Contribution
Developed an Accurate Detection Platform
We have developed a novel precision diagnostic platform capable of intelligent analysis for multiple nucleic acid targets. Both the platform itself and our development process will provide insights and experiences for other teams in the future.
Overview of detection principles
The working mechanism of our designed molecular computing system can be divided into steps such as recognition, weighting, summation, reporting, and decision-making, as shown in the diagram below:
Recognition The 3' end sequence of the Track strand is designed to be complementary to the miRNA target. Therefore, the weight probe can specifically bind to the target, completing the recognition of the input signal.
Weighting The Track strand is attached with weight molecules corresponding to the target DNA weight numbers. During the DNA polymerase-mediated DNA synthesis process, the weight molecules are displaced and released from the newly synthesized daughter strand. Through this process, the signal from the target is quantitatively converted into the number of weight molecules.
Summation During the chain displacement reaction, weight molecules of the same kind are displaced and released into the solution. The signal from the target is uniformly converted into weighted weight molecule signals, and the signals are automatically summed.
Reporting Free weight molecules are quantitatively converted into fluorescent signals through toe-mediated chain displacement reactions.
decision-making After obtaining the fluorescence signals, the subtracted result of the weighted sum of each target can be used to determine the outcome.
After obtaining the fluorescence signals, the difference between the calculated fluorescence signals for targets with positive weights (>0) and negative weights (<0) is calculated (represented as D). D>X indicates bacterial infection, D<Y indicates viral infection, and Y<D<X indicates a healthy condition.
Machine Learning Screening Targets
Check the Machine Learning Screening Targets for more details.
Our team has developed a target screening and weight allocation algorithm capable of quickly and accurately identifying nucleic acid detection targets.
Initially, we validated the effectiveness of this algorithm using human gene expression profiles from individuals infected with bacteria or viruses, achieving results consistent with existing literature.
Subsequently, we made slight adjustments to the algorithm for screening miRNA expression profiles associated with lung adenocarcinoma.
Again, feasible targets with assigned weights were obtained. This suggests that with minor modifications, the algorithm can be applied to most nucleic acid-based disease detections, demonstrating its broad application potential. It aids in expediting disease diagnosis, enabling personalized medicine, early prediction, and disease monitoring.
Furthermore, it supports in-depth exploration of disease mechanisms.
The automation features of this technology are poised to save time and resources, making significant contributions to the fields of medicine and biological research, advancing scientific frontiers, and promoting human health improvements.
Tutorial about Designing Molecular Probes
NUPACK is an open-source nucleic acid analysis and design software developed by the University of California. Due to its powerful functionality and accurate simulation results that closely resemble real-world scenarios, it is widely used.
- However, we noticed a lack of tutorials about NUPACK on Chinese websites. Therefore, our team has created instructional videos on how to use the software.
Checkout to our instructional video [Scientific Research Tutorial] NUPACK web version usage tutorial!
The tutorial primarily focuses on identifying the missing target sequences when given the desired DNA molecular structure and some partial sequences.
Kinetic models of molecular computational processes
Check the Simulation for more details.
Our team conducted dynamic modeling of the molecular computation process, validating the theoretical feasibility of the molecular computation-based diagnostic system.
- We refined the model using experimental data to accurately predict the kinetic characteristics of chemical reactions. This work not only provided a solid theoretical foundation for our research but also holds extensive influence and practical value.
Our model not only guided the current experiments but also serves as a valuable reference for future molecular computation experiments by other teams.
Through dynamic modeling, researchers can gain deeper insights into the dynamic changes of the molecular computation process, crucial for experimental design and optimization.
It is expected to assist other teams in more effectively managing experimental resources, reducing trial and error costs, enhancing the success rate of experiments, thereby promoting the feasibility and sustainability of research projects.
Molecular Classifier for Bacterial & Virus Detection
We chose 'Pathological Diagnosis of Human Acute Respiratory Tract Infections' to validate the feasibility of our molecular computation-based diagnostic system.
- We successfully conducted a proof-of-concept verification and developed a molecular classifier capable of distinguishing between bacterial and viral infections in acute respiratory tract infections. This classifier can detect and intelligently analyze various nucleic acid targets, providing decisive diagnostic results.
Parts
For details, please refer to page 🔗Parts.
The molecular classifier is composed of a series of weight probes, Bst DNA polymerase large fragment, and fluorescent reporting probes.
To validate its functionality, we synthesized the corresponding 7 target molecules which is available in the 🔗Parts page. All the parts have been uploaded to the Registry of Standard Biological Parts for further use. If anyone in the future needs to use this molecular classifier, they can simply need to obtain the sequences of these parts to construct the molecular classifier (Bst DNA polymerase large fragment is a commercial protein, other parts are in the form of ssDNA).
Brief introduction to the effect of the proof of concept
Using SDS Gel Electrophoresis to Verify the Successful Synthesis of the weight probes
The electrophoresis results indicate that all weight probes have been successfully synthesized.
Characterization Results of the Molecular Classifier
We have successfully completed the verification of each workflow of this molecular classifier.
The fluorescence characterization results of the molecular classifier, as shown in the diagram below, demonstrate a linear relationship between fluorescence intensity and the sum of weight numbers of the mixed targets.
- This indicates that the weight probes can specifically recognize their corresponding targets without responding to other targets. Signals from targets with the same symbol can be summed after weighting, and the summed results can be output in the form of fluorescence intensity signals.
After subtracting different fluorescence values, the judgment results of bacterial infection, viral infection, or healthy condition can be obtained. In the diagram below, D>X represents bacterial infection, D<Y represents viral infection, and Y<D<X indicates a healthy condition.
Molecular Classifier for Lung Adenocarcinoma Detection
After the successful experiments mentioned above, we further validated the universality and superiority of our molecular computing diagnostic system by choosing 'diagnosis of lung adenocarcinoma.' We have completed partial conceptual validation.
Parts
For details, please refer to page 🔗Parts.
We utilized machine learning algorithms to identify 9 miRNA targets. Therefore, our molecular classifier designed for lung adenocarcinoma diagnosis consists of 9 weight probes corresponding to these miRNA targets, Bst DNA polymerase large fragments, and fluorescence reporting probes. To validate its functionality, we synthesized the corresponding 9 target molecules.
- The parts included in the molecular classifier are listed in the table below, all of which have been uploaded to the Registry of Standard Biological Parts.
Characterization Results of the Molecular Classifier
The partial functionality of this classifier has been characterized.
The fluorescence characterization of the molecular classifier shows that the fluorescence intensity of the solution is linearly related to the target weights and concentrations. The weight probes can detect and weight the target signals.
An Engineered Strain that can Produce Bst DNA Polymerase, Large Fragment
In our designed DNA computing system, an enzyme with strong strand displacement ability is required to catalyze the isothermal amplification reaction of nucleic acids.
We chose the commonly used Bst DNA Polymerase, Large Fragment (BstpolLF) for isothermal amplification. It has been produced, sold, and widely used in the research field by biotechnology companies.
To reduce detection costs, we attempted to construct an engineered bacterium capable of producing BstpolLF. Additionally, BstpolLF has been submitted as a biological part to the Registry of Standard Biological Parts (BBa_K4612107).
The following is information about plasmids expressing BstpolLF engineering:
| Name | Content |
|---|---|
| Plasmid type | Cloneable expression plasmid |
| Source cell | E. coli |
| Selective marker | Ampicillin resistance gene |
| Functional parts | Ribosome binding site |
| Promoter | ATG,6xHis affinity tag,thrombin,RBS recognition and cleavage site, T7 tag, lac repressor, aminoglycoside phosphotransferase |
| Other parts | MCS, f1 bacteriophage origin of replication,Rop protein,plasmid size: 7105 bq, insert size: 1767 bp, |
| Notes | growth conditions: 37 C, shaking 300 rpm |
The Large Fragment of DNA Polymerase I possesses 5'→3' DNA polymerase activity but lacks 3'→5' and 5'→3' exonuclease activity.
Its optimal catalytic temperature is 45°C. Isothermal nucleic acid amplification catalyzed by this enzyme does not require thermal denaturation of DNA, avoiding the time loss caused by the three-temperature cycling used in traditional PCR.
Isothermal amplification also offers higher sensitivity, with detection limits two orders of magnitude lower than traditional PCR techniques. The average amplification factor can reach up to 200 times, producing significant amplification bands in the 500-2000 bp range.
In the field of PCR, it effectively reduces time and energy costs, while also providing advantages such as high specificity, simple operation, and easy result interpretation.
The 5'-3' polymerase activity domain of the enzyme is primarilyutilized in our molecular computing system. This domain performs thecatalytic polymerization function of the enzyme.
The plasmid is crucialin various applications such as loop-mediated isothermal amplification(LAMP), helicase-dependent isothermal gene amplification (HDA), multipledisplacement amplification (MDA), whole genome amplification (WGA),sequencing high-GC content DNA, rapid sequencing of picogram-level DNAtemplates, library construction sequencing, and other fields.
Reference
J. McClary, S. Y. Ye, G. F. Hong & F. Witney (1991) Sequencing with the large fragment of DNA polymerase I from Bacillus stearothermophilus, DNA Sequence, 1:3, 173-180, DOI: 10.3109/10425179109020768，https://freegenes.github.io/genes/BBF10K_003262.html