Engineering

We had to select biomarker microRNA for glaucoma to use in POIROT. To guide our decision on which biomarkers to choose, we consulted with Prof. Ochiya, a leading expert in miRNA biomarker research in Japan.
For more detail about Human Practices, click here:

Through conversation with Prof. Ochiya, we learned that conducting a meta-analysis on tear fluid miRNAs would not guarantee reliability. MiRNAs in tear fluid have not been extensively researched. Therefore, we decided not to use meta-analysis. Instead, we selected miRNAs that reported as glaucoma biomarkers based on highly reliable papers introduced to us by Prof. Ochiya ¹. After selecting the biomarkers, Dry Lab investigated whether any similar sequences to the chosen biomarkers exist in order to estimate the necessary sequence specificity for the amplification system.

We wrote a program in Python to search for the presence of biomarker-like sequences in tear fluids, using the results of a paper that profiled miRNAs predicted to be present in tear fluids.
For more detail about algorithm, click here:

We investigated whether miRNAs with sequences similar to those of hsa-miR-10b-5p, hsa-miR-375-5p, and hsa-miR-30d-5p, which we selected as biomarkers, were present in tear fluids.

miR-10b-5p

Hsa-miR-10a-5p, hsa-miR-100-5p, hsa-miR-99a-5p, and hsa-miR-99b-5p were found as miRNAs with sequences similar to hsa-miR-10b-5p. MiR-10a-5p consists of 23 mer, same with miR-10b-5p, with only the 12th base being different. MiR-100-5p, miR-99a-5p, and miR-99b-5p each consist of 22 mer, and when the 3' end is aligned with miR-10b-5p, mismatches occur on 4, 5, and 6 bases, respectively.

miR-375

As a result of the search for similar sequences, no miRNAs with sequences similar to hsa-miR-375 were found.

miR-30d-5p

Hsa-miR-30a-5p and hsa-miR-30e-5p were found to be miRNAs with sequences similar to hsa-miR-30d-5p. Both consist of 22 bases, with miR-30a-5p having a mismatch of one base and miR-30e-5p having a mismatch of two bases.

Dry Lab's analysis revealed that sequences similar to the biomarker miRNA used in POIROT are present in tear fluids. If the sequence specificity of the amplification system is low, miRNAs other than the miRNAs that should be amplified will be amplified. This can cause false positives and reduce the reliability of the detection system.
Therefore, high specificity is required for the Amplification System.

We aim to develop an amplification system in which the final dsDNA product contains CRISPR-Cas recognition sequences corresponding to the concentration of the target miRNA. Our first choice was to use a method that can amplify and quantify isothermally, without the use of enzymes.

We utilize the Hybridization Chain Reaction (HCR) ⁴, a method for producing dsDNA from single-stranded nucleic acid. For detection of the reaction products, gel electrophoresis is used. Since the resolution provided by agarose gel electrophoresis was insufficient, we considered using capillary electrophoresis for more precise identification.
The HCR method we use involves three types of DNA with hairpin structures as templates. This method needs annealing operations just before the reaction to induce the formation of hairpin structures. However, considering home use, the complex temperature changes required for annealing would be avoided.
If there is a long interval before starting the reaction, structural changes could occur. If no structural changes occur and the presence or absence of the annealing operation does not affect the results, the method could be considered for home use. However, if structural changes lead to the initiation of the reaction even in the Negative Control (NC), incorporating HCR into POIROT is impossible.
Therefore, we also conducted an investigation on the result without annealing operation.

With annealing operation

We conducted HCR using ssDNA as the target and performed agarose gel electrophoresis to confirm the reaction proceed. While we were able to confirm that hybridization occurred at higher target concentrations, we were unable to effectively distinguish and quantify the various reaction product DNAs. Multiple types of DNAs are expected to be found.

Without annealing operation

Reaction products were observed regardless of the presence or absence of the target. For more detail, see Results.

The detection limit of the MultiNA system is around 100 nM ⁵, making it difficult to observe whether reaction products are being generated at lower concentrations. The peaks obtained from the MultiNA analysis were attributed to nucleic acids that could potentially exist in the system. Taking into account measurement errors and the formation of incomplete dsDNA, we made the following attributions. The reaction products were, at most, oligomers with around n = 20. Even if hybridization occurred at low target concentrations, the amplification efficiency achieved is far below what we require.

In the experiment above, we used DNA that had been annealed just before the reaction. However, without annealing operation, amplification occurred even in the NC. The fact that the reaction proceeded independently of the target suggests a critical problem, as it could lead to false positives. Therefore, incorporating HCR into POIROT is not a practical option.
The lack of amplification efficiency is another issue. To address this shortcoming, we considered connecting HCR with other amplification systems, using HCR in the final stage to produce dsDNA. However, due to the inevitable structural changes in the hairpin, we concluded that the use of HCR itself is impractical for our purposes.

Furthermore, Cas12a and Cas3 are activated to some extent even by ssDNA containing a spacer sequence ⁶. In order for HCR to connect to CRISPR-Cas, the hairpin must have a spacer sequence, and the spacer sequence must be present in the solution at the start of the reaction. This suggests the possibility that CRISPR-Cas would be activated independently with the target miRNA.
For these reasons, enzyme-free systems are decided to be not suitable for use with POIROT, and it is needed to use an amplification mechanism that ensures sufficient amplification efficiency dependent on the target miRNA concentration.
→Cycle_2

We aimed to produce dsDNA by amplifying ssDNA with the same sequence as miRNA (sequence with U replaced by T) using EXPAR (Exponential Amplification Reaction) ⁷, which is expected to have sufficient amplification efficiency, and then performing an extension reaction using one of the target dsDNAs as a template.

Carter, J. G. et al. used EXPAR to obtain amplification products of the target concentration in about 10-20 minutes ⁷. We performed a preliminary experiment using the protocol described in their paper and were able to obtain the expected amplification products.

With annealing operation

Experiments with different sets of target, template, DNA polymerase, and nickase concentration.

The time it takes for amplification to occur under the same conditions has been reduced by about five times.

The shape of the amplification curve was highly dependent on the template concentration.
Regarding DNA polymerase and nickase concentrations, the higher the concentration, the faster the amplification, but above a certain concentration, it became stable. Further experiments were conducted to find the cause and solution of the large change in reaction rate under the same conditions.

Based on several previous studies that have done ODE modeling of EXPAR ^{8, 9, 10}, the ODE model of EXPAR in this project was constructed as follows.

Based on the figure, an ODE was set up and the fluorescence intensity was predicted as shown below.

Based on several previous studies that have done ODE modeling of EXPAR ^{9, 10, 11}, the ODE model of EXPAR in this project was constructed as follows.

A sudden change in amplification rate is undesirable from the view of robustness. To experimentally investigate which parameters affect the amplification rate, we investigate the extent to which amplification is affected by the purification grade of ssDNA and enzyme concentration.

To evaluate the influence of the purification grade of the template and target, experiments were performed using ssDNA purified by OPC (Oligonucleotide-Peptide Conjugates) and PAGE (Higher purification grade than OPC). We also investigated the behavior of the amplification rate when the polymerase and nickase concentrations were lowered after changing the lot.

Wet Lab conducted EXPAR using ssDNA from OPC and PAGE as the target and template, with the other conditions kept. No significant difference in the amplification rate due to the purification grade was confirmed.

When the polymerase concentration was reduced to 1/8 and the nickase concentration to 1/16 of the initial concentration, the amplification rate was almost the same as the initial result, and it was possible to distinguish between the result of 100 fM input and NC.

It was shown that EXPAR is a system that is highly dependent on the enzyme activity. When EXPAR is included in the amplification system, the overall robustness decreases, so we were needed to consider a different amplification system.
In addition, when considering amplification from tears, specificity becomes important. Namely, a system that can clearly distinguish similar sequences is desired. →Cycle_3 - The pursuit of robustness and specificity -

As discussed in Biomarker-Cycle_1, we found a similar sequence with a single base mutation for our target miR-10b-5p, and a similar sequence with a single and two base mutations for miR-30d-5p. Given the possibility that these similar sequences are present in tears, the system requires high sensitivity as well as high specificity.
POIROT uses tears, so the possibility that miRNAs with similar sequences may exist in the system cannot be denied. In other words, not only high sensitivity but also specificity is required. Chen, M. et al. have reported that high specificity for the target can be expected by using TWJ ¹¹. In order to utilize this high specificity in POIROT, we considered using it as the starting point of the amplification system.

We used the mechanism below ¹¹ that targets hsa-let-7b. This system uses SDA starting from the primer produced from TWJ to increase the amplification efficiency, and measures fluorescence using Molecular Beacon (MB).
We used DNA corresponding to let-7b, as well as the subspecies hsa-let-7a, 7c,...,7i, and let-7b with single base mutations at the 1st, 3rd, 5th, 7th, 9th, 11th, 13th, and 15th bases from the 5' end as targets, and investigated what kind of amplification curves would result.
We also performed a similar experiment with SDA that does not create TWJ, and compared the specificities of each.

The results are shown below.

TWJ can completely distinguish between subspecies, and can also distinguish between single-base mismatches. The fluorescence intensity after 40 minutes of incubation is shown in the graph below, confirming the high specificity of TWJ. In the case of single-base mismatches, specificity was slightly lower when mutations were added to the 1, 9, and 11 mers from the 5' end. However, there was a significant difference with let-7b in the case of mutations in the 9 and 11 mers.
The use of TWJ will ensure the sequence specificity of the POIROT target. For a detailed analysis, see TWJ-2cycle.

Similar to NJ-SDA, TWJ Amplification was also modeled to form a ternary complex by one-step binding, and the amplification of let-7b and mismatch-5 was compared. The result that NJ-SDA cannot distinguish between let-7b and mismatch-5 was reproduced, but the result that TWJ-SDA can distinguish between let-7b and mismatch-5 was not reproduced. Therefore, the TWJ-SDA ternary complex was modified to form in two steps and have multiple reaction pathways, and the above experimental results were reproduced.

Even if the model was simplified to have one reaction pathway, the experimental fact that TWJ Amplification can still distinguish between let-7b and mismatch-5 was reproduced. By moving the binding constant as a parameter in this model, it was suggested that the specificity of TWJ-SDA is due to the fact that the binding constant is kept small by the Target, Template, and Helper binding in multiple steps via a reaction intermediate. For more information, see Model_Specificity.

The experimental result that NJ-SDA cannot distinguish between let-7b and mismatch-5 was reproduced, but the experimental result that TWJ-SDA can distinguish between let-7b and mismatch-5 was not reproduced.

This suggests that this model is insufficient to show the characteristics of TWJ Amplification. Therefore, we decided to develop a new model.

Based on helpful advice from Human Practices, we created a new model. In the new model, a TWJ complex is formed through a two-step binding process, in order to more accurately replicate the physical phenomena.

We converted the above model into ODEs and simulated it in Python.

For more information, click here:

In our new model, TWJ Amplification can distinguish between let-7b and mismatch-5, which matches the experimental results obtained from the Wet Lab. What is crucial here is understanding why the change of the model allows us to replicate the specificity of TWJ-SDA Amplification. In order to identify the underlying reasons for the specificity of TWJ-SDA Amplification, we simplified the model to have one reaction pathway. Even if the model was simplified to have one reaction pathway, the experimental fact that TWJ-SDA Amplification can still distinguish between let-7b and mismatch-5 was reproduced. By moving the binding constant as a parameter in this model, it was suggested that the specificity of TWJ-SDA is due to the fact that the binding constant is kept small by the target, template, and helper binding in multiple steps via a reaction intermediate.
In Wet Lab, the results showed that specificity decreased when a mutation was introduced at the end of the target or near the junction structure. A model of binding in multiple stages reproduced these results. This suggests that the reason why specificity decreased when a mutation was introduced at the end or junction is because the thermodynamic stability of the ternary complex decreased. However, in the model, the difference in the predicted location of the mutation was not as significant as the results observed in Wet Lab. Analysis of the model that reproduced the results of Wet Lab revealed that the binding constant is important.

From the analysis of the model, it is expected that when a mutation is introduced near the 5' end (for example, mismatch-5), the binding constant between the template and target, and the binding constant between the target-helper complex and the template. Therefore, \(tt\) (the length of the binding region between the target and the template), \(ht\) (the length of the binding region between the helper and the template), are particularly effective in specificity. In addition, when a mutation occurs near the 3' end (e.g. mismatch-15), the binding constant between the target and helper, and the binding constant between the target and template complex and the helper are expected to be important factors. In other words, the length of the base pair where the target and helper bind, and \(ht\) are expected to be particularly important factors for specificity.

The high specificity of TWJ-SDA (Fig. 6) was demonstrated in Cycle_3-2. However, due to the nature of the reaction in which the target binds to the template and undergoes strand displacement, the amplification efficiency of TWJ-SDA is limited by the target concentration. We used a simulation with the TWJ-SDA ODE model to determine whether the amplification system for this project was sufficient. In Wet Lab experiments, the presence or absence of amplification can be checked by observing the fluorescence, but this modeling is essential because the actual concentration of the amplified product cannot be measured.

We built TWJ ODE model based on the TWJ mechanism ¹¹.

The simulation result with the initial target concentration being 100 fM is below.

It was shown that 40 min of TWJ-SDA only resulted in amplification of approximately 3.0 pM from a target of 100 fM. This alone is far from the amplification efficiency required by POIROT.
From the above, the ODE model suggests that for POIROT, the amplification efficiency of TWJ-SDA is insufficient. Reactions with higher amplification efficiency needed to be combined with TWJ-SDA. A detailed discussion, see Model.

The amplification efficiency of TWJ-Amp is insufficient, and it is necessary to combine TWJ-Amp with a reaction with higher amplification efficiency. Komiya et al. have developed a method to link multiple SDAs to improve amplification efficiency ¹². We call this amplification system that links multiple SDAs as "Multistep-SDA".

We conducted experiments by changing the number of steps using the mechanism and sequence of Komiya et al.'s research ¹². In the paper, no increase in fluorescence intensity is confirmed when the target is not present though the amplification proceeds slowly.
We confirmed the change in amplification efficiency by linking multiple SDAs.
In addition, in the paper, the 3' end of the template is modified with a quencher, but we planned first to conduct experiments without this modification. For each, we linked SDA at various step(s) and conducted experiments at various target concentrations.

As shown in the figure below, it was confirmed that the amplification efficiency increased with each step, but when two or more steps were linked, an increase in fluorescence intensity was confirmed even in the absence of a target.

When 3' end modification was modified with a quencher amplification in the absence of the target was suppressed, as shown in the graph on the right.

Both Wet Lab and Dry Lab confirmed that amplification could be achieved by linking multistep-SDA with a template modified with a quencher at the 3' end. Also, it was confirmed that the amplification efficiency improved with an increase in the number of steps.
Furthermore, Dry Lab simulation confirmed that the robustness was improved compared to EXPAR.
Furthermore, the dry simulation confirmed that the robustness was improved compared to EXPAR.
However, within 200 minutes of observation, amplification was confirmed at 100 pM or more input.
This is inappropriate to be used in the model because it ignores the binding of the template to the 5' end. Based on the wet experimental results, we needed to further bring a more appropriate model and optimize the template design and concentration.
To examine as an amplification system in POIROT, we will consider linking from TWJ and redesign the template to improve amplification efficiency.

In the following, the template for TWJ-SDA is abbreviated as template0, and the templates for subsequent SDA steps are abbreviated as template1, template2, and template3, and the ssDNA output from template-n is called primer-n.
When the template sequence used in Cycle_4A-1 was used as is, it did not work well, and no difference was observed between TWJ alone and when it was linked.
There are two regions that can hybridize for each primer. That is, the 5' end of the template from which the primer is output and the 3' end that is the starting point for the next step. For example, primer1 can hybridize to the 5' end of template1 and the 3' end of template2.
As shown here, it is more stable for each primer to bind to the 5' end of the template in the previous step than to hybridize to the 3' end of the template in the next step.

Considering the above, we started redesign of template1, template2, and template3.

To ensure amplification efficiency, it is necessary to encourage the primer released by strand displacement to hybridize preferentially to the 3' end of the template in the next step.
The sequence was changed as follows.

When an experiment was performed using the redesigned sequence starting from Biomarker 1, the shape of the amplification curve changed as follows

Calculations using NUPACK confirmed that the released primer hybridized more stably with the 3' end of the template in the next step.

By redesigning the template, it was shown that a series of TWJ > Multistep-SDA ligations can distinguish between the result of 1 fM - 100 fM input.
Since TWJ is used as the starting point, this mechanism is considered to be appropriate in terms of specificity. By using this method as an amplification system for POIROT, it will be possible to detect extremely low concentrations of miRNA.

We investigated the effectiveness of Multistep-SDA and experimentally examined whether the product of TWJ-SDA can be used as a target. Moreover, we constructed ODE models for TWJ-EXPAR, TWJ-SDA > 2step-SDA, and TWJ-SDA > 3step-SDA, and compared these amplification systems in terms of amplification efficiency, Positive to Negative Ratio (P/N Ratio), robustness to reaction time and to enzymes activity.

Dry Lab constructed the ODE model of TWJ-SDA > EXPAR, TWJ-SDA > 2step-SDA, and TWJ-SDA > 3step-SDA. The outline of model of TWJ-SDA > EXPAR is shown in the figure below. For more details on the ODE models of other amplification systems, see Model.

The amplification efficiency of each amplification system was predicted as follows. Simulations were conducted for multiple initial template concentrations for each amplification system. The red line indicates the desired product concentration.

First, based on the results of the ODE models for TWJ-SDA > 2step-SDA, it was shown that the amplification efficiency is insufficient. TWJ-EXPAR and TWJ-SDA > 3step-SDA showed enough ampification efficiency for POIROT. TWJ-SDA > 3step-SDA was proposed as the most suitable amplification system for POIROT from the perspective of amplification efficiency and robustness to enzymes. Furthermore, the optimal initial template concentration was proposed for TWJ-SDA > 3step-SDA.
For more details, see Model.

We conducted designing the helper and template of TWJ-SDA. The key factors in sequence design are target specificity and S/N ratio. Dry Lab aimed to predict the conditions for the helpers and templates that achieve target specificity and high S/N ratio using an ODE model. Wet Lab conducted TWJ-SDA using various templates and helpers and measured S/N ratio.

By changing miRNA-complementary sequence, they then recognize other miRNA.

For biomarker 1, 2, and 3, the hybridize lengths of template and target were 10, 11, and 12 bp, and the hybridize lengths of helper and template were 5, 6, 7, and 8 bp, respectively. Wet Lab conducted TWJ-SDA using 3 templates and 4 helpers, and measured S/N ratio.

The results for biomarker1 were plotted below:

for more detail, see Wet_Result_Designed-Sequence.

S/N ratio were shown below:

The larger this ratio, the more clearly the target concentration can be distinguished.
For Biomarker 1, when the condition was 12-5, that is, when the hybridize length between template and target was 12 bp, and the hybridize length between template and helper was 5 bp, the ratio was the largest.
Similarly, it was concluded that 10-5 or 12-7 for Biomarker2 and 12-6 for Biomarker3 would be optimal.

ODE Simulation was conducted using Python.
For detailed information, click Model

ODE simulation showed that sequence specificity in TWJ is achieved when the association constant (\(a_3\)) of the complex between target and helper is less than \(10^3\).

Furthermore, the ODE Model simulation was able to predict helpers and template that result in a high S/N ratio to some extent.

The horizontal axis shows the rank order of large and small S/N Ratio in the simulation and the vertical axis shows the rank order of large and small S/N Ratio in the experiment.

By using the ODE model above, Dry Lab developed software to predict the optimal templates and helpers. This software can assist in the design of templates and helpers for various miRNAs.
For more details about the software, click here:

We aim to complete the flow from Biomarker > TWJ-SDA > Multistep-SDA > Cas. Cas3 or Cas12a is activated by the PAM sequence in dsDNA and the spacer sequence in DNA. It has been reported that even incomplete PAM and spacer sequences can exhibit some collateral activity ³. By using the complementary strand of the spacer as the final template, we designed Cas would be only activated when amplification occurs.
Additionally, for the reasons mentioned in the Cycle_Detection, Cas3 is the most suitable Detection Module for use in POIROT.

We designed the template of the final-step and confirmed in silico that amplification occurs without problems. From the 3' end, the template includes the region complementary to the primer, the Cas recognition site.

We connect to Cas3 at various amplification stages, including only the final stage, from the middle of multistep-SDA to the final stage, and from miRNA to the final stage.

The result was shown below.

When connected ds-amplification > Cas3, the amplification curve showed dependence on target concentration. As the number of amplification steps increased, the amplification of the NC became non-negligible.
By tuning the concentration of template, we were able to distinguish NC and 1 fM or higher.

For more detail about tuning, Please click here:

As the number of amplification steps increases, the collateral activity of CRISPR-Cas independ on the target, becomes more prominent. Suppressing NC is a challenge for the future.
To achieve this, it is necessary to optimize the concentrations of polymerase, nickase, and template. Regarding template concentration, there are many kinds of templates in POIROT. Conducting Wet Lab experiments to independently modify each of them is not practical.
Therefore, it is essential to improve the model by obtaining additional parameters and use it to find the optimal conditions.

We had considered using CRISPR-Cas3 or Cas12a as a detection unit. They exhibit collateral activity and work at 37 °C with dsDNA, suitable for signal amplification and visualization of POIROT. By comparing the collateral activity of Cas3 and Cas12a, we determined which type of Cas to use for POIROT.

We added dsDNA recognized by Cas at various concentrations and tracked the cleavage rate of the FQ probe by measuring fluorescence intensity. These experiments were conducted for both Cas3 and Cas12a to evaluate the cleavage rate of collateral activity.

The result was shown below.

We calculated the cleavage rate from the slope. The slope of the amplification curve represents the cleavage rate of the collateral activity. We demonstrated the cleavage rate is proportional to dsDNA concentration within a certain range.

The cleavage rate is directly proportional to dsDNA concentration within a certain range. The proportional range is wider for Cas3. The cleavage rate for Cas12a is approximately five times faster than that of Cas3 at the same dsDNA concentration.
One of the weaknesses of the amplification system of POIROT is that target independent amplification is a challenge. Therefore, using Cas3 is considered more appropriate than Cas12a, as Cas12a triggers collateral activity even at low concentrations of the final product.

We assembled the LFA with reference to CONAN developed by Yoshimi et al. ³.
The strip used was pre-installed with AuNPs with rb anti FITC antibody on the conjugate pad, streptavidin on the control line (C-line), and anti rb antibody on the test line (T-line). FITC-ssDNA-biotin was used as a reporter ¹³. When CRISPR-Cas3 is activated, the ssDNA within the FITC-ssDNA-biotin is cleaved, producing FITC-labeled reporter fragments and biotin-labeled reporter fragments. Using this, both the T-line and C-line turn red if the test is positive, and only the C-line turns red if the test is negative.
Previous studies have shown that whether the T-line develops color in the same sample can depend on the flow rate and the amount of FITC-ssDNA-biotin added ¹³. In our designed device, it is possible to adjust the flow rate and reaction time.
Based on this, we aim to demonstrate the quantification potential of miRNA concentration (more precisely, the signal amplified from miRNA) by tuning the flow rate.
For details about the device, please click here:

We compared the conditions with and without the addition of 5% Polyethleneglycol (PEG). For each condition, we developed FITC-ssDNA-biotin solution and TE alone for 5 min. The case where FITC-ssDNA-biotin was added corresponds to the negative control, while the case where it was not added corresponds to the positive control.

Without adding PEG, both the C-line and T-line turned red regardless of whether FITC-ssDNA-biotin was added or not. In the case with 5% PEG, only the C-line turned red when uncleaved FITC-ssDNA-biotin was added (corresponding to a negative result), while both the C-line and T-line turned red when FITC-ssDNA-biotin was not added (corresponding to a positive result).
The upper strip had no PEG added to the buffer, and the lower strip had PEG added to the buffer.

It was confirmed that the viscosity of the solution is an important factor for LFA. Under our experimental conditions, adding 5% PEG to increase the viscosity seems to achieve the appropriate flow rate. When implementing POIROT, in addition to adjusting the viscosity, it will be necessary to fine-tune the width of the device's flow path to ensure accurate visualization at home.

To confirm that the combination of amplification, Cas3, and LFA allows for successful visualization for POIROT, we conducted experiments.

Under the conditions tuned in Cycle_5, we performed a comparison between the presence and absence of the target using the 1-step SDA > Cas3 > LFA system.

Under these conditions, fluorescence measurement using the FQ probe successfully distinguished between the presence and absence of the target.

In both cases, T-lines were not turned red.
The upper strip is NC and the lower strip is when primer2 was added.

The factors that prevented the distinction between positive and negative resultsare currently unknown. While the concentration of FITC-ssDNA-biotin and flow rate are likely contributing factors, it is also necessary to investigate further to determine whether sufficient cleavage of FITC-ssDNA-biotin occurred in the first place.
In addition to that, it is known that in many cases, miRNAs that serve as biomarkers are expressed in both glaucoma patients and healthy subjects. To distinguish between healthy individuals and those with the condition based on miRNA concentration, it is necessary to identify not just an "all or none" response, but variations in signal strength.
To achieve this, optimizing factors such as the concentration of FITC-ssDNA-biotin and flow rate is essential.