Proof of concept

Our hypothesis is that our designed molecular computing system based on polymerase-mediated strand displacement reactions can perform "weighted"and "mmation"erations on multidimensional biomarker input signals, thereby generating more intuitive low-dimensional assessment results.

We will break down the diagnostic process into several components as depicted in the diagram and intend to validate each of them successfully in a stepwise manner.

Molecular Classifier for Bacterial & Virus Diagnosis

The fluorescence reporter probe responds to the weight molecules through the TMSD reaction

We initially validate whether our designed fluorescence reporter probe can produce a quantitative response to weight molecules.

To perform this validation, we reacted weight molecules with different concentration gradients, along with an excess of fluorescence reporter probes, in a solution. We recorded the final fluorescence intensity in the solution after the reaction was complete and established an equation correlating fluorescence intensity with weight molecule concentration.

After the fluorescence curve stabilized, there was a linear correlation between fluorescence intensity and weight molecule concentration. This demonstrates that the fluorescence reporter probes can produce a linear response to weight molecules at different concentrations, enabling a quantitative conversion from weight molecule signals to fluorescence intensity signals.

Detection and Weighting of Weight Probes for Target Signals

Next, we validate the functionality of weight probes. Weight probes should be able to bind to target molecules through base complementary pairing, thereby initiating a polymerase-mediated strand displacement reaction, resulting in the release of weight molecules corresponding to the target weight number.

We conducted reactions of target molecules at concentrations of 2M, 5nM, and 10nM with an excess of weight probes in a solution containing Bst polymerase, dNTPs, and other components. We then employed the previously validated fluorescent reporter probes to quantitatively respond to the weight molecules displaced from the track strand, monitoring changes in the fluorescence intensity in the solution.

As shown in the figure, the results indicate that after the fluorescence signal stabilizes, the fluorescence intensity in the solution exhibits a linear relationship with both the target weight and the target concentration. The weight probes enable the detection and weighting of target signals.

Summation of the weighted signals

Next, we validated the mixed detection of targets with the same sign of weight. The weight probes should be capable of specific recognition of their corresponding targets while not responding to other targets. The signals from targets with the same sign of weight should be able to undergo weighting and subsequent summation. The summed results are output in the form of fluorescence intensity signals, with separate signals for detecting viral and bacterial infections.

We first prepared a mixture solution containing one or more targets according to the seven possible target mixtures that could occur (each target concentration was 10 nM). These seven target solutions were mixed with a solution containing weight probes, fluorescence reporter probes, Bst DNA polymerase, dNTPs, and so on. The changes in fluorescence intensity of the solution were then monitored.

The results indicate that after the reaction, the fluorescence intensity exhibits a linear relationship with the sum of the weight numbers of the mixed targets. There is no mutual interference between the various targets in terms of binding with the weight probes, and the summation of the weighted signals is achievable.

We initially prepared a set of mixed solutions containing one or more targets out of the 8 possible combinations (each target concentration was 10 nM).
These 8 target solutions were then mixed with a solution containing weight probes, fluorescence reporter probes, Bst polymerase, dNTPs, and other necessary components.
Subsequently, we monitored changes in fluorescence intensity within the solutions.

Subtraction of fluorescence values and making a judgment

We further validate the feasibility of the step "subtracting the results obtained after weighting and summing each target to derive a decisive outcome."

Calculate the difference in fluorescence signals between the two types in each combination (denoted as D). D > X indicates bacterial infection, D < Y indicates viral infection, and Y < D < X indicates a healthy condition.

After subtracting the different fluorescence values, we can obtain a judgment result indicating "bacterial infection,""ral infection," "thy condition."

Molecular Classifier for Lung Adenocarcinoma Diagnosis

Detection and Weighting of Weight Probes for Target Signals

we validate the functionality of weight probes. Weight probes should be able to bind to target molecules through base complementary pairing, thereby initiating a polymerase-mediated strand displacement reaction, resulting in the release of weight molecules corresponding to the target weight number.

We reacted 5nM of the target with an excess of weight probes in a solution containing Bst polymerase, dNTPs, and other components. Subsequently, we utilized the successfully validated fluorescence reporter probe to quantitatively respond to the displaced weight molecules, monitoring the changes in fluorescence intensity in the solution.

As shown in the figure, the results indicate that after the fluorescence signal stabilizes, the fluorescence intensity in the solution exhibits a linear relationship with the target weight. The weight probes enable the detection and weighting of target signals.