Rapid, antibiotic incubation-free determination of tuberculosis drug resistance using machine learning and Raman spectroscopy.

Tuberculosis (TB) is the world's deadliest infectious disease, with over 1.5 million deaths and 10 million new cases reported anually. The causative organism Mycobacterium tuberculosis (Mtb) can take nearly 40 d to culture, a required step to determine the pathogen's antibiotic susceptibility. Both rapid identification and rapid antibiotic susceptibility testing of Mtb are essential for effective patient treatment and combating antimicrobial resistance. Here, we demonstrate a rapid, culture-free, and antibiotic incubation-free drug susceptibility test for TB using Raman spectroscopy and machine learning. We collect few-to-single-cell Raman spectra from over 25,000 cells of the Mtb complex strain Bacillus Calmette-Guérin (BCG) resistant to one of the four mainstay anti-TB drugs, isoniazid, rifampicin, moxifloxacin, and amikacin, as well as a pan-susceptible wildtype strain. By training a neural network on this data, we classify the antibiotic resistance profile of each strain, both on dried samples and on patient sputum samples. On dried samples, we achieve >98% resistant versus susceptible classification accuracy across all five BCG strains. In patient sputum samples, we achieve ~79% average classification accuracy. We develop a feature recognition algorithm in order to verify that our machine learning model is using biologically relevant spectral features to assess the resistance profiles of our mycobacterial strains. Finally, we demonstrate how this approach can be deployed in resource-limited settings by developing a low-cost, portable Raman microscope that costs <$5,000. We show how this instrument and our machine learning model enable combined microscopy and spectroscopy for accurate few-to-single-cell drug susceptibility testing of BCG.

Rapid, antibiotic incubation-free determination of tuberculosis drug resistance using machine learning and Raman spectroscopy.

Tuberculosis (TB) is the world's deadliest infectious disease, with over 1.5 million deaths annually and 10 million new cases reported each year. The causative organism, Mycobacterium tuberculosis (Mtb) can take nearly 40 days to culture, a required step to determine the pathogen's antibiotic susceptibility. Both rapid identification of Mtb and rapid antibiotic susceptibility testing (AST) are essential for effective patient treatment and combating antimicrobial resistance. Here, we demonstrate a rapid, culture-free, and antibiotic incubation-free drug susceptibility test for TB using Raman spectroscopy and machine learning. We collect few-to-single-cell Raman spectra from over 25,000 cells of the MtB complex strain Bacillus Calmette Guerin (BCG) resistant to one of the four mainstay anti-TB drugs, isoniazid, rifampicin, moxifloxacin and amikacin, as well as a pan susceptible wildtype strain. By training a neural network on this data, we classify the antibiotic resistance profile of each strain, both on dried samples and in patient sputum samples. On dried samples, we achieve >98% resistant versus susceptible classification accuracy across all 5 BCG strains. In patient sputum samples, we achieve ~79% average classification accuracy. We develop a feature recognition algorithm in order to verify that our machine learning model is using biologically relevant spectral features to assess the resistance profiles of our mycobacterial strains. Finally, we demonstrate how this approach can be deployed in resource-limited settings by developing a low-cost, portable Raman microscope that costs <$5000. We show how this instrument and our machine learning model enables combined microscopy and spectroscopy for accurate few-to-single-cell drug susceptibility testing of BCG.

Combining Acoustic Bioprinting with AI-Assisted Raman Spectroscopy for High-Throughput Identification of Bacteria in Blood.

Identifying pathogens in complex samples such as blood, urine, and wastewater is critical to detect infection and inform optimal treatment. Surface-enhanced Raman spectroscopy (SERS) and machine learning (ML) can distinguish among multiple pathogen species, but processing complex fluid samples to sensitively and specifically detect pathogens remains an outstanding challenge. Here, we develop an acoustic bioprinter to digitize samples into millions of droplets, each containing just a few cells, which are identified with SERS and ML. We demonstrate rapid printing of 2 pL droplets from solutions containing S. epidermidis, E. coli, and blood; when they are mixed with gold nanorods (GNRs), SERS enhancements of up to 1500× are achieved.We then train a ML model and achieve ≥99% classification accuracy from cellularly pure samples and ≥87% accuracy from cellularly mixed samples. We also obtain ≥90% accuracy from droplets with pathogen:blood cell ratios <1. Our combined bioprinting and SERS platform could accelerate rapid, sensitive pathogen detection in clinical, environmental, and industrial settings.

Plasmonic and Electrostatic Interactions Enable Uniformly Enhanced Liquid Bacterial Surface-Enhanced Raman Scattering (SERS).

Surface-enhanced Raman spectroscopy (SERS) is a promising cellular identification and drug susceptibility testing platform, provided it can be performed in a controlled liquid environment that maintains cell viability. We investigate bacterial liquid-SERS, studying plasmonic and electrostatic interactions between gold nanorods and bacteria that enable uniformly enhanced SERS. We synthesize five nanorod sizes with longitudinal plasmon resonances ranging from 670 to 860 nm and characterize SERS signatures of Gram-negative Escherichia coli and Serratia marcescens and Gram-positive Staphylococcus aureus and Staphylococcus epidermidis bacteria in water. Varying the concentration of bacteria and nanorods, we achieve large-area SERS enhancement that is independent of nanorod resonance and bacteria type; however, bacteria with higher surface charge density exhibit significantly higher SERS signal. Using cryo-electron microscopy and zeta potential measurements, we show that the higher signal results from attraction between positively charged nanorods and negatively charged bacteria. Our robust liquid-SERS measurements provide a foundation for bacterial identification and drug testing in biological fluids.

Toward rapid infectious disease diagnosis with advances in surface-enhanced Raman spectroscopy.

In a pandemic era, rapid infectious disease diagnosis is essential. Surface-enhanced Raman spectroscopy (SERS) promises sensitive and specific diagnosis including rapid point-of-care detection and drug susceptibility testing. SERS utilizes inelastic light scattering arising from the interaction of incident photons with molecular vibrations, enhanced by orders of magnitude with resonant metallic or dielectric nanostructures. While SERS provides a spectral fingerprint of the sample, clinical translation is lagged due to challenges in consistency of spectral enhancement, complexity in spectral interpretation, insufficient specificity and sensitivity, and inefficient workflow from patient sample collection to spectral acquisition. Here, we highlight the recent, complementary advances that address these shortcomings, including (1) design of label-free SERS substrates and data processing algorithms that improve spectral signal and interpretability, essential for broad pathogen screening assays; (2) development of new capture and affinity agents, such as aptamers and polymers, critical for determining the presence or absence of particular pathogens; and (3) microfluidic and bioprinting platforms for efficient clinical sample processing. We also describe the development of low-cost, point-of-care, optical SERS hardware. Our paper focuses on SERS for viral and bacterial detection, in hopes of accelerating infectious disease diagnosis, monitoring, and vaccine development. With advances in SERS substrates, machine learning, and microfluidics and bioprinting, the specificity, sensitivity, and speed of SERS can be readily translated from laboratory bench to patient bedside, accelerating point-of-care diagnosis, personalized medicine, and precision health.

Nanophotonic Platforms for Chiral Sensing and Separation.

Chirality in Nature can be found across all length scales, from the subatomic to the galactic. At the molecular scale, the spatial dissymmetry in the atomic arrangements of pairs of mirror-image molecules, known as enantiomers, gives rise to fascinating and often critical differences in chemical and physical properties. With increasing hierarchical complexity, protein function, cell communication, and organism health rely on enantioselective interactions between molecules with selective handedness. For example, neurodegenerative and neuropsychiatric disorders including Alzheimer's and Parkinson's diseases have been linked to distortion of chiral-molecular structure. Moreover, d-amino acids have become increasingly recognized as potential biomarkers, necessitating comprehensive analytical methods for diagnosis that are capable of distinguishing l- from d-forms and quantifying trace concentrations of d-amino acids. Correspondingly, many pharmaceuticals and agrochemicals consist of chiral molecules that target particular enantioselective pathways. Yet, despite the importance of molecular chirality, it remains challenging to sense and to separate chiral compounds. Chiral-optical spectroscopies are designed to analyze the purity of chiral samples, but they are often insensitive to the trace enantiomeric excess that might be present in a patient sample, such as blood, urine, or sputum, or pharmaceutical product. Similarly, existing separation schemes to enable enantiopure solutions of chiral products are inefficient or costly. Consequently, most pharmaceuticals or agrochemicals are sold as racemic mixtures, with reduced efficacy and potential deleterious impacts.Recent advances in nanophotonics lay the foundation toward highly sensitive and efficient chiral detection and separation methods. In this Account, we highlight our group's effort to leverage nanoscale chiral light-matter interactions to detect, characterize, and separate enantiomers, potentially down to the single molecule level. Notably, certain resonant nanostructures can significantly enhance circular dichroism for improved chiral sensing and spectroscopy as well as high-yield enantioselective photochemistry. We first describe how achiral metallic and dielectric nanostructures can be utilized to increase the local optical chirality density by engineering the coupling between electric and magnetic optical resonances. While plasmonic nanoparticles locally enhance the optical chirality density, high-index dielectric nanoparticles can enable large-volume and uniform-sign enhancements in the optical chirality density. By overlapping these electric and magnetic resonances, local chiral fields can be enhanced by several orders of magnitude. We show how these design rules can enable high-yield enantioselective photochemistry and project a 2000-fold improvement in the yield of a photoionization reaction. Next, we discuss how optical forces can enable selective manipulation and separation of enantiomers. We describe the design of low-power enantioselective optical tweezers with the ability to trap sub-10 nm dielectric particles. We also characterize their chiral-optical forces with high spatial and force resolution using combined optical and atomic force microscopy. These optical tweezers exhibit an enantioselective optical force contrast exceeding 10 pN, enabling selective attraction or repulsion of enantiomers based on the illumination polarization. Finally, we discuss future challenges and opportunities spanning fundamental research to technology translation. Disease detection in the clinic as well as pharmaceutical and agrochemical industrial applications requiring large-scale, high-throughput production will gain particular benefit from the simplicity and relative low cost that nanophotonic platforms promise.

Rapid identification of pathogenic bacteria using Raman spectroscopy and deep learning.

Raman optical spectroscopy promises label-free bacterial detection, identification, and antibiotic susceptibility testing in a single step. However, achieving clinically relevant speeds and accuracies remains challenging due to weak Raman signal from bacterial cells and numerous bacterial species and phenotypes. Here we generate an extensive dataset of bacterial Raman spectra and apply deep learning approaches to accurately identify 30 common bacterial pathogens. Even on low signal-to-noise spectra, we achieve average isolate-level accuracies exceeding 82% and antibiotic treatment identification accuracies of 97.0±0.3%. We also show that this approach distinguishes between methicillin-resistant and -susceptible isolates of Staphylococcus aureus (MRSA and MSSA) with 89±0.1% accuracy. We validate our results on clinical isolates from 50 patients. Using just 10 bacterial spectra from each patient isolate, we achieve treatment identification accuracies of 99.7%. Our approach has potential for culture-free pathogen identification and antibiotic susceptibility testing, and could be readily extended for diagnostics on blood, urine, and sputum.

Nanoscopic control and quantification of enantioselective optical forces.

Circularly polarized light (CPL) exerts a force of different magnitude on left- and right-handed enantiomers, an effect that could be exploited for chiral resolution of chemical compounds as well as controlled assembly of chiral nanostructures. However, enantioselective optical forces are challenging to control and quantify because their magnitude is extremely small (sub-piconewton) and varies in space with sub-micrometre resolution. Here, we report a technique to both strengthen and visualize these forces, using a chiral atomic force microscope probe coupled to a plasmonic optical tweezer. Illumination of the plasmonic tweezer with CPL exerts a force on the microscope tip that depends on the handedness of the light and the tip. In particular, for a left-handed chiral tip, transverse forces are attractive with left-CPL and repulsive with right-CPL. Additionally, total force differences between opposite-handed specimens exceed 10 pN. The microscope tip can map chiral forces with 2 nm lateral resolution, revealing a distinct spatial distribution of forces for each handedness.

Grating-flanked plasmonic coaxial apertures for efficient fiber optical tweezers.

Subwavelength plasmonic apertures have been foundational for direct optical manipulation of nanoscale specimens including sub-100 nm polymeric beads, metallic nanoparticles and proteins. While most plasmonic traps result in two-dimensional localization, three-dimensional manipulation has been demonstrated by integrating a plasmonic aperture on an optical fiber tip. However, such 3D traps are usually inefficient since the optical mode of the fiber and the subwavelength aperture only weakly couple. In this paper we design more efficient optical-fiber-based plasmonic tweezers combining a coaxial plasmonic aperture with a plasmonic grating coupler at the fiber tip facet. Using full-field finite difference time domain analysis, we optimize the grating design for both gold and silver fiber-based coaxial tweezers such that the optical transmission through the apertures is maximized. With the optimized grating, we show that the maximum transmission efficiency increases from 2.5% to 19.6% and from 1.48% to 16.7% for the gold and silver structures respectively. To evaluate their performance as optical tweezers, we calculate the optical forces and the corresponding trapping potential on dielectric particles interacting with the apertures. We demonstrate that the enahncement in the transmission translates into an equivalent increase in the optical forces. Consequently, the optical power required to achieve stable optical trapping is significantly reduced allowing for efficient localization and 3D manipulation of sub-30 nm dielectric particles.

Toward efficient optical trapping of sub-10-nm particles with coaxial plasmonic apertures.

Optical trapping using focused laser beams has emerged as a powerful tool in the biological and physical sciences. However, scaling this technique to nanosized objects remains challenging due to the diffraction limit of light and the high power levels required for nanoscale trapping. In this paper, we propose plasmonic coaxial apertures as low-power optical traps for nanosized specimens. The illumination of a coaxial aperture with a linearly polarized plane wave generates a dual optical trapping potential well. We theoretically show that this potential can stably trap dielectric particles smaller than 10 nm in diameter while keeping the trapping power level below 20 mW. By tapering the thickness of the coaxial dielectric channel, trapping can be extended to sub-2-nm particles. The proposed structures may enable optical trapping and manipulation of dielectric particles ranging from single proteins to small molecules with sizes previously inaccessible.