Machine learning at the edge for AI-enabled multiplexed pathogen detection.

Multiplexed detection of biomarkers in real-time is crucial for sensitive and accurate diagnosis at the point of use. This scenario poses tremendous challenges for detection and identification of signals of varying shape and quality at the edge of the signal-to-noise limit. Here, we demonstrate a robust target identification scheme that utilizes a Deep Neural Network (DNN) for multiplex detection of single particles and molecular biomarkers. The model combines fast wavelet particle detection with Short-Time Fourier Transform analysis, followed by DNN identification on an AI-specific edge device (Google Coral Dev board). The approach is validated using multi-spot optical excitation of Klebsiella Pneumoniae bacterial nucleic acids flowing through an optofluidic waveguide chip that produces fluorescence signals of varying amplitude, duration, and quality. Amplification-free 3× multiplexing in real-time is demonstrated with excellent specificity, sensitivity, and a classification accuracy of 99.8%. These results show that a minimalistic DNN design optimized for mobile devices provides a robust framework for accurate pathogen detection using compact, low-cost diagnostic devices.

Fast custom wavelet analysis technique for single molecule detection and identification.

Many sensors operate by detecting and identifying individual events in a time-dependent signal which is challenging if signals are weak and background noise is present. We introduce a powerful, fast, and robust signal analysis technique based on a massively parallel continuous wavelet transform (CWT) algorithm. The superiority of this approach is demonstrated with fluorescence signals from a chip-based, optofluidic single particle sensor. The technique is more accurate than simple peak-finding algorithms and several orders of magnitude faster than existing CWT methods, allowing for real-time data analysis during sensing for the first time. Performance is further increased by applying a custom wavelet to multi-peak signals as demonstrated using amplification-free detection of single bacterial DNAs. A 4x increase in detection rate, a 6x improved error rate, and the ability for extraction of experimental parameters are demonstrated. This cluster-based CWT analysis will enable high-performance, real-time sensing when signal-to-noise is hardware limited, for instance with low-cost sensors in point of care environments.

Optofluidic Flow-Through Biosensor Sensitivity - Model and Experiment.

We present a model and simulation for predicting the detected signal of a fluorescence-based optical biosensor built from optofluidic waveguides. Typical applications include flow experiments to determine pathogen concentrations in a biological sample after tagging relevant DNA or RNA sequences. An overview of the biosensor geometry and fabrication processes is presented. The basis for the predictive model is also outlined. The model is then compared to experimental results for three different biosensor designs. The model is shown to have similar signal statistics as physical tests, illustrating utility as a pre-fabrication design tool and as a predictor of detection sensitivity.

Optofluidic multiplex detection of single SARS-CoV-2 and influenza A antigens using a novel bright fluorescent probe assay.

The urgency for the development of a sensitive, specific, and rapid point-of-care diagnostic test has deepened during the ongoing COVID-19 pandemic. Here, we introduce an ultrasensitive chip-based antigen test with single protein biomarker sensitivity for the differentiated detection of both severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and influenza A antigens in nasopharyngeal swab samples at diagnostically relevant concentrations. The single-antigen assay is enabled by synthesizing a brightly fluorescent reporter probe, which is incorporated into a bead-based solid-phase extraction assay centered on an antibody sandwich protocol for the capture of target antigens. After optimization of the probe release for detection using ultraviolet light, the full assay is validated with both SARS-CoV-2 and influenza A antigens from clinical nasopharyngeal swab samples (PCR-negative spiked with target antigens). Spectrally multiplexed detection of both targets is implemented by multispot excitation on a multimode interference waveguide platform, and detection at 30 ng/mL with single-antigen sensitivity is reported.

Optofluidic Amplification-free Multiplex Detection of Viral Hemorrhagic Fevers.

Infectious disease outbreaks such as Ebola and other Viral Hemorrhagic Fevers (VHF) require low-complexity, specific, and differentiated diagnostics as illustrated by the recent outbreak in the Democratic Republic of Congo. Here, we describe amplification-free spectrally multiplex detection of four different VHF total RNA samples using multi-spot excitation on a multimode interference waveguide platform along with combinatorial fluorescence labeling of target nucleic acids. In these experiments, we observed an average of 8-fold greater fluorescence signal amplitudes for the Ebola total RNA sample compared to three other total RNA samples: Lake Victoria Marburg Virus, Ravn Marburg Virus, and Crimean-Congo Hemorrhagic Fever. We have attributed this amplitude amplification to an increased amount of RNA during synthesis of soluble glycoprotein in infection. This hypothesis is confirmed by single molecule detection of the total RNA sample after heat-activated release from the carrier microbeads. From these experiments, we observed at least a 5.3x higher RNA mass loading on the Ebola carrier microbeads compared to the Lake Victoria Marburg carrier microbeads, which is consistent with the known production of soluble glycoprotein during infection.

Optofluidic detection of Zika nucleic acid and protein biomarkers using multimode interference multiplexing.

The recent massive Zika virus (ZIKV) outbreak illustrates the need for rapid and specific diagnostic techniques. Detecting ZIKV in biological samples poses unique problems: antibody detection of ZIKV is insufficient due to cross-reactivity of Zika antibodies with other flaviviruses, and nucleic acid and protein biomarkers for ZIKV are detectable at different stages of infection. Here, we describe a new optofluidic approach for the parallel detection of different molecular biomarkers using multimode interference (MMI) waveguides. We report differentiated, multiplex detection of both ZIKV biomarker types using multi-spot excitation at two visible wavelengths with over 98% fidelity by combining several analysis techniques.

Integration of sample preparation and analysis into an optofluidic chip for multi-target disease detection.

Detection of molecular biomarkers with high specificity and sensitivity from biological samples requires both sophisticated sample preparation and subsequent analysis. These tasks are often carried out on separate platforms which increases required sample volumes and the risk of errors, sample loss, and contamination. Here, we present an optofluidic platform which combines an optical detection section with single nucleic acid strand sensitivity, and a sample processing unit capable of on-chip, specific extraction and labeling of nucleic acid and protein targets in complex biological matrices. First, on-chip labeling and detection of individual lambda DNA molecules down to concentrations of 8 fM is demonstrated. Subsequently, we demonstrate the simultaneous capture, fluorescence tagging and detection of both Zika specific nucleic acid and NS-1 protein targets in both buffer and human serum. We show that the dual DNA and protein assay allows for successful differentiation and diagnosis of Zika against cross-reacting species like dengue.