2025.6.7_Discussion with Wang Lab

@Inspiring meeting with our PI, Professor Jiangyun Wang.

Professor Jiangyun Wang is from the Institution of Biophysics, CAS, and specializes in synthetic biology. One of the meetings with him that deserved particular attention was the one after the initial conception of the project prototype and experimental scheme, when we visited his lab and sought his advice on improvements, especially regarding the selection of our second input and final output. We benefited a lot from it.

Why did we contact?

• Prof. Wang has rich experience in synthetic biology. As our PI, he can offer us some guidance on how to come up with iGEM projects. Some of his students had also participated in iGEM as undergraduates and can give us practical lectures to learn from.
• In May, we finally selected our topic, PROTATO, for this year's iGEM project, from three topics we've been researching. With the goal of detecting tomato disease, we were looking for a proper output signal and a smart design of a handy device.
• What aspects should we consider when selecting inputs? At that time, our input choices were AvrPphB (another protease secreted by Pseudomonas syringae pv. tomato) and hydrogen peroxide (broad-spectrum plant immune response signal). Are they reasonable?
• What kind of output form is more in line with our project? We read about split luciferase in the article, with a detection duration of one hour. However, luciferase requires substrates to function, increasing additional cost. We converted our attention to split GFP (spGFP). But if we choose spGFP as the output, we would have to create a device that can detect the fluorescence signal, which increases cost and lowers accessibility. Is there a better choice?
• How should we design our hardware to make it useful to farmers, except for the output signal mode?

What information did we acquire?

Prof. Wang, together with his team members, affirmed the feasibility of our project and put forward many suggestions for improvement.

• During the incubation period, the plant may not produce hydrogen peroxide, which means our second input may be invalid. It's also possible that the pathogen only secretes a small amount of ArvPphB, leading to difficulty in signal amplification.
• Some proteases perform cleavage function after recognizing the overall structure of their substrates, while others only need a specific peptide segment to target and cut. Thus, attention should be paid to the protease AvrPphB's cleavage site on the substrate PBS1 since the cleavage sequence is important for the design of a flexible linker.
• As for the detection device, a student in Wang Lab who is familiar with the paper disc detecting method introduced keys in designing the outputs, mentioning an example of using trehalase to transform signal. Prof. Wang encouraged us to look for more output forms, including split RFP and tripartite spGFP.
• The selection of the output form will greatly influence the convenience of our hardware. Prof. Wang urged us to talk to farmers face-to-face for a better understanding of what their real needs are.

What did we do after reflection?

For Wet Lab

• We looked for choices to replace hydrogen peroxide, utilizing the protein-protein interaction mechanism.
• We added a positive feedback circuit in our protease input circuit to amplify the signal.
• We searched for the basis of AvrPphB's recognition. Unfortunately, it requires the full structure of its substrate to cut. But during the literature review, we found another protease produced by the pathogen, AvrRpt2, which only needs 7 amino acids to recognize and cleave its substrate, greatly simplifying the design of the linker. It ended up as the final choice for our first input.
• We looked up information on split RFP, split YFP, and tripartite split GFP. A paper introducing a sensor design based on tripartite spGFP aroused our interest. We decided to use this form of split GFP instead of bipartite spGFP owing to its quick self-assembly.

For Dry Lab

• We began to focus on the design of our hardware device and look for advice from possible users.

For HP

• We started seeking proactive conversations with tomato growers.

2025.6.13_Conversation with tomato grower Mr. Xu

@Get in touch with farmers' real needs according to the discussion with Mr. Xu.

Yunfeng Xu is an expert in tomato growing in Qingdao, Shandong Province. Limited by our spatial distance, we contacted him online and introduced our project to him.

Why did we contact?

• Mr. Xu has been growing tomatoes for more than three years. He was willing to get to know our project.
• We expected more specific knowledge about tomato growing and the disease we are researching, not from literature but from real-life experience. What's the frequency and severity of this disease? What's the impact of it on tomato production? Are there any detection methods or treatments for it currently, and if there are, what's the effect of these actions?
• We were curious about the attitude of farmers towards our device. Would they see it as necessary? Would they find the device difficult to use?

What information did we get?

• Greenhouse cultivation has a higher prevalence of this disease, affected by factors like weather (the disease favours constant warm, cloudy days with high humidity), planting density, and the management level of peasant households.
• The influence of this disease depends on when farmers discover it and the degree of its infection. The initial symptoms manifest on the leaves as small spots. Farmers have to spray broad-spectrum bactericides or remove the whole plant to prevent its spread. Overall, the earlier it is discovered, the more effective the treatment will be. The disease will certainly reduce the production and quality of the tomato, no matter its taste or its appearance.
• Prevention is more important than treatment. Farmers spray bactericides from time to time to enhance the resistance of plants. Farming operations such as removing old leaves and increasing air permeability are also necessary. As for detection, farmers can only distinguish based on their experience.
• Root rot disease and grey mold disease are also severe problems that tomato farmers are facing. Mr. Xu asked if our system could do something with them.

What did we do after reflection?

For Wet Lab

• It was inspiring to know our project is quite meaningful in farming practice to boost productivity. We reflected on the platform nature of our project. Our input can be replaced by other proteases specifically secreted from other pathogens. To make sure the detection is precise, the second input we were looking for should also be specific.

For Dry Lab

• We decided to use tomato leaves as our detection subjects since the symptom appears on the leaves first.

2025.6.15_Conversation with technician and tomato grower Mr. Liu

@A milestone for our output signal, thanks to our conversation with Mr. Liu.

Jie Liu used to be a farmer in Inner Mongolia Autonomous Region, with extensive experience in agriculture. Although he was on a business trip, he was glad to have an online discussion with us.

Why did we contact?

• Besides being a grower, Mr. Liu is also one of the Chief Technology Officers in Inner Mongolia Zehui Xiwang Biotechnology Company Limited. He has both perspectives from a farmer and the market side.
• We still wanted to know more about our target disease and pathogen. Farmers from different regions may have different habits and methods.
• We hoped to learn from Mr. Liu about his market mindset. How to design our hardware according to our target users?

What information did we acquire?

• The feedback from Mr. Liu agrees with Mr. Xu. Whether open-air or greenhouse cultivation, the disease is impossible to control without the help of bactericides, especially on cloudy and rainy days. However, chemicals in bactericides affect the quality and health values of tomatoes.
• Another fact about this disease is that it can spread via water drops or mist. Therefore, plants have a possibility of getting infected when farmers are spraying foliar fertilizers or bactericides targeting other diseases.
• As for our hardware, Mr. Liu suggested that if we equip every plant with one device, it would cost too much. Instead, we should focus on developing compact and convenient equipment for each field where farmers can collect leaves and put them into the device at certain intervals. He took glucometers as an example to explain what kind of tools can be accepted by farmers.

What did we do after reflection?

For Wet Lab

• We were surprised and also excited when Mr. Liu mentioned glucometers, since we've been researching the feasibility of the trehalase circuit after the meeting with Prof. Wang, and the results were positive. So we decided to replace spGFP design with split trehalase. Trehalase degrades its substrate trehalose into glucose, turning input signal into glucose concentration that can be detected with glucometers.

For Dry Lab

• Since the output signal was determined, our hardware design changed, with software design starting to cooperate with it.
• We knew that Pst can spread rapidly by rain splash and aerosols, especially when there are injuries on leaves. Model members started to develop a model to simulate its spread.

2025.6.19_Communication with UCAS Professor Ying Wang

@Reflecting on details in experimental and hardware design.

Professor Ying Wang, at the University of Chinese Academy of Sciences, is an expert in plant developmental biology and synthetic biology. We held an online interview with her to introduce our project and look for advice for our second input.

Why did we contact?

• We've been tackling the problem of the second input, selecting from the effectors Pst secreted via its Type III Secretion System (TTSS). In June, we excluded the proposal using protease AvrPtoB. We wished to know more about the TTSS of Gram-negative bacteria and how it works when infecting plants.

What information did we acquire?

• Prof. Wang offered literature on effectors like AvrPtoB and encouraged us to screen relevant effector proteins.
• She also asked about our trehalase output and doubted whether the substrate trehalose would influence the detection of glucose.

What did we do after reflection?

For Wet Lab

• We read about reviews on effector proteins in Pseudomonas syringae. However, it seemed that they couldn't recognize certain peptide sequences to achieve cleavage.

For Dry Lab

• We learnt about the mechanism of the glucometer and made sure that trehalose wouldn't interfere with the detection of glucose.

2025.6.19_Communications with CASS Professor Ali Chai

@Catching up on common detection methods in the past.

Professor Ali Chai is a researcher at Institute of Vegetables and Flowers, Chinese Academy of Agricultural Sciences (CASS). We also had an online meeting with her.

Why did we contact?

• Prof. Chai focused on the research on the detection, prevention and control technologies of major vegetable diseases, as while as the aerosol transmission mechanism of vegetable diseases and the application technologies of fungicides. Thus, we intended to consult her about current detection methods they use.
• We also looked forward to information on plant pathology to help us better understand the process of Pst infection.

What information did we acquire?

• Prof. Chai told us that the current detection methods involve identifying specific primers for the target bacterium with PCR or RPA (Recombinase Polymerase Amplification, an advanced detection method that has the advantage of simplified procedures, a relatively short reaction time, and a low cost). However, they still rely on specific genes of the pathogen. None of these detection systems can be applied in the field. She praised our method as highly innovative using a synthetic biology-based approach. She believes that it will overcome current difficulties and will have a strong application prospect.
• Since DC3000 is a well-recognized standard strain of Pst, the research on effectors related to bacterial speck of tomato is quite thorough. We were presented with the basic characteristics of the RIN4 protein, substrate of protease AvrRpt2, its role in plant immunity, and its interaction with pathogenic effectors.
• She also recommended several possible candidate proteins to optimize the first input, including AvrPto, AvrPtoB, AvrRpm1, and HopF2.

What did we do after reflection?

For Wet Lab

• It's great to hear that our project outstands current detection methods. We then did a more detailed research on these methods and summarized their pros and cons to compare with our system.
• We investigated her recommended effectors. However, they can only modify RIN4 through phosphorylation or ubiquitination, which means they can not be used for the first input. We still took AvrRpt2 as the input.

For HP

• We noticed the methods provided by Prof. Chai are not suitable for practice in the field. Also, her attitude towards some detailed designs of our project is actually opposite compared with that of the tomato growers we've talked with before. Looking back on our investigations, we came to a profound realization that farmers and researchers have a distinctive gap in the way of thinking. We must seek professional advice from experts and dive deep into real life at the same time. We summarized this perception into 'Beyond lab', the second discipline in our 'A-F' principles for HP.

2025.9.26_Conversation with Professor Peng Zhou

@Optimizing the future direction of our predict and design model.

Professor Peng Zhou is an associate professor from the University of Electronic Science and Technology of China. He mainly engages in theoretical calculation research on the sequence, structure, activity, and function of biological peptides, as well as the design and development of peptide drugs. We attempted to get in touch with him via email.

Why do we contact?

• Prof. Zhou experts in the field of quantitative structure-activity relationships of peptides (QSAR). He and his research group built a QSAR database called PepQSAR, which is the first database dedicated to QSAR. At the same time, the idea of our prediction model is actually the same as QSAR models.
• We expected advice on our prediction and design model, and effective methods for dealing with high-dimensional and small-sample situations in the prediction part.

What information did we acquire?

• Prof. Zhou gave a very detailed reply. In the predictive model section, he suggested that in the case of such a high-dimensional and small-sample dataset, we could change the encoding method of the peptide chain (for example, using amino acid descriptors, orthogonal encoding, etc.) and the regression method (for example, partial least squares regression, PLS), and conduct simulations using different combinations of encoding methods and regression methods respectively to compare the effects of different combination models to obtain a better performance. Additionally, he mentioned that it is possible to enlarge our database by replacing regression model with discriminative model.
• As for the design model section, he advised conducting a statistical analysis of the results of multiple runs of the design model, which would be more likely to obtain more valuable information about the mutation design. He also pointed out that the most reliable method is still molecular simulation.

What did we do after reflection?

For Dry Lab

• Due to the lack of information and time shortage, we did not further improve the encoding method of the peptide chain, but attempted to fit it using other regression models. And we did attempt to change the model's pattern into discrimination and achieved certain results. (For more details, please refer to the engineering and Model part).
• We have modified the design model part to a method that runs the GA algorithm multiple times and conducts statistical analysis on the mutation position and the type of amino acids after the change in each result. We hope to design a new mutation design method based on molecular simulation in the future, which can be faster, more automated, and have more accurate results.