Overview
Early detection of diseases is crucial. To achieve this, we can target disease biomarkers for detection to assist in early diagnosis. Currently, clinical tests for disease biomarkers, such as ELISA, RIA, and mass spectrometry, are widely used but tend to be time-consuming. Many have developed biosensors and rapid diagnostic tools (RDT) for this purpose. However, there are still many challenges in detecting disease biomarkers, primarily due to the complexity of samples and the insufficient sensitivity and specificity of detection technologies. To address these issues, our team has proposed a new strategy this year, utilizing a pre-trained Natural Language Processing (NLP) model to identify potential disease biomarkers. Through this approach, we aim to overcome the current bottlenecks in detection technologies, providing a more efficient and cost-effective pathway for early diagnosis and timely treatment of diseases. This not only helps to shorten diagnostic times but also improves diagnostic accuracy and patient treatment outcomes.
Model to predicted peptide
This year, our project focuses on developing a model capable of identifying the relationships between diseases and proteins to accurately pinpoint potential disease biomarkers. The model systematically reviews all articles within the database to identify and label significant associations between proteins and diseases.
Later, we listed the top ten highly relevant diseases and chose leukemia as the target disease to prove the feasibility of our project. As a result, we chose CD97 as a key underlying marker for leukemia.
We aim to determine the expression level of CD97 through electrochemical detection, which requires identifying the receptor for CD97. From research papers, we identified that the protein CD55 can bind to CD97. [2] Furthermore, our goal goes beyond just developing a "search engine" for identifying the relationships between proteins and diseases. Therefore, we have also incorporated generative artificial intelligence (Generative AI) to design short peptides that can bind to potential markers, specifically 2BOU-2 and 2BOU-3. These peptides will be utilized in our detection system to enhance its functionality.
If you want to see more detailed descriptions, please go to the Model and Software page.
Construction of BELO with CD97 and predicted peptides
We aim to utilize electrochemical detection methods to detect the target biomarker. To achieve this, we will use the Lpp-OmpA-GS linker sequence, followed by a receptor that can bind to the target biomarker, allowing the receptor to be expressed on the surface of E. coli. [3] Subsequently, we will detect the electrical signals generated when the receptor binds to the biomarker and analyze the data to determine whether the patient has the disease.
Lpp-OmpA (BBa_K5151019)
Lpp-OmpA protein is the expression system of outer membrane protein. The lipoprotein (Lpp) possesses the function of targeting to the outer membrane. While The OmpA domain constitutes eight-stranded β-barrel to construct an anchor on the outer membrane, providing the stable expression of the protein on the outer membrane.
Lpp-OmpA-GS Linker-GFP (BBa_K5151009)
To check our protein can be expressed by Lpp-OmpA, we chose to use Green Fluorescent Protein (GFP) as an example of a functional protein. GFP was chosen because of its unique fluorescent properties, which can be used to test Lpp-OmpA functionality and ensure that our experimental aims can be clearly verified.
CD97-His (BBa_K5151013)
Based on the results generated by the model, we selected CD97 as the biomarker for leukemia from all proteins associated with the disease. We then aim to use an electrochemical detector to measure the expression level of CD97 to determine whether the patient has the disease.
Lpp-OmpA-CD55 (BBa_K5151010)
We aim to determine the expression level of CD97 through electrochemical detection, which requires identifying the receptor for CD97. Therefore, we found a protein that binds to CD97 from research papers, namely CD55.
Lpp-OmpA-2BOU-2 (BBa_K5151006)
In our project, we aim to determine the expression level of CD97 through electrochemical detection, which requires identifying the receptor for CD97. In addition to the CD55 protein found in research papers, we have also incorporated Generative Artificial Intelligence (Generative AI) to design short peptides that can bind to potential markers, specifically 2BOU-2. We use these peptides in our detection system to enhance its functionality.
Lpp-OmpA-2BOU-3 (BBa_K5151015)
In our project, we aim to determine the expression level of CD97 through electrochemical detection, which requires identifying the receptor for CD97. In addition to the CD55 protein found in research papers, we have also incorporated Generative Artificial Intelligence (Generative AI) to design short peptides that can bind to potential markers, specifically 2BOU-3. We use these peptides in our detection system to enhance its functionality.
If you want to see more detailed descriptions, please go to the Part page.
If you want to see more detailed descriptions, please go to the Experiments page.
To view the results of the experiment, please go to the Results page.
Proof of BELO with CD97 & predicted peptide through functional test
Enzyme-linked immunosorbent assay (ELISA)
We will confirm the binding of biomarkers and receptors and validate our results. The experiment will use PB buffer to wash away nonspecific bindings, followed by the detection of successful protein binding to the antibody using TMB reagent (indicated by a color change to blue). Finally, we will evaluate the results by measuring the OD630 value.
Electical chemical analysis - current analysis
We will confirm the binding of biomarkers and receptors and validate our results. The experiment will involve mixing the liquid containing biomarkers and receptors together, and then centrifuging to precipitate the biomarkers that have successfully bound to the receptors. Since the biomarkers have a 6X His-tag attached, we will add Anti-His conjugated with HRP. Finally, we will add TMB reagent to detect the current change at different biomarker concentrations.
If you want to see more detailed descriptions, please go to the Experiments page.
To view the results of the experiment, please go to the Results Page.
These experiments were designed to prove the application of lipoprotein (Lpp)–OmpA fusion vehicle and to confirm that the protein attached to Lpp-OmpA is expressed and functions properly.
Summary
Early diagnosis of diseases is crucial for improving patient treatment outcomes. To address the challenges of insufficient sensitivity and specificity in existing detection technologies, our team has proposed an innovative strategy that utilizes a pre-trained Natural Language Processing (NLP) model to identify potential disease biomarkers. This approach not only helps overcome the bottlenecks in current technologies but also provides a more efficient and cost-effective solution for early diagnosis and timely treatment of diseases. Our research aims to shorten diagnostic times, enhance diagnostic accuracy, and ultimately improve patient prognosis and quality of life.
Reference
- GeeksforGeeks. (n.d.). What is a bipartite graph? GeeksforGeeks. https://www.geeksforgeeks.org/what-is-bipartite-graph/
- Otoupal, P. B., Dahl, K. M., Baltrus, D. A., & Way, J. C. (2021). Towards environmentally resistant synthetic microbes: Evolution, biomanufacturing, and resilience. Journal of Biological Chemistry, 296, 100610. https://doi.org/10.1016/j.jbc.2021.100569
- Kelly, B. G., Vesper, S. J., & Stark, J. M. (1987). Plasmid-based bacterial detection systems. Advances in Applied Microbiology, 32, 47-74. https://doi.org/10.1016/S0076-6879(87)02607-2