Rethinking massive multiplexing in whispering gallery mode biosensing
Abstract
Accurate, label-free quantification of multiple analytes in complex biological media remains a major challenge due to limited multiplexing, signal cross-correlations, and inconsistency across sensor samples and measurement runs. We introduce a multiplexed whispering-gallery-mode (WGM) biosensing framework that overcomes these barriers by jointly advancing photonic integration and data analytics. Our glass-chip platform enables massive, parallelized and flexible multiplexing of >10000 microresonators organized into up to 100 sensing channels, with universal and modular chip design and detection hardware, while maintaining loaded Q-factors of 10^6. Our novel hybrid deep-learning framework BioCCF that integrates domain adaptation with cross-channel fusion enables harmonization of responses across sensing chips and extraction of nonlinear correlations in complex mixtures. Using a highly heterogeneous dataset comprising over 200 hours of sensing data acquired from nine chips with different channel configurations, biological replicates, and repeated regeneration cycles, we demonstrate recalibration-free identification of solution (99.3\% accuracy) and quantification of immunoglobulin G components with relative prediction error of 10^-4 under 5 min. The affordability and modularity of the platform enable distributed data acquisition and aggregation into shared repositories, providing a pathway toward continuously improving model generalization, cross-validation and a scalable, community-driven paradigm for biosensing.
Summary
This paper presents a novel multiplexed whispering-gallery-mode (WGM) biosensing platform that addresses the challenges of accurate, label-free quantification of multiple analytes in complex biological media. The core innovations are a massively parallelized glass-chip platform with >10,000 microresonators organized into up to 100 sensing channels, combined with a hybrid deep-learning framework called BioCCF. BioCCF integrates domain adaptation to harmonize sensor responses across different chips and cross-channel fusion to extract nonlinear correlations from complex mixtures. The platform maintains high loaded Q-factors of 10^6. The authors demonstrate the platform's capabilities using a heterogeneous dataset comprising over 200 hours of sensing data from nine chips with varying channel configurations, biological replicates, and regeneration cycles. They achieve recalibration-free identification of solutions with 99.3% accuracy and quantification of immunoglobulin G (IgG) components with a relative prediction error of 10^-4 in under 5 minutes. The affordability and modularity of the platform facilitate distributed data acquisition and aggregation, paving the way for a scalable, community-driven approach to biosensing. This work tackles the limitations of existing WGM biosensors, which often suffer from limited multiplexing capabilities, signal cross-correlations, and inconsistencies across sensor samples. The AI/ML driven system allows for a sensor platform which can more easily adapt to new analytes and complex mixtures.
Key Insights
- •The glass-chip platform enables massive multiplexing with >10,000 microresonators, significantly exceeding the capabilities of existing WGM biosensors.
- •The BioCCF framework leverages domain adaptation to harmonize responses across different chips, addressing a critical challenge in achieving consistent and reliable biosensing results. TCA domain adaptation achieved an RMSE of 0.037 with 100 microresonators and 0.01 in the "Inf" case.
- •Cross-channel fusion in BioCCF allows for the extraction of nonlinear correlations in complex mixtures, improving the accuracy and robustness of analyte quantification.
- •The system achieves recalibration-free identification of solutions with 99.3% accuracy and quantification of IgG components with a relative prediction error of 10^-4 under 5 min.
- •The platform's modularity and affordability enable distributed data acquisition and aggregation, fostering a community-driven paradigm for biosensing and continuous model improvement.
- •The detection limit (LOD) of the sensor is estimated as ≈6 pM for both IgGR and IgGH using an analytical model.
- •The hybrid DL-framework BioCCF improves the prediction accuracy by at least 4 times compared to more common regressor architectures.
Practical Implications
- •The platform can be used for rapid and accurate quantification of multiple analytes in complex biological samples, enabling applications in personalized medicine, environmental monitoring, and drug discovery.
- •Researchers and engineers can leverage the platform's modular design and data analytics framework to develop customized biosensing solutions for specific applications.
- •The platform's affordability and ease of use make it accessible to a broader range of researchers and practitioners, fostering innovation and collaboration in the field of biosensing.
- •The distributed data acquisition and aggregation capabilities of the platform enable the creation of large, diverse datasets that can be used to train and validate more robust and generalizable AI/ML models.
- •Future research directions include optimizing biorecognition layers, integrating different microresonator geometries and materials, and extending the platform's capabilities to other optical biosensing technologies such as surface plasmon resonance (SPR).