Improving biosensor accuracy and speed using dynamic signal change and theory-guided deep learning.

False results and time delay are longstanding challenges in biosensing. While classification models and deep learning may provide new opportunities for improving biosensor performance, such as measurement confidence and speed, it remains a challenge to ensure that predictions are explainable and consistent with domain knowledge. Here, we show that consistency of deep learning classification model predictions with domain knowledge in biosensing can be achieved by cost function supervision and enables rapid and accurate biosensing using the biosensor dynamic response. The impact and utility of the methodology were validated by rapid and accurate quantification of microRNA (let-7a) across the nanomolar (nM) to femtomolar (fM) concentration range using the dynamic response of cantilever biosensors. Data augmentation and cost function supervision based on the consistency of model predictions and experimental observations with the theory of surface-based biosensors improved the F1 score, precision, and recall of a recurrent neural network (RNN) classifier by an average of 13.8%. The theory-guided RNN (TGRNN) classifier enabled quantification of target analyte concentration and false results with an average prediction accuracy, precision, and recall of 98.5% using the initial transient or entire dynamic response, which is indicative of high prediction accuracy and low probability of false-negative and false-positive results. Classification scores were used to establish new relationships among biosensor performance characteristics (e.g., measurement confidence) and design parameters (e.g., inputs and hyperparameters of classification models and data acquisition parameters) that may be used for characterizing biosensor performance.

Improving the Accuracy of Physics-Based Hydration-Free Energy Predictions by Machine Learning the Remaining Error Relative to the Experiment.

The accuracy of computational models of water is key to atomistic simulations of biomolecules. We propose a computationally efficient way to improve the accuracy of the prediction of hydration-free energies (HFEs) of small molecules: the remaining errors of the physics-based models relative to the experiment are predicted and mitigated by machine learning (ML) as a postprocessing step. Specifically, the trained graph convolutional neural network attempts to identify the "blind spots" in the physics-based model predictions, where the complex physics of aqueous solvation is poorly accounted for, and partially corrects for them. The strategy is explored for five classical solvent models representing various accuracy/speed trade-offs, from the fast analytical generalized Born (GB) to the popular TIP3P explicit solvent model; experimental HFEs of small neutral molecules from the FreeSolv set are used for the training and testing. For all of the models, the ML correction reduces the resulting root-mean-square error relative to the experiment for HFEs of small molecules, without significant overfitting and with negligible computational overhead. For example, on the test set, the relative accuracy improvement is 47% for the fast analytical GB, making it, after the ML correction, almost as accurate as uncorrected TIP3P. For the TIP3P model, the accuracy improvement is about 39%, bringing the ML-corrected model's accuracy below the 1 kcal/mol threshold. In general, the relative benefit of the ML corrections is smaller for more accurate physics-based models, reaching the lower limit of about 20% relative accuracy gain compared with that of the physics-based treatment alone. The proposed strategy of using ML to learn the remaining error of physics-based models offers a distinct advantage over training ML alone directly on reference HFEs: it preserves the correct overall trend, even well outside of the training set.

Reduction of Biosensor False Responses and Time Delay Using Dynamic Response and Theory-Guided Machine Learning.

Here, we provide a new methodology for reducing false results and time delay of biosensors, which are barriers to industrial, healthcare, military, and consumer applications. We show that integrating machine learning with domain knowledge in biosensing can complement and improve the biosensor accuracy and speed relative to the performance achieved by traditional regression analysis of a standard curve based on the biosensor steady-state response. The methodology was validated by rapid and accurate quantification of microRNA across the nanomolar to femtomolar range using the dynamic response of cantilever biosensors. Theory-guided feature engineering improved the performance and efficiency of several classification models relative to the performance achieved using traditional feature engineering methods (TSFRESH). In addition to the entire dynamic response, the technique enabled rapid and accurate quantification of the target analyte concentration and false-positive and false-negative results using the initial transient response, thereby reducing the required data acquisition time (i.e., time delay). We show that model explainability can be achieved by combining theory-guided feature engineering and feature importance analysis. The performance of multiple classifiers using both TSFRESH- and theory-based features from the biosensor's initial transient response was similar to that achieved using the entire dynamic response with data augmentation. We also show that the methodology can guide design of experiments for high-performance biosensing applications, specifically, the selection of data acquisition parameters (e.g., time) based on potential application-dependent performance thresholds. This work provides an example of the opportunities for improving biosensor performance, such as reducing biosensor false results and time delay, using explainable machine learning models supervised by domain knowledge in biosensing.

Physics-Guided Deep Learning for Drag Force Prediction in Dense Fluid-Particulate Systems.

Physics-based simulations are often used to model and understand complex physical systems in domains such as fluid dynamics. Such simulations, although used frequently, often suffer from inaccurate or incomplete representations either due to their high computational costs or due to lack of complete physical knowledge of the system. In such situations, it is useful to employ machine learning (ML) to fill the gap by learning a model of the complex physical process directly from simulation data. However, as data generation through simulations is costly, we need to develop models being cognizant of data paucity issues. In such scenarios, it is helpful if the rich physical knowledge of the application domain is incorporated in the architectural design of ML models. We can also use information from physics-based simulations to guide the learning process using aggregate supervision to favorably constrain the learning process. In this article, we propose PhyNet, a deep learning model using physics-guided structural priors and physics-guided aggregate supervision for modeling the drag forces acting on each particle in a computational fluid dynamics-discrete element method. We conduct extensive experiments in the context of drag force prediction and showcase the usefulness of including physics knowledge in our deep learning formulation. PhyNet has been compared with several state-of-the-art models and achieves a significant performance improvement of 7.09% on average. The source code has been made available*.