effectively use our biosensor. On top of the fact that individual citizens typically have lower purchasing power compared with larger organizations, we strategically decided to concentrate our efforts on the other three user groups, which clearly demonstrate a clear and genuine need for our product, leading to more focused development and a greater chance of successful adoption. This approach of defining target groups to ensure product features align with users’ needs contribute to creating successful products.
The timeline below summarizes our Human Practices journey. It visualizes how continuous engagement with different stakeholders has directly influenced and improved our project at key stages. From the initial conception of our whole-cell biosensor to its potential transition to a cell-free system, each interaction provides critical insights that help us guide our project’s development.
1. Build a diverse team
2. Explore context
The widespread application of pesticides could lead to significant and irreversible detrimental impact on both environment and public health. According to a science literacy report in 2024 from ScienceDirect, chronic exposure to pesticide residues in food is linked to increased risk of certain cancers, hormonal disruption, and neurological disorders like Parkinson's disease. Children are especially more vulnerable due to their developing bodies and dietary habits. [2] Apart from health concerns, a scientific literature on National Library of Medicine stated that the application of pesticides directly causes environmental degradation, primarily through the contamination of surface and ground water which may compromise the safety of our drinking water, the deterioration of soil quality, and the indiscriminate harm inflicted upon non-target organisms. [3]
Based on our literature review, we questioned if pesticide application remains an inevitable practice in modern food production, and whether we could formulate a solution addressing the issue from our perspective as biology students. Thus, the first step of our investigation was to assess the necessity of a novel biosensor for pesticide detection in food. To learn more about the use of pesticides in farming and to understand if our solution is feasible, we consulted Mr. Jeff Lun, founder of BSF Innovation Limited.
Throughout the interview, Mr. Lun mentioned several common pesticides, such as glyphosate and malathion. He noted we should prioritize testing pesticides which are more commonly used in farming or pesticides which are more toxic to the human body. Hence, he encouraged us to further conduct literature review to identify the more commonly used pesticides or the more toxic pesticides out there, such that we may select these pesticides as our biosensor’s target molecules.
To validate our biosensor technology, Mr. Lun suggested that we collaborate with authorized pesticide residue testing organizations. By conducting parallel testing where the same food samples are analyzed using both our rapid biosensor and the standard laboratory methods, we can generate the authoritative, comparative data essential for building market trust.
This direct comparison would provide an undeniable measure of our biosensor's accuracy and reliability against the established regulatory benchmark. Such collaboration would not only strengthen our technical validation but also significantly enhance our credibility with farmers, regulators, and consumers. This step is fundamental to transitioning our innovation from a promising concept into a trusted tool for the public.
Growing public anxiety over food safety presents a significant marketing opportunity for our biosensor. While consumer concern on pesticide residues is a powerful driver, Mr. Lun advised us to strategically focus our messaging on our product's efficiency and accuracy to maximize market penetration.
However, upon knowing that we planned to genetically modify E.coil as our biosensor, Mr. Lun noted that the public may be reluctant and not willing to use a product made with bacteria, even though there are harmless strains of E.coil. In light of this, we were advised to consider replacing E.coil bacteria so as to increase the public acceptance of and trust in our biosensor.
The consultation with Mr. Lun substantiates the premise that pesticide application remains an inescapable component of contemporary agriculture, thereby cementing the necessity of our proposed biosensor as a pivotal tool for risk mitigation.
Our intervention is posited to deliver significant value to a wide range of stakeholders: For the public, the biosensor serves as an instrument of empowerment, providing accessible data to inform dietary choices and, through collective consumer pressure, catalysing market-driven demand for transparent and cleaner food production practices. For agricultural practitioners, the biosensor serves as an incentive to adopt more sustainable practices and reduce the use of pesticides. For accredited pesticide testing laboratories and organic farm inspectors, our biosensor serves as a screening test that identifies crops that may have potentially violated pesticide limits, which enables them to concentrate their more advanced technology into analyzing those potentially problematic crops.
Regarding the design of our biosensor, Mr. Lun highlighted the safety concern of using E. coli bacteria and prompted us to explore alternatives later in our project. After discussing with other experts such as Dr. Kwong, we would decide to explore the transition from our whole-cell biosensor to a cell-free expression system to mitigate biohazard risks. This shift not only addresses potential safety concerns but also enhances the practicality and stability of our biosensor, making it more suitable for field use.
Throughout the interview, Mr. Lun emphasized that our product must be user-centric. This motivated us to conduct our first street interview. Indeed, while expert analysis has established the technical feasibility and strategic necessity of our biosensor, it primarily operates from an expert perspective. To ensure that our biosensor suits the needs of a wide range of potential target users, one of them being the general public, it is imperative to complement expert perspective with general user perspective. To gather qualitative and quantitative data on public perception on pesticide use and our proposed biosensor, we decided to carry out our first street interview after our interview with Mr. Lun. We believe that through direct public engagement, we may mitigate adoption risks and align the product design with authentic user needs, rather than assumed requirements.
1. Build a diverse team
2. Explore context
Public concern over pesticide residues in food is growing in Hong Kong. According to a news report in 2020 by The Standard, 70% of vegetables were tainted with pesticides residue, with more than half exceeding EU Maximum Residue Limit, heightening consumer anxiety about food safety. Hong Kong Organic Resource Centre from the Hong Kong Baptist University found a high percentage of conventional and self-proclaimed organic vegetables were found containing pesticide residue. [4] A 2019 study by HKBU shows that exposure to these chemicals poses a significant threat to health through endocrine disruption, increased cancer risk, neurotoxicity, and severe reproductive and developmental consequences. [5] These findings highlight an urgent need for accessible detection technology. To assess public awareness and market readiness for a practical solution, our team conducted street interviews (n=128) to evaluate consumer concerns and gauge interest in our biosensor-based pesticide detection kit.
(1) Public perception of pesticide
The data reveals that the public is highly concerned about pesticide residues but possesses remarkably little specific knowledge about them. A majority of respondents (58.6%) reported being moderately to extremely concerned about pesticides in their food, primarily driven by health risks and environmental worries. Despite this concern, a stark knowledge gap exists. A large majority of respondents were unaware of what specific pesticides are used (64.8%) or how much is applied (82.8%). Without knowing which specific pesticides are present, consumers cannot make informed decisions to avoid substances linked to specific health risks (e.g., neurotoxins like chlorpyrifos or probable carcinogens like glyphosate), nor can they effectively advocate for stricter regulation of high-risk chemicals. This leaves the public unable to take targeted action to protect their health or hold producers accountable.This fuels the demand for a biosensor that can provide them insights on the safety of their food.
(2) Biosensor features appealing to public
Our team had conceptualized a preliminary bacterial whole-cell biosensor in hopes of addressing public concerns regarding pesticide residues in food. Public reception to the concept was tested in the street interview and was found to be cautiously optimistic. Nearly half of the respondents (47.7%) would use a test with engineered bacteria only if proven safe, indicating that safety is the non-negotiable barrier to adoption. A significant portion (31.3%) is immediately open to the idea, suggesting a strong potential market if safety can be demonstrably assured. For uncertain individuals, their primary concerns are the biosensor’s long-term safety (57.0%) and affordability (38.3%). Any product development and marketing campaign must directly and credibly address these two points to achieve widespread adoption. The path to market success requires a product that is first and foremost safe and trustworthy, followed by being affordable and accurate.
The most desired features are specificity (63.3%) and affordability (52.3%). This indicates that consumers demand a product that is not just inexpensive, but also scientifically reliable and informative. They want to know exactly what chemicals are present (e.g., glyphosate, chlorpyrifos), not just a generic positive/negative result. This demonstrates a sophisticated market that seeks empowerment through detailed information, not just reassurance.
Our street interview findings have directly shaped both the design and promotion of our biosensor project. Widespread public concern about pesticides, paired with low specific knowledge, confirmed the need for a tool that identifies specific chemicals. Safety and affordability emerged as non-negotiable priorities, leading us to develop a closed-system test with inactivated bacteria and cost-effective production.
For promotion, health concerns emerged as the primary motivator, steering our messaging toward family protection and consumer control but not technical details. The strong public demand for safety assurance shaped our plan to offer demonstrations and transparent educational content. Understanding that the media heavily shapes opinions, we will focus our outreach on visual storytelling and media engagement to build trust and close the information gap.
1. Build a diverse team
2. Explore context
3. Brainstorm broadly
Findings from our street interview highlighted the need for a safe, accurate and quick detection tool that enables the detection of multiple pesticides for the general public. Thus, we embarked on our journey to designing a range of whole-cell biosensors, each of which can detect a specific pesticide. To enable speedy results, we decided to engineer an E.coli with EGFP as our reporter gene. Furthermore, we incorporated a lac operon system to control recombinant protein expression.
However, we could not come up with a reliable molecular recognition element that can cause a dose response between pesticide and protein expression. In addition, we were unsure of which promoter to choose from the many options.
Realizing we have limited knowledge on plasmid design, we turned to Dr. Keith Wai Yeung Kwong for guidance. As the founder of the biotech company DreamTec, Dr. Kwong is an expert in the development of valuable recombinant proteins, and we believed his experience would help us translate our idea into a practical design.
Dr. Kwong proposed incorporating pesticide-binding aptamer into our plasmid construct. After transcription, the aptamers will be transcribed into mRNA and can recognize pesticides. When the target pesticide molecule binds to the aptamer, a conformational change will be induced in the mRNA, leading to the blockage of ribosome binding site (RBS). Subsequently, ribosomes’ interactions with RBS will be less effective, leading to reduced protein translation. In other words, a higher concentration of pesticide should cause reduced fluorescent signals.
There are many promoter options, with T7 promoter being the most popular. While Dr. Kwong acknowledged the strength of T7 promoter in enhancing our protein transcription, he was concerned that the promoter’s excessive strength may lead to significant leakage expression, causing the bacteria to glow even in the absence of IPTG induction, which would waste cellular resources, reduce biosensor sensitivity, and lead to inaccurate results.
To solve the issue, Dr. Kwong gave us two choices. First, we could use a weaker promoter such as the LacUV5 promoter. Alternatively, we could keep the T7 promoter, but culture our bacteria at room temperature and without shaking. After further discussion with Dr. Kwong, we opted for the second option at that stage of our investigation, as it enables our biosensor to be used without a shaking incubator. This allows our biosensor to be used at home by the general public.
Dr. Kwong noted that, if we were to lyse plant cells in our sample to test the pesticide within the cells, there would be a huge amount of cellular debris. To enhance test sensitivity and accuracy, Dr. Kwong recommended that we filter our cell extractions before adding our biosensor into the food sample. Given the fact that pesticide molecules tend to be much smaller than cellular proteins, he believed an ultrafiltration step would help. However, he also warned us that as ultrafiltration is technical and requires centrifugation, the step would be a potential hurdle in marketing our biosensor to the general public. Instead, he noted our biosensor could be a useful tool for government agencies or specialized laboratories.
While Dr. Kwong gave a plethora of suggestions to improve our whole-cell biosensor design, he also pointed out it is difficult to standardize our biosensor due to the variabilities, such as differences in the amount of enzymes, between bacteria. Therefore, it is possible that a sample causes dimmer fluorescent signals not due to the high concentration of pesticides, but due to a lack of protein synthesis materials. In light of the issue, Dr. Kwong suggested we look into cell-free expression systems and try transitioning our whole-cell biosensor into cell-free biosensors after obtaining success with our E.coli sensors.
Prior to the interview, we had planned to use a biosensor that emits more fluorescent signals in more concentrated pesticide, which requires a complicated plasmid design and stumped our project early on. Dr. Kwong’s advice to utilize the interactions between pesticide, aptamer and ribosome-binding site offered an effective solution to the problem. Instead of emitting more signals in higher pesticide concentration, our new design will emit less signals in higher pesticide concentration.
After the interview, pesticide-binding aptamers were identified through rigorous literature review and modeling were carried out to identify aptamers with the higher affinity to our target pesticides. Our modelling team found out RNA aptamers bind more strongly to pesticide than DNA aptamers do. In other words, pesticides will bind to the transcribed mRNA instead of the plasmid construct directly. This lent weight to the theory that our biosensor works by blocking translation in the presence of target pesticides.
Dr. Kwong’s reminder of the strength of T7 promoter enlightened us to try culturing our whole-cell biosensor in room temperature and without shaking. If successful, this will make our design an ideal choice as an at-home detection kit as no shaking incubator is required for using the biosensor. Thus, after designing our plasmids and transforming bacteria with them, experiments comparing the testing efficiency and accuracy of bacterial biosensors cultured in different temperatures, with and without shaking, were conducted to standardize the culture process.
In addition to the use of aptamer and the bacterial culture environment, Dr. Kwong also suggested that we add an ultrafiltration step during food sample preparation. While the removal of lysis debris enhances test sensitivity and accuracy, the step is technical and requires centrifugation, a process that is impossible to carry out in a typical household. Hence, Dr. Kwong urged us to reconsider our choice of users if we indeed adopt the ultrafiltration step. Instead of the general public, our biosensor might be useful for government agencies, specialized laboratories or organic farms as a screening test prior to more labor-intensive tests requiring more advanced laboratory equipment.
Although Dr. Kwong gave many ideas to improve our whole-cell biosensor design, he also highlighted the challenge of standardizing a whole-cell biosensor. Thus, he suggested we explore cell-free expression systems in the future. This advice simultaneously addressed Mr. Jeff Lun’s concern over the use of E. coli bacteria as our biosensor, and may help alleviate user concern regarding the potential biohazard of whole-cell biosensors. As our team continued with our design of the whole-cell biosensor, we also sourced a cell-free expression test kit to study its potential in replacing the whole-cell biosensor. It is in our hope that, ultimately, we can eliminate the need for a bacterial host and standardize our test kit materials to maximize safety and accuracy.
Our interview with Dr. Kwong informed our plasmid design, bacterial culture environment and pushed us to think more in depth regarding sample preparation method, target users and a potential transition from our current whole-cell biosensor to a cell-free sensing system.
1. Build a diverse team
3. Brainstorm broadly
Following Dr. Kwong’s advice, we started searching for pesticide-recognizing aptamers that can be incorporated into our plasmids. Based on rigorous literature review, we selected potential aptamers and modelled their interactions with target pesticides. However, during our modelling process, we faced a significant setback in our molecular docking (MD) simulations: the aptamer-pesticide complex system was not electrically neutral, leading to constant simulation failures. This fundamental error originated from incorrect ligand parameterization, which rendered all attempted runs unusable for our core goals, such as conformational analysis and binding affinity calculations. As such, we decided to seek help from modelling experts, one of them being Mr. Ricky Leung Chung Ki.
Mr. Leung recommended utilizing Acpye, a tool designed to generate standardized GROMACS topology files. By reprocessing the complex's topology files through Acpye, we effectively corrected the faulty ligand parameterization, ensuring the system's electrical neutrality and thus stabilizing the MD runs. This successful intervention allowed us to smoothly complete the analysis of conformational changes and generate the high-quality trajectory data essential for identifying the best aptamers and validating the biosensor's operating mechanism.
With Mr. Leung’s advice, we successfully stabilized our MD runs, enabling us to identify the best aptamers to be incorporated into our plasmid.
4. Document progress
5. Integrate insights
While experts from biotech startups and mathematical modelling had given us concrete advice on the components required in our genetic circuit, we would like to know more about genetic engineering. Thus, we secured an interview with Dr. Wong Tsz Yeung, an expert in molecular virology and immunology and experienced Biology teacher, to seek advice on our plasmid design.
Dr. Wong highlighted the importance of plasmid selection in our project. He noted that, while most plasmids are used for cloning or gene expression, our project requires the regulation of reporter gene expression according to the presence of target pesticides. Hence, the plasmid must be a bacterial expression vector with the flexibility to change the region before the ribosome binding site (RBS). He suggested looking into components that may modify our promoter to suit the needs of our detection system. This aligned with Dr. Kwong’s advice and lent weight to our decision of using the interactions between pesticide, aptamer and RBS to modify reporter gene expression signals.
Upon learning our idea of using an aptamer, Dr. Wong noted that the position of the aptamer must be decided carefully. The aptamer system is used to bind the target chemicals and then interferes with the binding of ribosome for translation in mRNA. Therefore, it is crucial to check compatibility and ensure the aptamer sequence does not interfere with plasmid replication, transcription, or translation.
Given the many available reporter genes, we were initially lost on the type of reporter gene we should employ. Dr. Wong advised us to try out different types of reporter genes to determine the type that has a suitable signal range, which should be strong or sensitive enough to take up the role. He suggested that we try out GFP, RFP and luciferase in our study. In addition, he reminded us to consider if our laboratories had the appropriate equipment for signal detection.
Given the fact that our team decided to use E. coli as our whole-cell biosensor, a bacterial promoter must be used. Dr. Wong confirmed Dr. Kwong’s advice regarding T7 promoter and lacUV5 promoter, and noted that, while we were designing the plasmid, we should be aware of the possibility of changing the promoter with the help of common restriction enzymes to modify the vector easily if necessary. In other words, we should plan ahead such that if the promoter we initially choose doesn’t work well, we can replace it without redesigning the whole plasmid.
Since Dr. Wong emphasized on the importance of ensuring our chosen aptamer sequence does not interfere with plasmid replication, transcription and translation. Subsequently, we made plans to conduct appropriate experiments to confirm this.
After the interview, we also re-considered our choice of reporter gene and ultimately decided on using GFP and RFP. With Dr. Wong’s reminder in mind, we purchased a blue light box with an orange filter to detect fluorescent signals.
Most importantly, Dr. Wong’s modular design approach inspired us to choose restriction enzyme sites around the promoter region, such that we can easily swap out the promoter later if needed.
5. Integrate insights
After our interviews with Dr. Kwong, Mr. Leung and Dr. Wong, we successfully designed our plasmid. But before initiating any laboratory procedures involving genetically modified E. coli, we consulted the opinions of Vice Principal Chan Kwan Wai, an experienced Biology teacher, to better understand the government regulations regarding the use of GMOs in research.
Upon learning that our project involved the use of genetically modified E. coli, Vice Principal Chan recommended the use of E. coli BL21 strain, a non-pathologenic strain classified as Risk Group 1 agents, which means they do not infect the human body. However, he noted that while the BL21 strain is safer than other E. coli strains, we must still handle the bacteria with good industrial hygiene.
Throughout the interview, Vice Principal Chan emphasized the importance of laboratory safety, especially when dealing with biological materials.
Since our experiments involved the use of LB medium, which all types of bacteria thrive in, he reminded us to incorporate appropriate antibiotic resistance gene in our plasmid and culture our transformed bacteria in LB medium with said antibiotic. That way, other types of bacteria without that antibiotic resistance gene would not grow in the LB media and contaminate our E. coli culture. In particular, Vice Principal Chan highlighted that we must use fresh antibiotics, as expired antibiotics may lead to a reduced effectiveness in killing bacteria.
Apart from the use of appropriate antibiotics, Vice Principal Chan also underlined the importance of sterilizing our experimental utensils, LB medium and other materials, so as to eliminate potential contaminations. In addition, he reminded us to utilize the UV light function of our biosafety cabinet to kill any leftover bacteria on the cabinet’s working surfaces.
Vice Principal Chan stressed that such genetically modified microorganisms as the E. coli used in our biosensor project must not be released into the environment. Bacterial leakage could lead to various issues, including the spread of infectious disease and disruption of natural ecosystems.
In addition, Vice Principal Chan reminded us to wear personal protective equipment (PPE) such as lab coats, gloves and masks to protect against direct contact with harmful substances or microorganisms during our experiments.
Vice Principal Chan noted that all laboratory practices in Hong Kong, including our biosensor project, must comply with the Genetically Modified Organisms (Control of Release) Ordinance, Cap. 607. For school laboratories, we should strictly follow the “Safety Guidelines on Microbiology and Biotechnology Experiments in School Laboratories” issued by The Hong Kong Education Bureau. After a visit to our laboratories, Vice Principal Chan confirmed that our laboratories are classified under Biosafety Level (BSL) 2, meaning they are equipped to handle organisms with moderate individual risk. In other words, we were well-equipped to conduct our project in full compliance with government guidelines and ordinances, ensuring that all experimental procedures meet the highest safety and regulatory standards for educational biotechnology projects.
Our interview with Vice Principal Chan reminded us of the paramount role safety plays in our biosensor project. To prevent contamination, infection and ecological hazards, we must always use protective gears, sterilize equipment and handle waste properly. As we carried on with our experiments, we made sure our laboratory practices complied with Hong Kong’s specific laboratory guidelines and ordinances and worked in a BSL-2 lab that provides a secure environment for handling modified bacteria. These safety measures could protect both us and the environment.
4. Document progress
5. Integrate insights
Following Dr. Kwong’s and Mr. Leung’s advice, we selected the best pesticide-recognizing aptamers and incorporated them into our plasmids. We also decided to use the strong T7 promoter such that our whole-cell biosensor can function without a shaking incubator, which enhances its practicality for real-world applications. However, during initial wet-lab experiments, we encountered a significant issue: our biosensors exhibited fluorescent signals even in the absence of IPTG induction, indicating unintended “leaky” expression. Additionally, we sought to explore more efficient reporter gene options besides EGFP to improve signal detection.
To troubleshoot the leakage issue and evaluate alternative reporter systems, we consulted BGI Genomics, a world-leading genomics and synthetic biology research institution. BGI’s extensive expertise in genetic engineering, biosensor development, and large-scale genomic applications made them ideal advisors for addressing our technical challenges. Their experience in designing genetic circuits and optimizing reporter systems—particularly in practical, real-world contexts—provided critical insights applicable to our project. We secured an interview with synthetic biologists Mr. Li Ying Shan and Mr. Lin Tao to leverage their specialized knowledge in overcoming these obstacles.
This addition highlights BGI’s qualifications and explicitly connects their expertise to your project’s needs, strengthening the rationale for seeking their guidance.
Initially, EGFP was selected as the reporter gene due to its cost-effectiveness and convenience, given the fact that no additional substrate is required for it to emit light signals. During our interview, BGI experts acknowledged the advantages of EGFP but recommended that we consider luciferase as an alternative.
Our lab protocol at the time employed the use of a blue light box with orange-tinted glass. However, BGI experts pointed out that our blue light box produces a wide range of wavelengths instead of the 488nm (excitation wavelength for EGFP) specifically, which may lead to reduced accuracy and stability in the emission signal. Because a luciferase reporter gene system does not require specific excitation signals, it may avoid the issue.
While luciferase requires an additional, exogenous substrate, the signal it produces is significantly stronger and can be more easily detected. The fact that visual lights are produced in a luciferase reporter gene system further highlights its ease of detection. Therefore, BGI experts concluded that we may consider replacing our EGFP reporter gene system with a luciferase one to enhance detection sensitivity and ease.
Recognizing the trade-off between detection sensitivity and user convenience, we tested both reporter systems in parallel. This approach allowed us to empirically determine which system offers the optimal balance of performance and practicality.
In our wet lab sessions, we found transformed bacterial colonies exhibiting strong fluorescence even in the absence of IPTG induction. BGI experts had kindly gone through our protocol and plasmid design, and identified the cause of the issue as the close proximity between our aptamer sequence and our lac operator. Due to the lack of spacers, the aptamer sequence had likely disrupted the lac operator’s function. The experts recommended inserting more spacers between these genetic elements to prevent interference and restore proper regulatory control.
The consultation with BGI experts proved invaluable to our project's development, providing specific solutions to critical wet-lab challenges. Their recommendation to test luciferase as an alternative reporter led us to directly compare luciferase and EGFP systems in laboratory settings, evaluating both signal strength and practical usability. Additionally, their diagnosis of our constitutive expression problem prompted us to incorporate spacers between the aptamer and lac operator in our revised plasmid design. After the interview, we ordered new plasmids and transformed new colonies, which no longer exhibited abnormal protein expressions.
4. Document progress
5. Integrate insights
Following BGI’s suggestion, we added spacers between our aptamer and our lac operator and successfully fixed the previous problem with leakage expression. In addition, we took BGI’s advice and designed more plasmid constructs using the luciferase gene in place of the fluorescent protein gene to test out the sensitivity of these two different optical systems.
Unfortunately, induction tests using 0.1 mM and 0.5 mM IPTG with our initially chosen T7 promoter and EGFP system failed to produce a proportional decrease in fluorescence. The standard curves generated across 12, 24 and 36 hours yielded low R-squared values, which means the tests failed to show a dosage response in our bacteria. To solve this issue, we conducted an interview with Mr. Wong Yik Teng, an experienced Biology teacher with many years of experience in conducting biotechnology experiments, to brainstorm potential solutions.
Upon learning our problem, Mr. Wong suggested that we replace the strong T7 promoter with the moderately-tuned PlacUV5-MB7 promoter (Part:BBa_K4941056). As characterized in the iGEM Registry, this component is engineered for an optimal expression intensity that enhances target gene expression while minimizing metabolic burden on the E. coli host.
Upon repeating the experiments with the PlacUV5-MB7 promoter system, we successfully observed a clear, proportional decrease in fluorescence with increased pesticide concentration across all test conditions (different durations, different IPTG concentration).
When we compared different test conditions, we realized the 12-hour induction time yielded the highest R-squared value. However, the actual difference in fluorescence observed between consecutive data points was small, which means there is poor signal resolution for the data generated from 12-hour induction time. In other words, the data generated from this induction time is not reliable for constructing a robust standard curve.
Meanwhile, the group with 24-hour induction time and with 0.5mM IPTG gave the most optimal performance. It maintained a high R-squared value of 0.95, while also providing large, unambiguous differences in fluorescence between each concentration interval. Thus, we decided to use 0.5mM IPTG and 24-hour induction as our test condition to generate our biosensor’s standard curve.
6. Close the loop
7. Present evidence
Our wet lab findings confirmed that increased pesticide concentration dampens fluorescent signals. However, we would like to seek ways to validate the mechanism of our biosensor, ensuring that it is indeed interactions between aptamers and pesticide that leads to dampened signals. In addition, we would like to ensure that our complicated and lengthy pesticide detection protocol can be streamlined into a user-friendly experience.
To address these concerns, we sought expert guidance from Dr. Hon-Ming Lam. His research focuses on genomic studies of soybean abiotic stress tolerance and disease resistance, and he has extensive experience in synthetic biology applications, particularly in developing practical solutions for agricultural challenges. His remarkable studies include resequencing wild and cultivated soybean genomes to identify stress-tolerant genes. His expertise in integrating molecular biology with real-world agricultural applications makes him uniquely positioned to guide our biosensor project, particularly in improving scientific rigor and ensuring practical usability.
While our E.coli exhibited dose response between fluorescent signals and pesticide concentrations, we were concerned that the dose response might not be entirely due to the interactions between aptamer, pesticide and ribosome-binding site (RBS). Thus, we asked Dr. Lam for ways to validate the mechanism of our biosensor. In light of our concern, Dr. Lam proposed several control set-ups to rule out aptamer-irrelevant causes of dose response.
One of our concerns was that pesticides might inhibit bacterial growth, leading to the loss of our whole-cell biosensor and, subsequently, dampened fluorescent signals. To rule out the possibility that our bacteria’s dose response was due to pesticide’s effect on bacterial growth instead of aptamer-pesticide-RBS interactions, Dr. Lam advised us to culture our bacteria on agar plates with increasing pesticide concentrations and compare the bacterial growth rates with untreated controls. This can confirm that the observed fluorescent signal changes observed in our biosensors are not due to pesticide-caused deaths or growth inhibition of E.coli.
The second possible cause of dose response was pesticide’s effect on bacterial transcription. If pesticides affect mRNA synthesis, the observed fluorescence reduction in pesticide-treated bacteria would not be specifically due to inhibition of translation caused by interactions between aptamer, pesticide and RBS. To rule out this alternative hypothesis, Professor Lam advised incorporating the following controls in our experiments:
To observe pesticide’s effects on bacterial transcription, the same reporter gene (EGFP) is included in the plasmids used in all set-ups.
In set-up 1, bacteria with control plasmid (without aptamer sequence) are cultured without pesticide. In set-up 2, bacteria with the control plasmid (without aptamer sequence) are cultured with pesticide. In set-up 3, bacteria with the experimental plasmid (with aptamer sequence) are cultured without pesticide.
If pesticide does not affect transcription, the bacteria in the negative controls 1 and 2 should exhibit similarly high levels of fluorescence. Furthermore, if the aptamer sequence does not affect transcription, the bacteria in the negative controls 1 and 3 should exhibit similarly high levels of fluorescence. There should only be significant signal reduction in the positive control, which demonstrates that pesticide and aptamer alone do not interfere with protein synthesis in bacteria.
Professor Lam further highlighted the risk of off-target effects in complex soil matrices, where impurities may bind to the aptamer or the RBS directly, potentially causing false-positive or false-negative signals. To evaluate this, he recommended a critical experiment that can be conducted by preparing soil samples with graded concentrations of the target pesticide added. Their corresponding light intensity is then measured and compared. If there are no soil impurities that interfere with the light intensity signal output, increases in pesticide concentration should have an additive effect on the reduction of light intensity. Significant deviation would suggest matrix interference, which requires additional sample purification before testing.
In order to quantify the results of our wet lab, a calibration curve is required. It can be produced by plotting the values of light intensity against pesticide concentrations, yet we are struggling to determine which type of function graph is the most suitable choice. Dr.Lam suggested fitting our experimental results in a linear graph first. Given the computational simplicity of linear function, the fit may allow direct calculation of concentrations from light intensity values. The logarithmic function graphs shall be considered as a second option. To decide between linear and logarithmic graphs, we would evaluate their R-squared values to identify the best-fit curve for our data. If neither graph has a good R-squared value, a gradational scale will be implemented to provide semi-quantitative readouts.
To ensure the accuracy of our biosensor’s output, Dr. Lam emphasized the necessity to standardize the food sample. He advised us to rinse a certain surface area of vegetable leaves in a fixed volume of distilled water, then use the rinsed water as the sample, allowing users to check any presence of pesticide on vegetable leaves. By eliminating the variabilities in arbitrary sample extraction, this method produces consistent quantification of the food sample prepared for the experiments, ensuring the accuracy of the experimental results, without requiring complex laboratory extraction steps.
Based on Professor Lam’s advice, we have immediately redesigned our experiments and rescheduled our experimental timeline to incorporate his suggestions. Our web lab team conducted the control experiments as advised to test the pesticide effects on bacteria growth and bacterial transcription, successfully proving that the E.coli bacteria will not be killed by pesticide and that bacterial transcription will not be affected by pesticide. The outcomes of these experiments are foundational to our project validation, ensuring that our biosensor operates as designed and providing clear data to support our conclusions. Professor Lam’s insights were instrumental in refining our experimental approach to ensure the reliability and accuracy of our biosensor’s performance.
8. Connect and share
9. Carry it forward
Following the development of our PestiGuard biosensor prototype and its accompanying analysis software, our team shifted focus towards user-centric refinement. While our initial street interviews (n=128) confirmed significant public concern over pesticide residues in Hong Kong’s food supply, this subsequent phase aims to directly evaluate the usability and effectiveness of our software interface. The successful adoption of our technology relies not only on the biosensor’s accuracy but also on a seamless and intuitive user experience when using the biosensor’s analysis app. This street interview (n=307) is designed to gather critical feedback from users and stakeholders on the app’s functionality, workflow, and design, ensuring our final product is both powerful and accessible for the end-user.
(1) Holistic Usability and Workflow
The evaluation findings confirm strong performance in the application's fundamental operational design. Users reported the core workflow to be logical and well-structured, enabling an efficient analytical process. This conclusion is supported by consistently positive feedback regarding the application's procedural flow, which respondents described as coherent and manageable. Moreover, the interface's operational components received favorable assessments for their guidance quality and transition smoothness. Users particularly noted the fluid movement between analytical stages, allowing continuous progress through the detection procedure. The systematic organization of the workflow ensures users can navigate from initial sample processing to final results without encountering significant obstacles. This integrated operational framework demonstrates the application's effectiveness in presenting complex analytical procedures through a structured and approachable interface design.
(2) Navigation and Layout Clarity
The survey results indicate a significant success in the domain of Navigation and Layout Clarity. Users found the app’s organization to be intuitive and straightforward, which facilitates a seamless user experience. This is primarily evidenced by the high scores for the app’s overall layout, which was reported as clear and easy to navigate. Furthermore, the instructions and labels embedded within the user interface were perceived as clear and helpful, effectively guiding users through the analytical process. The functionality dedicated to accessing past results, the "Analysis History" page, was highlighted as particularly user-friendly, allowing users to retrieve historical data with ease. This cohesive and logical organization is further validated by the effortless navigation reported between the three core functions of the app: "Pesticide Analysis," "Analysis History," and "Settings.
(3) Visual & Interface Design
Similarly, the app received strong positive feedback regarding its Visual and Interface Design. Respondents expressed a favorable view of the app's overall aesthetic, including its color scheme, typography, and iconography. A key strength identified was the design of the functional icons, which were intuitively understood by users, enabling them to easily discern each icon's purpose without confusion. Complementing this, the typographical choices—specifically the font size and color contrast—were rated highly for readability and visual comfort, ensuring that on-screen text was easy on the eye during extended use. This thoughtful attention to visual design elements has contributed to an interface that is not only aesthetically pleasing but also functionally clear and accessible.
Our street interview successfully engaged 307 participants. The feedback collected has become a direct input for shaping the future development and marketing strategy of our biosensor software.
The overwhelming positive feedback confirms that our application is successfully meeting user needs with its intuitive design and reliable performance. While respondents were hugely satisfied by our app’s visual design, we noted that it received relatively more neutral feedback compared with our app’s other aspects. Hence, for the next step on product refinement, we shall improve the visual design. By focusing on this target, we will ensure users have a pleasant experience using our app and significantly upgrade the overall experience.
For our promotional strategy, these results provide a clear direction. Our marketing will highlight the seamless user experience as our primary strength, assuring potential customers of the software's ease of use and reliability. We will craft our messaging to emphasize this proven success, building trust and strengthening our position in the market.
Throughout our project, we completed a full human practice cycle to guide our project forward and connect to the real world.
Our project undergoes significant changes driven by consulting and interviewing different stakeholders, highlighted by three key milestones. First, experts’ guidance is crucial for troubleshooting issues with molecular docking and protein expression, resolving our core technical challenges. Second, discussion with experts allowed us to strategically narrow our target user from the general public to specialized laboratories and organic farms, positioning our application as a reliable screening tool. Third, after the completion of our E. coli biosensor, we decided to explore the potential transition from our whole-cell system to a cell-free expression system, a transformative change that may enhance usability and practicality for PestiGuard’s real-world implementation.
Overall, our Human Practices journey reflects our team’s continuous exploration of the relationship between biotechnology, agriculture and food safety. Through active interactions with a wide range of stakeholders,from end-users to experts from multiple fields,we considered multifaceted perspectives and navigated the successful implementation of PestiGuard. We constantly strived to improve social and technical inadequacies by educating the public on pesticide risks and incorporating experts’ advice on biosafety, practicality and usability, such that we may contribute to the enhancement of food safety and public health. By consistently delving into the essence of sustainable innovation, from ensuring the reliability of our scientific claims to promoting fair access to detection technologies, we have contributed to close the divide between laboratory research and practical field application. Carrying forward the lessons from this journey, we will persistently seek out innovative approaches to balance agricultural productivity with public health.
1. Pesticide use. (2022, October 28). Landgeist https://landgeist.com/2022/11/01/pesticide-use/
2. Ahmad, M. F. (2024, April 15). Pesticides Impacts on Human Health and the Environment with Their Mechanisms of Action and Possible Countermeasures. ScienceDirect. https://www.sciencedirect.com/science/article/pii/S2405844024051594
3. Aktar, M. W., Sengupta, D., & Chowdhury, A. (2009). Impact of pesticides use in agriculture: their benefits and hazards. Interdisciplinary toxicology, 2(1), 1–12. https://doi.org/10.2478/v10102-009-0001-7
4. Study shows more than half of vegetables tainted with pesticides residue | The Standard https://www.thestandard.com.hk/breaking-news/article/146463/ere
5. Prenatal exposure to organochlorine pesticides and its association with birth outcomes - Fang, Jing https://scholars.hkbu.edu.hk/ws/portalfiles/portal/55045687/OA-0673.pdf