Analyte detection using surface-enhanced raman spectroscopy

Abstract

An example method includes receiving, on a testing surface, a sample; outputting excitation light to the testing surface; and detecting, from the testing surface, scattered light including Raman scattering of the sample on the testing surface. The example method further includes generating at least one spectrum representing an intensity of the scattered light with respect to wavelengths of photons in the scattered light; and determining whether the sample includes at least one analyte by analyzing the at least one spectrum.

Description

ANALYTE DETECTION USING SURFACE-ENHANCED RAMAN SPECTROSCOPYCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of Indian Application No. 202411000282, filed on January 2, 2024, and U.S. Application No. 63 / 641 ,849, filed on May 2, 2024, each of which is incorporated by reference herein in its entirety.TECHNICAL FIELD

[0002] This application relates generally to techniques for detecting analytes using surface-enhanced Raman spectroscopy (SERS), such as techniques for detecting the presence of bacteria using SERS.BACKGROUND

[0003] In 2024, the World Health Organization (WHO) estimated that 600 million people fall ill due to food contamination every year. For some individuals, foodborne illness can result in serious medical complications or even death. Pathogens can be introduced to food products at various stages, particularly during preparation, packaging, and distribution. Entities involved with handling food, such as food distribution companies, can prevent the spread of foodborne illness by regularly testing facilities for the presence of specific pathogens, such as E. coll, Listeria monocytogenes, S. aureus, and Salmonella.

[0004] In various cases, entities can test for the presence of pathogens by obtaining a sample in a food handling environment. These entities send the sample to a remote laboratory that can detect for the presence of the pathogens by various enhancement and detection methodologies including performing a polymerase chain reaction (PCR)-based analysis on the sample. However, this process introduces a significant delay (e.g., hours, days, or weeks) in the identification of the presence of the pathogens. In some cases, food handled in the environment can be contaminated and distributed before the entities are aware of the presence of the pathogens in their facilities. In some cases, these delays can result in individuals being infected with serious foodborne illnesses after consuming the contaminated food.DESCRIPTION OF THE FIGURES

[0005] The following figures, which form a part of this disclosure, are illustrative of described technology and are not meant to limit the scope of the claims in any manner.

[0006] FIG. 1 illustrates an example environment for detecting for the presence of one or more analytes using SERS.

[0007] FIGS. 2A and 2B illustrate examples of hydrophobic surfaces that can be utilized to enhance Raman spectra in order to increase the accuracy of identifying the presence or absence of an analyte in a sample.

[0008] FIG. 3 illustrates an example of utilizing strain engineering to enhance a testing surface for obtaining Raman spectra of samples.

[0009] FIG. 4 illustrates an example environment for training and / or utilizing a predictive model to detect the presence or absence of one or more analytes.

[0010] FIG. 5 illustrates an example process for detecting the presence of an analyte in a sample using a testing surface.

[0011] FIG. 6 illustrates one or more devices configured to perform various operations described herein.

[0012] FIG. 7A illustrates examples of cartridges that can be utilized for analyte testing. FIG. 7B illustrates another example of a cartridge that can be utilized for analyte testing.

[0013] FIG. 8 illustrates the efficacy of a SERS-based technique for detecting the presence and concentration of a virus in a sample.

[0014] FIG. 9 illustrates an example scheme for preparation of a bacterial library.

[0015] FIG. 10 illustrates an example of the detection of E. coli'm urine of patients with urinary tract infections (UTIs).

[0016] FIG. 11 illustrates an example pipeline for better resolution of multiplexed bacterial detection.

[0017] FIG. 12 illustrates an example of a UMAP-based pipeline for better resolution of bacteria in clinical UTI samples.DETAILED DESCRIPTION

[0018] There is a need to develop a rapid and accurate test for pathogens (e.g., in food handling environments) that can be performed at or near the point of sampling.

[0019] Implementations of the present disclosure relate to apparatuses, systems, and methods for collecting samples, detecting the presence of one or more materials in the samples, managing data related to the detection of the material(s) in the samples, and storing the samples for extended periods of time. Various techniques described herein enable label-free detection of the material(s). Moreover, various techniques described herein enable users without specialized training to detect and / or track the material(s) in the samples. In particular cases, the material(s) include potential pathogens (e.g., bacteria) in food samples, but implementations are not so limited.

[0020] In some examples, a sample is collected using a swab. For instance, the swab can be rubbed on a surface, such as the surface of a food preparation or packaging apparatus. In various cases, the swab is connected to the cap of a vial that contains a fluid (e.g., saline) that suspends residue from the surface that has been attached to the swab when the cap is disposed on the vial. Moreover, the vial itself may include a dropper mechanism that outputs a precise volume of the mixture of the fluid and the residue, such as by operation of a button on the vial by a user. In some aspects, the same device dispenses and analyzes the sample. The mixture of the fluid and the residue, for instance, may be referred to as the "sample.”

[0021] In various implementations described herein, the sample is disposed on a testing surface (also referred to as a "substrate,” "testing substrate,” or "chip” and the like) for further processing. The testing surface is suitable for surface- enhanced Raman spectrophotometry (SERS), for example. The testing surface, in various cases, includes a base substrate and various nanoparticles (e.g., plasmonic nanoparticles, such as dendrites, nanofractals, and the like). After the sample (e.g., the precise volume of the sample) is deposited on the testing surface, the sample may be subjected to Raman spectroscopy. In some cases, analysis of the sample on the testing surface can be further enhanced using strain engineering. In various cases, the testing surface is hydrophobic (e.g., superhydrophobic) and / or configured to receive stimuli that enhance its SERS properties.

[0022] For example, the testing surface can include a SERS chip (e.g., silver nano-structure sensor, such as disclosed in Label-Free Spectroscopic SARS-CoV-2 Detection on Versatile Nanoimprinted Substrates by Paria et al., Nano Lett. 2022, May 11 ; 22(9): 3620-27, which is incorporated herein by reference in its entirety. Additional suitable substrates include gold and silver based nano structures. In some embodiments, the testing surface omits aluminum.The testing surface, for instance, may include any SERS substrate described in WO 2023 / 164207, which is incorporated by reference herein in its entirety. For example, the testing surface may include an electrically conductive surface and metal nanofractals (e.g., silver nanofractals). Optionally, the testing surface includes graphene layers.

[0023] In some cases, the testing surface is part of a cartridge. For example, a housing (e.g., containing a polymer) may be at least partially enclosed around the testing surface. Some testing surfaces can be fragile, such that components in the cartridge provide rigidity to the testing surface as well as improving ease of handling of the testing surface. In various cases, multiple cartridges can be stacked together in an efficient manner (e.g., within a testing system). For instance, the housings of the respective cartridges may include grooves or other structures that enable the cartridges to be efficiently and stably stacked in groups. In some cases, the housing prevents dust and other contaminants from being deposited on the testing surface. Accordingly, the sample may be dried and stored in the cartridge for an extended period of time (e.g., days, weeks, months, or even years) and may be accurately tested after long-term storage.

[0024] In some implementations, a device can be utilized to perform Raman spectroscopy (e.g., SERS) on the sample (e.g., dried or undried) deposited on the testing surface. For instance, the device includes at least one light source (e.g., a laser) and at least one photodetector. The light source(s) outputs first photons at a first energy level onto the sample deposited on the testing surface. The sample, in response, outputs second photons of various energy levels to the photodetector(s). The second photons represent Raman scattering of the sample. In various cases, device includes circuitry and / or a processor configured to generate one or more spectra representing the Raman scattering of the sample. In various cases, the testing surface enhances the one or more spectra. According to various implementations of the present disclosure, the device can further detect the presence of the material(s) in the sample by analyzing the one or more spectra. In some examples, one or more external devices (e.g., one or more servers) perform the analysis of the one or more spectra after receiving data indicative of the one or more spectra from the device.

[0025] The one or more devices analyze the one or more spectra using one or more processors. For instance, the processor(s) execute non-transitory instructions stored in memory and / or one or more non-transitory computer- readable media. In some cases, the device(s) utilize a pretrained machine learning (ML) model to detect the presence of the material(s) in the sample. For instance, parameters of the ML model may be optimized based on training data including previously obtained Raman spectra of previous samples as well as ground truth classifications of the presence of the material(s) in the previous samples. Various types of ML models can be utilized for this purpose.

[0026] Techniques described herein can detect various materials, such as bacteria, in a label-free fashion with high accuracy. Moreover, in cases wherein the Raman spectroscopy is performed on-site near the place of sample acquisition, results can be obtained rapidly (e.g., within minutes or hours) after sample acquisition. Accordingly, various implementations of the present disclosure can enable fast and accurate detection of pathogens, such as bacteria, in food distribution and packaging environments, thereby preventing widespread foodborne illness from packaged foods. Moreover, various implementations of the present disclosure can be utilized to test for the materials without damaging the materials themselves, such that samples can be tested, stored, and retested.

[0027] Various implementations of the present disclosure will be described in detail with reference to the drawings, wherein like reference numerals present like parts and assemblies throughout the several views. Additionally, anysamples set forth in this specification are not intended to be limiting and merely set forth some of the many possible implementations.

[0028] FIG. 1 illustrates an example environment 100 for detecting for the presence of one or more analytes using SERS. For instance, a sample 102 is obtained. In some examples, the sample 102 includes residue 104 obtained from a sample surface 106. For instance, the residue 104 is captured by sweeping or otherwise contacting a swab 108 on the sample surface 106. In some implementations, the sample surface 106 is in a food preparation and / or packaging environment. For example, the sample surface 106 has come into contact with food being prepared (e.g., a Zone 1 surface, such as a surface of a utensil, a sink, a food storage bin, etc.), is in close proximity to a food contact surface (e.g., a Zone 2 surface, such as a surface of a machine, floor, or trash can in a food handling environment), a surface that is adjacent to a food handling area but is not designed to come into contact with food being prepared (e.g., a Zone 3 surface, such as a surface of a drain or restroom), or a non-food contact surface in an area remote from a food handling area (e.g., a Zone 4 surface, such as a surface of an office, restroom, or break room). In some cases, the sample surface 106 is a food preparation surface. According to some examples, the swab 108 is configured to receive the residue 104 by coming into contact with a subject (e.g., a human), a fluid obtained from the subject (e.g., a liquid biopsy sample, such as blood, plasma, urine, mucus, saliva, or any combination thereof), or a tissue obtained from the subject (e.g., a tissue biopsy sample). In some cases, the swab 108 comes into contact with a sampling site of the subject, such as a nasal cavity, oral cavity, skin surface, or a wound. For example, the environment 100 may be utilized for diagnostic purposes, such as to determine whether the subject has a pathological infection. Optionally, the residue 104 is obtained without a swab.

[0029] In some implementations, the sample 102 is obtained by introducing the residue 104 to a buffer 110. The buffer 110 includes a fluid, such as an aqueous solution, configured to suspend at least a portion of the residue 104. In some cases, the buffer 110 includes a saline solution. The buffer 110, in various cases, is configured to suspend at least a portion of the residue 104 without substantially lysing cells, viruses, vesicles, or other structures within the residue 104. Accordingly, the structures within the residue 104 are substantially maintained after being suspended within the buffer 110.

[0030] In some cases, the swab 108 and / or buffer 110 are integrated into a disposable kit. For example, the buffer 110 may be prepackaged and contained within a vessel 112, the swab 108 may be physically attached to a cap 114, and the cap 114 may be configured to be removably coupled to the vessel 112. For instance, the vessel 112 and the cap 114 include conformationally complementary threads, the cap 114 is configured to be press-fit onto the vessel 112, or the like. In some examples, when the cap 114 is coupled with the vessel 112, the swab 108 is configured to contact the buffer 110 within the vessel 112. In various cases, the vessel 112 and / or the cap 114 include a polymer material, such as polystyrene. In some cases, the swab 108 includes a porous material, such as cotton, polyester fibers, or the like. The swab 108 may be sterile prior to use.

[0031] In various implementations, the sample 102 containing at least a portion of the residue 104 and / or the buffer 110 is obtained from the vessel 112. For example, a dispenser 116, which is optionally integrated with the vessel 112 and / or the kit, is configured to output a volume of the sample 102 from the interior of the vessel 112. In some examples, the dispenser 116 is a dropper, pipet, or other apparatus configured to dispense a fluid. For example, the dispenser 116 is configured to output a precise volume of the sample 102, such as a precise volume in a range of 1 microliter (piL) to 10 milliliters (mL).

[0032] Various alternative techniques can be used to collect the sample 102. For instance, a microfluidic device may be configured to obtain the sample 102, which can enable precise control and manipulation of small volumes, enabling rapid and automated sample preparation for analysis. In some cases, a filtration technique is utilized to obtain the sample 102. Filtration techniques such as membrane filtration or size-based separation can isolate target analytes from complex sample matrices, followed by concentration for enhanced detection sensitivity. A droplet-based system (e.g., utilizing droplet microfluidics) can be utilized to encapsulate analytes within picoliter to nanoliter-sized droplets, facilitating high-throughput screening and analysis of the sample 102. In some examples, a centrifuge (e.g., a centrifugal microfluidic device) is configured to separate one or more portions of the sample 102 based on their density, mass, or size, prior to analysis. According to some instances, laser ablation techniques can aerosolize at least a portion of the sample 102 prior to analysis.

[0033] The sample 102 is deposited on and / or in a cartridge 118 for further testing. In various cases, the cartridge 118 includes a testing substrate 120 at least partially enclosed in a housing 122. For example, a testing surface 124 of the testing substrate 120 is exposed by a trench in the housing 122. In some cases, the trench is tapered, such that the area of the testing surface 124 that is exposed by the trench is smaller than an area exposed by the trench on the top surface of the housing 122. When the sample 102 is deposited on the testing surface 124, an analyte 126 within the sample 102 is also deposited on the testing surface 124. In some cases, at least a portion of the buffer 110 evaporates from the testing surface 124 after the sample 102 is deposited on and / or in the cartridge 118. In some examples, optical tweezers or laser trapping techniques are utilized to manipulate individual particles (e.g., cells) within the sample 102, allowing for the precise positioning of at least a portion of the sample 102 on the testing surface 124.

[0034] In various cases, components of the environment 100 are used to detect for the presence of the analyte 126 in the sample 102 by detecting the presence of the analyte 126 on the testing surface 124. In various cases, the analyte 126 includes one or more cells. For instance, the analyte 126 includes one or more bacteria, such as Bacillus spp. (e.g., Bacillus cereus), Campylobacter spp., Clostridium spp. (e.g., Clostridium botulinum, Clostridium perfringens, etc.) Escherichia spp. (e.g., E. Coli), Klebsiella spp. (e.g., K. pneumoniae), Listeria spp. (e.g., L. monocytogenes), Mycobacterium spp. (e.g., Mycobacterium tuberculosis), Salmonella spp., Shigella spp., Staphylococcus spp. (e.g., Staphylococcus aureus), Pseudomonas spp. (e.g., Pseudomonas aeruginosa), Vibrio spp. (e.g., Vibrio cholerae), or any combination thereof. The analyte 126 may be, or include, other types of cells, such as parasites Toxoplasma gondii, Cryptosporidium, Giardia, and the like. In some examples, the analyte 126 includes a virus, such as a norovirus, a hepatitis virus (e.g., Hepatitis A), influenza virus, a coronavirus, or any combination thereof. The analyte 126 may include a component of a biological tissue, a microorganism, a biofilm, or any combination thereof. In some cases, the analyte 126 includes one or more biomolecules, such as proteins, peptides, lipids, nucleic acids, amino acids, carbohydrates, metabolites, and the like.

[0035] In some cases, the analyte 126 includes a molecule or chemical. For instance, the analyte 126 may include an organic or medically relevant material, such as a food, a beverage, a pharmaceutical, a biomolecule, a water contaminant, an environmental pollutant, a natural product, a biological sample, a polymer, a forensic sample, an explosive, a petrochemical, or any combination hereof. For instance, the analyte 126 may include nitrosamine, a food additive, a flavor, a nutrient, a pesticide, an herbicide, a drug, an antibiotic, a pharmaceutical, a pollutant (e.g., an organic pollutant), an essential oil, a plant extract, a tetrahydrocannabinol (THC), a cannabinoid (CBD), an herbalmedicine, a synthetic polymer, a plastic, a polymer, a drug of abuse, an explosive material, a hydrocarbon, crude oil, a petroleum product, or the like. In some aspects, the analyte 126 includes an inorganic material, such as a catalyst, a metal ion, a mineral, a nanomaterial, an industrial material, a coordination compound, an environmental contaminant, a crystal, a mineral, or any combination thereof. Examples of the analyte 126 include an inorganic nanoparticle (e.g., metal and / or semiconductor nanoparticle, a quantum dot, a nanowire, a nanocomposite, etc.), a catalytic intermediate, a transition metal, a heavy metal, a silicate, a carbonate, a sulfate, an oxide, a halide, a metal alloy, a coating, a film, a metal complex, a coordination polymer, a chelate, a heavy metal, a metalloid, a pollutant, a microplastic, a natural crystal, a synthetic crystal, a gemstone, a mineral deposit, or any combination thereof.

[0036] A spectrometer 128 is configured to generate spectra 130 based on the sample 102 disposed on the testing surface 124. In various implementations, the spectrometer 128 is configured to generate the spectra 130 by performing Raman spectroscopy on the testing surface 124, such that the spectra 130 include Raman spectra. In various cases, one or more light sources 132 are configured to output excitation light 134 onto the testing surface 124 of the cartridge 118. In various cases, the light source(s) 132 include one or more lasers, one or more light-emitting diodes (LEDs), or a combination thereof. For example, the excitation light 134 includes electromagnetic radiation having one or more frequencies in a range of 1012to 1018Hz. In some cases, the excitation light 134 has a wavelength in a range of about 1 to 800 nanometers (nm). In some cases, the light source(s) 132 are configured to emit the excitation light 134 at gradually changing frequencies with respect to time (e.g., the excitation light 134 may scan through a range of frequencies). In various cases, the excitation light 134 is monochromatic at any particular time point when the excitation light 134 is emitted by the light source(s) 132. In some cases, the light source(s) 132 are configured to scan different positions on the testing surface 124 over time.

[0037] Photons within the excitation light 134 are configured to interact with the sample 102 disposed on the testing surface 124. In various cases, at least a portion of the photons are inelastically scattered by the sample 102 disposed on the testing surface 124, which results in an energy shift of the at least portion of the photons (also referred to as "Raman scattering”). According to some cases, the testing surface 124 is plasmonically active. In some examples, some of the photons within the excitation light 134 are also elastically scattered, without an energy shift (also referred to as "Raleigh scattering”). In various cases, the scattered photons are detected by one or more detectors 136 in the spectrometer 128 as scattered light 138. The scattered light 138, in various cases, includes Raman-scattered photons with various energy levels (e.g., frequencies and / or wavelengths) that are different from the energy levels of the excitation light 134. For example, the Raman-scattered photons have higher and / or lower frequencies than the photons in the excitation light 134. According to various cases, the detector(s) 136 include one or more charge coupled devices (CCDs), InGaAs detectors, CMOS devices, or any combination thereof. Although not specifically illustrated, in some cases, the spectrometer 128 includes one or more lenses, one or more mirrors, one or more gratings, one or more optical filters, or any combination thereof, that are configured to transmit, scatter, focus, absorb, or reflect the excitation light 134 and / or the scattered light 138. According to various implementations, the spectrometer 128 is further configured to generate the spectra 130 based on the detected Raman-scattered photons in the scattered light 138. For instance, the spectra 130 represent the detected intensity of the scattered light 138 with respect to different wavelengths and / or frequencies of the photons in the scattered light 138.

[0038] In various cases, an analyzer 140 is configured to determine whether the analyte 126 is present in the sample 102 by analyzing the spectra 130. According to various examples, characteristics of the spectra 130 (e.g., intensities, peak intensities, wavelengths and / or frequencies at peak intensities, etc.) are indicative of the presence or absence of the analyte 126 on the testing surface 124. Thus, the analyzer 140 may determine whether the analyte 126 is present by identifying the presence and / or absence of these relevant characteristics in the spectra 130. However, depending on the composition of the testing substrate 120 and the testing surface 124, the relevant characteristics of the spectra 130 may be difficult to identify by the analyzer 140. For instance, peaks of the spectra 130 that indicate the presence or absence of the analyte 126 on the testing surface 124 may be relatively small, and difficult to identify in view of noise and artifact within the spectra 130. Accordingly, it can be difficult to for the analyzer 140 to detect the presence of the analyte 126 using Raman spectroscopy with high accuracy (e.g., high sensitivity and / or specificity).

[0039] According to various implementations of the present disclosure, the relevant characteristics of the spectra 130 can be enhanced due to a composition of the testing surface 124. In various cases, the composition of the testing surface 124 enhances the Raman scattering of the analyte 126, which increases the Raman signal in the spectra 130. That is, a signal-to-noise ratio (SNR) of the spectra 130 can be increased based on the structure, composition, and state of the testing surface 124. In various cases, the relevant characteristics in the spectra 130 and associated with the presence of the analyte 126 can be more easily discerned by the analyzer 140, as compared to spectra obtained using other types of testing surfaces. For example, the testing surface 124 may include a surface described in International Publication No. WO 2023 / 164207 or Paria et al., Nano Lett. 2022, May 11 ; 22(9): 3620-27, each of which is incorporated by reference herein in its entirety.

[0040] In various cases, the testing substrate 120 includes a base substrate 142. In some examples, the base substrate 142 includes at least one of silicon, germanium, glass, quartz, a polymer (e.g., polydimethylsiloxane, polyimide, etc.), or a gel. In some cases, the base substrate 142 is flexible. For instance, the base substrate 142 may include polyimide.

[0041] The testing substrate 120, in some implementations, further includes a conductive layer 144 disposed on the base substrate 142. The conductive layer 144 is electrically conductive. In some examples, the conductive layer 144 includes a metal, such as gold, silver, copper, platinum, aluminum, or any combination (e.g., alloy) thereof. In some cases, the conductive layer 144 includes graphene, M0S2, black phosphorus, WS2, or any combination thereof. In various cases, the conductive layer 144 is electrochemically deposited, deposited via Physical Vapor Deposition (PVD), deposited via electron beam deposition, and / or electroplated on a surface of the base substrate 142. In some cases, the conductive layer 144 includes an electrically conductive oxide (e.g., indium tin oxide, zinc oxide, gallium oxide, indium oxide, tin oxide, etc.), a doped semiconductor, a conducting polymer, or any combination thereof. In various cases, the conductive layer 144 is deposited on a flat surface of the base substrate 142. In some cases, the surface of the base substrate 142 includes ridges or is otherwise patterned.

[0042] Nanostructures 146 are disposed on the conductive layer 144 and / or the base substrate 142, in various cases. In various cases, the nanostructures 146 include one or more metals, such as gold, silver, copper, platinum, aluminum, or any combination thereof. In some cases, the nanostructures 146 include at least one noble metal. The nanostructures 146, for instance, have a different material composition than the conductive layer 144. For example, if the conductive layer 144 includes silver, the nanostructures 146 may include gold, or vice versa. In various cases, thenanostructures 146 include at least one dimension that is shorter than 1 micrometer (pm). In some implementations, the nanostructures 146 include at least one dimension that is shorter than 20pm. In some cases, the dimension is greater than 0.001 nanometer (nm). The nanostructures 146, for instance, include fractals, spheres, rods, dendrites, dendrimers, or any combination thereof. For instance, the nanostructures 146 include a branched shape. In various cases, the nanostructures 146 are formed by electrochemical deposition, electroplating, or the like.

[0043] In some cases, an antibody-based immunoassay can selectively capture the analyte 126 from the sample 102. In some cases, the nanostructures 146 are functionalized to enhance capture of the sample 102 and / or the analyte 126. In some cases, the nanostructures 146 are bound to chemical structure configured to specifically bind the analyte 126. In various cases, the functionalized nanostructures 146 cause immobilization of the analyte 126 on at least a portion of the testing surface 124.

[0044] In various examples, the conductive layer 144 and / or the nanostructures 146 enhance the portion of the scattered light 138 produced by Raman scattering from the analyte 126. In some cases, the relevant characteristics in the spectra 130 are further enhanced by applying one or more stimuli to the testing surface 124 (also referred to as "strain engineering”). For example, the stimuli may be configured to control (e.g., change) the shape of the nanostructures 146, thereby enhancing the portion of photons in the scattered light 138 that are Raman scattered by the analyte 126. In some cases, the stimuli include a force applied to the testing substrate 120 and / or the testing surface 124. For example, the testing substrate 120 may be bent, compressed, or stretched while being illuminated with the excitation light 134. For instance, the housing 122 may be configured to apply a force (e.g., tensile force) to the testing substrate 120. In some examples, the stimuli includes a pH difference applied to the testing surface 124. For example, an acidic fluid (e.g., an aqueous solution including greater than a threshold concentration of H+ ions) or a basic fluid (e.g., an aqueous solution including less than a threshold concentration of H+ ions) can be disposed on the testing surface 124 in order to adjust the pH of the environment on the testing surface 124.

[0045] Other characteristics of the testing substrate 120 and / or testing surface 124 may impact the quality of the spectra 130 and may enhance the relevant characteristics of the spectra 130. In various implementations, the testing surface 124 may be hydrophobic and / or superhydrophobic. In various cases, a contact angle of a droplet of water (or an aqueous solution) on the testing surface 124 may have a contact angle that is greater than 90 degrees. In some implementations, the testing surface 124 is superhydrophobic. The increased hydrophobicity of the testing surface 124 decreases an amount of area of the testing surface 124 on which the volume of the sample 102 is disposed, thereby increasing the concentration of the analyte 126 on the testing surface 124. In various cases, a branched structure of the nanostructures 146 can be engineered using self-assembly and / or templating techniques. These hierarchical structures can exhibit multiscale roughness, increasing surface area and creating air pockets that enhance the hydrophobicity of the testing surface 124.

[0046] In some cases, a hydrophobic coating may be disposed on the conductive layer 144, on the nanostructures 146, between the nanostructures 146, or any combination thereof. The hydrophobic coating, for instance, may include a silicone, polydimethylsiloxane (PDMS), polytetrafluoroethylene (PTFE), a fluoropolymer, fluorinated silane, calcium carbonate, or any combination thereof. In some cases, the hydrophobic coating includes nanoparticles or other types of nanostructures, such as titanium oxide nanoparticles, a manganese oxide polystyrene nanocomposite, a zinc oxide polystyrene nanocomposite, a silica nanocoating, carbon nanotubes, or may combination thereof. In some cases, thehydrophobic coating is applied to the conductive layer 144 and / or the nanostructures 146 by spraying a fluid onto the testing surface 124 and initiating polymerization, nanostructure assembly, drying, or any combination thereof, on the fluid. In various cases, the hydrophobic coating itself includes the nanostructures 146. The coating, for instance, can be applied by spin-coating, dip-coating, or spray-coating. In various cases, the hydrophobic coating (optionally including the nanostructures 146) creates a conformal layer with superior adhesion to other layers of the testing substrate 120, resulting in enhanced hydrophobicity.

[0047] According to some cases, hydrophobic groups are bound (e.g., conjugated and / or adsorbed) to at least a portion of the nanostructures 146. 1 n various cases, side chains of the nanostructures 146 (e.g. , side chains of dendrites in the nanostructures 146) undergo chemical conjugation and / or physical adsorption with the hydrophobic groups. For example, the hydrophobic groups include at least one of alkyl groups, alkane groups, alkene groups, alkyne groups, fluorinated groups, phenyl groups, ester groups, nonpolar amino acids, or any combination thereof. In some implementations, the hydrophobic groups include inorganic materials, such as tenorite (CuO), silicon oxide, titanium oxide, aluminum oxide, iron oxide, organic-inorganic hybrid materials, or any combination thereof. In various implementations, the hydrophobic groups are functionalized onto one or more metals (e.g., gold and / or silver) in the nanostructures 146.

[0048] In various implementations, a portion of the scattered light 138 includes scattering (e.g., Raman scattering) from the hydrophobic coating and / or hydrophobic groups. In some cases, the portion of the spectra 130 corresponding to scattering by the hydrophobic coating and / or hydrophobic groups is removed from the spectra 130 (e.g., by a digital filter) by the spectrometer 128 and / or the analyzer 140. In some cases, the hydrophobic coating and / or the hydrophobic groups are selected to have substantially nonoverlapping Raman scattering characteristics with respect to the analyte 126.

[0049] The spectrometer 128 itself may be adapted for automated use. In some implementations, the spectrometer 128 can hold (e.g., store and / or test) multiple cartridges. In some cases, the spectrometer 128 includes multiple (e.g., two) removable racks configured to hold multiple cartridges including the cartridge 118. One rack can be loaded with unused cartridges (e.g., cartridges that have not been loaded with samples, such as the sample 102). Another rack can be used for the storage of used cartridges (e.g., the cartridge 118 after the sample 102 has been disposed on the testing surface 124).

[0050] In some implementations, the spectrometer 128 is configured to autonomously shuttles a cartridge from the rack of unused cartridges into a position within the system where the cartridge 118 deposits a sample on the cartridge and interrogates the sample for the presence and identification of chemical and biological structures. For instance, the dispenser 116 may be integrated with the spectrometer 128. After interrogation, the spectrometer 128 may shuttle the used cartridge into the used cartridge storage rack. The spectrometer 128 can be programmed to sample and test cartridges at user-defined increments, for instance.

[0051] In various cases, the spectrometer 128 is configured to autonomously report results (e.g., spectra including the spectra 130) after each sample is processed and, in some cases, archives the data. For instance, the spectrometer 128 may at least temporarily store the data in memory. This cartridge rack design, in some cases, facilitates a user being able to remove the racks for archive purposes and to place new racks for analysis by the spectrometer 128. The spectrometer 128 can be designed to autonomously collect samples through the use of a robotic arm, the use of anair collector, or through direct liquid sample collection. In various cases, the spectrometer 128 includes various actuators configured to move racks and / or cartridges through the spectrometer 128.

[0052] In various implementations, spectrometer 128 is designed to collect and analyze multiple chemical and biological sample types in an autonomous fashion. For instance, the spectrometer 128 may be prescheduled to analyze samples at a predetermined sampling frequency and / or schedule, which may extend throughout the 24 hours of each day, and / or each day of a 7-day week. In some cases, the spectrometer 128 is configured to be placed in-line at a fixed position for automated sampling and detection, or can be integrated onto an automated platform that has the ability to relocate the system to desired points of sampling. For instance, the spectrometer 128 may be configured to detect and / or modify its position (e.g., within a food processing environment) using geolocation technologies (e.g., global positioning system (GPS)), for instance.

[0053] In some cases, the spectrometer 128 includes a benchtop or portable system. For instance, the spectrometer 128 (and optionally the analyzer 140) be integrated in a single unit which contains all components including power supply, computer, lasers, BLUETOOTH™, near-field communication (NFC), and WI-FI™, and other necessary components to process the cartridge 118. For instance, the spectrometer 128 can communicate with a cloud-based or internal, self-contained sample library to identify samples detected on the cartridge 118. In various cases, the spectrometer 128 is designed to accept only a proprietary cartridge shape, such that it cannot be utilized with 3rd party cartridges or sample acquisition techniques.

[0054] In various cases, the spectra 130 generated by the spectrometer 128 is represented in one or more forms. In some cases, the spectra 130 includes a dataset indicating intensity relative to wavelength of the scattered light 138. In some examples, the spectra 130 is pre-filtered to remove signal that is irrelevant to the presence or absence of the analyte 126 on the testing surface 124. For example, components corresponding to a portion of the scattered light 138 scattered by a hydrophobic coating (and / or the testing substrate 120 itself) on the testing surface 124 may be removed. In some cases, the spectra 130 includes data indicating the wavelengths and / or intensities at one or more intensity peaks detected from the scattered light 138. In some examples, the spectra 130 includes an image (e.g., a two- dimensional (2D) digital image) representative of the dataset.

[0055] According to various implementations, the processing performed by the analyzer 140 may further be enhanced by the use of a predictive model 148. In some examples, the predictive model 148 includes one or more machine learning (ML) models. In some implementations, the predictive model 148 includes at least one supervised model. In some cases, the predictive model 148 includes at least one unsupervised ML model. Various examples of ML models that can be included in the predictive model 148 include transformers, artificial neural networks (ANNs) (e.g., feedforward neural networks, multi-layer perceptrons (MLPs), convolutional neural networks (CNNs), backpropagation models, etc.), nearest neighbor models (e.g., k-nearest neighbor models), support vector machines (SVMs), regression analysis models, clustering models (e.g., k-means clustering models, mean-shift clustering, agglomerative hierarchical clustering, expectation-maximization (EM) clustering, etc.), gradient boosting models, random forest models, or any combination thereof.

[0056] Various techniques can be utilized to optimize parameters of the predictive model 148 in advance of the predictive model 148 receiving the spectra 130, such that the predictive model 148 can identify the relevant characteristics of the spectra 130 associated with the analyte 126 with high accuracy. In various cases, the predictivemodel 148 is pretrained. For example, at least a portion of the predictive model 148 is trained using a supervised learning process. In some cases, at least a portion of the predictive model 148 is trained using an unsupervised learning process. During training, various parameters of the predictive model 148 are optimized based on training data.

[0057] In particular cases, the training data includes multiple spectra that are generated in the form of Raman spectral images (e.g., raw images of obtained Raman spectrum or spectra). The analyzer 140, for instance, performs preprocessing on the spectra 130. In some cases, the preprocessing is in the form of baseline correction and / or noise reduction to enhance data quality. Techniques such as smoothing and filtering can be applied for further noise reduction and to improve the signal-to-noise ratio. In some cases, the preprocessed data can be divided into training, validation, and testing sets, with spectral images converted into suitable input formats for the predictive model 148 (e.g., convolutional neural network (CNN) and residual neural network (e.g., ResNet) models). Data augmentation techniques can also be employed to augment the training dataset.

[0058] For model architectures within the predictive model 148, in particular examples, CNN and / or ResNet architectures can be chosen based on their effectiveness in feature extraction from image data, with CNNs including convolutional layers, pooling layers, and fully connected layers. In contrast, ResNet introduces residual connections to address a vanishing gradient problem. Model training, for instance, can involve tuning hyperparameters like learning rate and batch size to optimize performance, followed by evaluation on validation and testing datasets to assess classification performance using metrics such as accuracy, precision, recall, and F1-score. Additionally, confusion matrices and receiver operating characteristic (ROC) curves may be generated for visualizing classification results and evaluating model robustness. Furthermore, components can express distributed dominant features in the spectral image range, such that localization based deep learning algorithms like 1-D ConvNets, RegNetsm and ResNets can be used to differentiate between analyte positive and negative samples and regression for analyte load (e.g., concentration) calculations.

[0059] The predictive model 148 is configured to generate an analyte indicator 150 based on the spectra 130. For example, the analyte indicator 150 is output by the analyzer 140 and / or the predictive model 148 in response to receiving the spectra 130. In various cases, the predictive model 148 is configured to perform various transformations and / or computations on data within the spectra 130, wherein the transformations and / or computations are defined according to the parameters of the predictive model 148. In various cases, the analyte indicator 150 is generated based on the result of the transformations and / or computations. The analyte indicator 150, for instance, includes data representing a predicted presence or absence of the analyte 126 in the sample 102. In some cases, the analyte indicator 150 indicates an amount (e.g., count, concentration, etc.) of the analyte 126 in the sample 102. In some examples, the analyte indicator 150 includes a probability and / or certainty that the analyte 126 is present (or not present) within the sample 102. In some examples, the analyte indicator 150 includes a predicted presence or absence (or probability or certainty thereof) of one or more types of analytes, including the analyte 126, in the sample 102.

[0060] The analyte indicator 150 can be utilized in various ways. In some examples, the analyte indicator 150 is output to a user (e.g., a user that obtained the residue 104). For example, the user may take action to suspend use of the sample surface 106, such as for food preparation. In some examples, the user is a care provider that can diagnose and / or treat a condition of a subject from which the residue 104 is obtained. For instance, the care provider may administer a treatment (e.g., an antibiotic) to the subject in response to perceiving the analyte indicator 150.

[0061] The environment 100 may be utilized for various purposes. Various implementations of the present disclosure enable testing for one or more of the following purposes, such as food safety (e.g., detection of contaminants, additives, pesticides, and adulterants in food products to ensure quality and safety), biomedical diagnostics (e.g., detection of biomolecules, proteins, DNA, and RNA for disease diagnosis, drug discovery, and personalized medicine diagnostics), drinking water quality (detection of contaminants, additives, pesticides, and microorganisms to ensure quality and safety), clinical diagnostics (point-of-care testing, disease screening, and monitoring of biomarkers for early detection and management of diseases such as cancer, diabetes, and infectious diseases), circulating tumor cell characterization (e.g., identification of tissue origin of circulating tumor cells), environmental monitoring (e.g., identification and quantification of pollutants, heavy metals, pesticides, and toxins in air, water, and soil samples), forensic analysis (e.g., identification of trace evidence, drugs, explosives, and biological samples in criminal investigations), pharmaceutical analysis (e.g., quality control, authentication, and characterization of pharmaceutical ingredients, formulations, and drug delivery systems), detection of materials in supply chain (e.g., detection of trace amount of cross contamination in supply vessels), materials science (e.g., surface analysis, characterization of nanostructures, thin films, coatings, and catalysts for research and development in materials science and nanotechnology), chemical sensing (e.g., detection and quantification of chemical species, including gases, organic compounds, and pharmaceuticals, in industrial processes, environmental monitoring, and security applications), biochemistry and molecular biology (e.g., study of molecular interactions, conformational changes, and biochemical processes at the nanoscale level), nanotechnology (e.g., development of advanced sensors, devices, and diagnostic platforms for nanoscale imaging, manipulation, and sensing applications), identification of organic and inorganic compounds in extraterrestrial / planetary exploration, or any combination thereof.

[0062] FIGS. 2A and 2B illustrate examples of hydrophobic surfaces that can be utilized to enhance Raman spectra in order to increase the accuracy of identifying the presence or absence of an analyte in a sample. FIG. 2A illustrates a testing surface 200 that includes nanostructures 202 functionalized with hydrophobic groups 204. For instance, the testing surface 200 may be the testing surface 124 described above with reference to FIG. 1. In various cases, the nanostructures 202 include at least one metal, such as such as gold, silver, copper, platinum, aluminum, or any combination (e.g., alloy) thereof. The hydrophobic groups 204, in various cases, may include carbon-containing groups, such as alkyl groups, alkane groups, alkene groups, alkyne groups, fluorinated groups, phenyl groups, ester groups, nonpolar amino acids, or any combination thereof. For instance, one or more of the carbon-containing groups include at least 5 carbons. In some cases, the hydrophobic groups 204 include CuO, silicon oxide, titanium oxide, aluminum oxide, iron oxide, organic-inorganic hybrid materials, or any combination thereof. The testing surface 200 further includes a conductive layer 206 and a base substrate 208. For instance, the conductive layer 206 is disposed between the nanostructures 202 and the base substrate 208. In various implementations, the conductive layer 206 includes at least one electrically conductive material, such as a metal and / or graphene. In various cases, the base substrate 208 includes at least one of silicon, germanium, glass, quartz, a polymer (e.g., polydimethylsiloxane, polyimide, etc.), or a gel. In some cases, the base substrate 208 is flexible. For instance, the base substrate 208 may include polyimide.

[0063] FIG. 2B illustrates a testing surface 210 that includes a hydrophobic coating 212. In some examples, the testing surface 210 is the testing surface 124 described above with reference to FIG. 1. The hydrophobic coating 212, for instance, may include a silicone (e.g., PDMS), a fluoropolymer (e.g., PTFE), fluorinated silane, calcium carbonate,titanium oxide nanoparticles, a manganese oxide polystyrene nanocomposite, a zinc oxide polystyrene nanocomposite, a silica nanocoating, carbon nanotubes, or may combination thereof. For example, the hydrophobic coating 212 may be disposed on a conductive layer 214 of the testing surface 210. In various implementations, the conductive layer 214 includes at least one electrically conductive material, such as a metal and / or graphene. In some cases, the hydrophobic coating 212 is disposed on and / or between nanostructures 216 within the testing surface 210. In various cases, the nanostructures 216 include at least one metal, such as such as gold, silver, copper, platinum, aluminum, or any combination (e.g., alloy) thereof. The testing surface 210 further includes a base substrate 218. For instance, the conductive layer 214 is disposed between the nanostructures 216 and the base substrate 218. In various cases, the base substrate 218 includes at least one of silicon, germanium, glass, quartz, a polymer (e.g., polydimethylsiloxane, polyimide, etc.), or a gel. In some cases, the base substrate 218 is flexible. For instance, the base substrate 208 may include polyimide.

[0064] FIG. 3 illustrates an example of utilizing strain engineering to enhance a testing surface 300 for obtaining Raman spectra of samples. For instance, the testing surface 300 is the testing surface 124 described above with reference to FIG. 1.

[0065] Strain engineering, in some implementations, can be performed to further enhance the hotspots. A controlled mechanical or chemical stress to dendrite molecules or assemblies can be induced by specific structural changes. For instance, the plasmonic properties of nanostructures (e.g., dendrite nanofractals) can be enhanced by employing controlled mechanical or chemical stimuli, which may optimize their morphology, surface roughness, aggregation behavior, and surface functionalization. This optimization can lay the foundation for advanced molecular sensing platforms applicable in biosensing, environmental monitoring, and biomedical diagnostics. By subjecting nanostructures to mechanical stress or chemical stimuli, conformational changes in their molecular structure can be induced, potentially altering the arrangement of functional groups or branches and influencing overall morphology and surface characteristics. Controlled strain engineering, for instance, can induce surface roughening or fractal-like features, creating more active sites for plasmonic enhancement and thereby enhancing the sensitivity of analyses utilizing Raman spectra from various testing surfaces described herein.

[0066] Mechanical or chemical strain can promote nanostructure aggregation, forming hierarchical nanostructures with enhanced plasmonic properties, resulting in closely spaced clusters or interparticle gaps that significantly amplify the local electromagnetic field and Raman signal. Strain engineering can also enable selective modification of the surface chemistry of nanostructures, allowing precise control over their interaction with analyte molecules by strategically introducing or manipulating functional groups or ligands. Additionally, dynamic strain engineering approaches can be utilized to achieve reversible changes in structure and properties, employing stimuli-responsive dendrites capable of undergoing reversible conformational changes in response to external triggers like force, pH, temperature, or light, thus tuning plasmonic behavior for on-demand SERS sensing applications.

[0067] In various cases, the testing surface 300 illustrated in FIG. 3 includes a base substrate 302, a conductive layer 304, and nanostructures 306. For instance, the conductive layer 304 is disposed between the base substrate 302 and the nanostructures 306. According to some cases, the nanostructures 306 have branched shapes. For instance, the nanostructures 306 include nanofractals and / or dendrites.

[0068] According to various implementations, the nanostructures 306 are manufactured in a first state (upper box). In the first state, the nanostructures 306 have a first shape. However, when a stimulus 308 is applied to the testing surface 300, the nanostructures 306 transform into a second state (lower box). In the second state, the nanostructures 306 have a second shape.

[0069] The stimulus 308, in some cases, includes a force applied to the testing surface 300. For instance, a force can be applied to the base substrate 302, the conductive layer 304, the nanostructures 306, or any combination thereof. In some cases, the force is a tensile force, a shear force, a normal force, a compressive force, or any combination thereof. In some examples, the force causes twisting, bending, compression, extension, or any combination thereof, of a testing substrate including the testing surface 300.

[0070] In some cases, the stimulus 308 is a chemical stimulus. For instance, a fluid can be disposed on the testing surface 300 in order to alter and / or control a pH in an environment including the testing surface 300. In some cases, an acid is disposed on the testing surface 300. In some examples, a base is disposed on the testing surface 300. The change in pH, for instance, can be utilized to tune the shape of the nanostructures 306.

[0071] In some aspects, the stimulus 308 changes a temperature of the testing surface 300. For instance, the stimulus 308 may heat the testing surface 300 to greater than or equal to a threshold temperature. In some examples, the stimulus 308 includes electromagnetic radiation (e.g., light) applied to the testing surface 300. In various implementations, the stimulus 308 optimizes the shape of the nanostructures 306.

[0072] In various implementations, the second shape of the nanostructures 306 can be utilized to obtain an optimized Raman spectra from the testing surface 300 and / or a sample (not illustrated) disposed on the testing surface 300. In various cases, the optimized Raman spectra includes characteristics that can be readily and accurately identified as indicative of the presence or absence of one or more analytes on the testing surface 300. Accordingly, the stimulus 308 can be used to increase the accuracy of the detection of the analyte(s) in the sample.

[0073] FIG. 4 illustrates an example environment 400 for training and / or utilizing a predictive model 402 to detect the presence or absence of one or more analytes. The predictive model 402 is at least a portion of an analyzer 404. For instance, the predictive model 402 corresponds to the predictive model 148, and the analyzer 404 corresponds to the analyzer 140.

[0074] The analyzer 404 includes a trainer 406 configured to train the predictive model 402 based on training data 408. In particular instances, the trainer 406 is configured to set, modify, optimize, or any combination thereof, various parameters 409 within the predictive model 402 based on the training data 408. In various cases, the predictive model 402 includes one or more ML models.

[0075] The parameters 409, in various implementations, are defined based on a type of at least one model in the predictive model 402. In some cases, the predictive model 402 includes an ANN including an input layer, at least one hidden layer, and an output layer, wherein each layer is configured to perform operations (e.g., convolution and / or cross-correlations) on data input to the layer. For example, the input layer, the hidden layer(s), the output layer, or any combination thereof, are connected to one another in series and / or parallel. In some cases, the output of one layer is input into another layer within the model. In some cases, the parameters 409 include weights, thresholds, scalar values, filters, kernels, or other data objects that define the operations of the layers in the ANN.

[0076] In particular cases, the predictive model 402 includes a clustering model. In some cases, various data points (e.g., defined by the training data 408) are mapped to an n-dimensional feature space. Various clusters of the data points can be defined and associated with class labels. For instance, new data points that are defined within a cluster can be inferred to be within the class label associated with the cluster. According to some cases, the data points and / or the feature space are represented using one or more dimensionality reduction techniques, such as uniform manifold approximation and projection (UMAP), multidimensional scaling, principal component analysis (PCA), or any combination thereof. In some cases, the parameters 409 include the number of neighbors within the feature space that are defined by a class label, a threshold distance in which a cluster is defined, a density of data points used to classify a cluster, locations of clusters within the feature space, or the like.

[0077] In some examples, the predictive model 402 includes a regression analysis model. For instance, the predictive model 402 may be defined by a regression function that indicates a relationship between one or more independent variables and one or more dependent variables. For example, the order of the regression function, coefficients of the regression functions, and the like, are included in the parameters 409.

[0078] In some cases, the predictive model 402 includes one or more decision trees. For instance, the predictive model 402 includes a random forest model. The decision tree(s), for instance, are defined by the parameters 409.

[0079] In some examples, the predictive model 402 includes a PCA model. In various cases, the model defines a group of principal components of unit vectors within a feature space based on a data set (e.g., the training data 408). Various weights of the model, in various cases, are defined by the parameters 409.

[0080] According to various implementations, the trainer 406 is configured to optimize the parameters 409 of the predictive model 402 based on example spectra 410 within the training data 408. The example spectra 410, in various implementations, include representations of Raman spectra obtained from example samples. For example, the example spectra 410 are obtained by illuminating a type of testing surface (e.g., the testing surface 124) with excitation light, and detecting scattered light 138 from various example samples disposed on the testing surface 124. In some cases, the example spectra 410 include spectra obtained from samples having different combinations and / or concentrations of one or more analytes-of-interest. In some examples, the example spectra 410 include spectra obtained from samples that omit the analyte(s)-of-interest. For instance, the analyte(s)-of-interest include cells (e.g., types of bacteria, viruses, or parasitic organisms), molecules (e.g., biomolecules, organic molecules, inorganic substances), or other types of substances described herein.

[0081] The example spectra 410 may be represented in one or more forms of data. In some cases, the example spectra 410 include Raman spectra datasets obtained from a spectrometer, such as datasets defined by intensity relative to wavelength, wavenumber, wavelength shift (e.g., relative to a wavelength of excitation light), or any combination thereof. In some examples, portions of Raman spectra are included within the example spectra 410, such as portions of Raman spectra defined between predetermined ranges of wavelength, wavenumber, or wavelength shift. In some cases, the example spectra 410 indicate one or more peaks within the Raman spectra, such as the wavelengths, wavenumbers, or wavelength shifts associated with greater than a threshold intensity. In some cases, the intensities of the peak(s) are included within the example spectra 410. According to some examples, the example spectra 410 include digital images of at least portions of Raman spectra. In some instances, the example spectra 410 include indications of the shapes of various Raman spectra obtained from the samples.

[0082] In some cases, the trainer 406 optimizes the parameters 409 utilizing an unsupervised learning technique. For example, the trainer 406 may define, within an n-dimensional feature space, data points corresponding to the example spectra 410. In some cases, the trainer 406 is configured to extract features (e.g., features of the peak(s) in the Raman spectra) from the example spectra 410 and to plot them within the feature space. In some examples, the trainer 406 is configured to define clusters of the data points and / or features of the example spectra 410. Each cluster may be associated with a given class that can be used for future classification, for instance. In some cases, each class and / or cluster is associated with features indicative of the presence and / or absence of the analyte(s) of interest.

[0083] According to some cases, the trainer 406 is configured to optimize the parameters 409 utilizing a supervised learning technique. In some cases, the training data 408 includes example indicators 412 that serve as labels (e.g., tags) representing a known presence and / or absence of the analyte(s)-of-interest in the example samples. For instance, each of the example spectra 410 may be associated with a corresponding one of the example indicators 412. In some cases, the example indicators 412 are generated by analyzing the samples using a non-Raman technique, such as mass spectrometry, chromatography, electrochemical detection, immunoassays, PCR tests, histopathological studies, or any combination thereof.

[0084] In various implementations, the trainer 406 is configured to input at least a portion of the example spectra 410 into the predictive model 402 and may receive at least one predicted analyte indicator. The trainer 406 may compute a loss (e.g., determine a discrepancy) between at least a portion of the example indicators 412 and the predicted analyte indicator(s). Further, the trainer 406 may alter the parameters 409 in order to minimize the loss. In various cases, the trainer 406 optimizes the parameters 409 iteratively based on the entire set of the training data 408. In some cases, the parameters 409 are optimized until the loss is below a particular threshold.

[0085] In various cases, the optimized parameters 409 enable the predictive model 402 to identify relevant features (e.g., predictive attributes) of the example spectra 410 that are correlated to, or otherwise associated with, the example indicators 412. For example, the predictive model 402 may be configured to determine that a predetermined pattern of peaks, or the shape of a Raman spectrum, is associated with the presence of a particular type of bacteria in the sample from which the Raman spectrum is obtained.

[0086] Once the predictive model 402 is trained, the analyzer 404 may be configured to classify testing spectra 410 (or a "testing spectrum,” in certain examples). In various cases, the testing spectra 410 are obtained by a spectrometer 416. In some cases, the spectrometer 416 was also utilized to obtain at least one spectrum of the example spectra 410 in the training data 408. The testing spectra 410 includes at least one Raman spectra obtained from a testing sample disposed on a testing surface (e.g., the testing surface 124). In various cases, the testing spectra 410 include the same data format as the example spectra 410 in the training data 408. In some examples, the testing sample is obtained from a surface (e.g., a food preparation surface), a subject (e.g., a patient), or the like.

[0087] The analyzer 404 is configured to input the testing spectra 410 into the predictive model 402. The predictive model 402 is configured to perform one or more computations on data within the testing spectra 410, wherein the computation (s) are defined by the optimized parameters 409. In various cases, the predictive model 402 outputs results of the computation (s) as one or more analyte indicators 414. In various cases, the analyte indicator(s) 414 indicate the predicted presence, or absence, of the analyte(s)-of-interest in the testing example.

[0088] FIG. 5 illustrates an example process 500 for detecting the presence of an analyte in a sample using a testing surface. The process 500 is performed by an entity, which may include a computing device, at least one processor, a medical device, a kit, a portable device, a spectrometer (e.g., the spectrometer 128), an analyzer (e.g., the analyzer 140), or any combination thereof.

[0089] At 502, the entity identifies scattered light including Raman scattering from sample disposed on a testing surface. The sample, for instance, is obtained from a surface and / or a subject. In some examples, the sample includes residue from a surface. In some cases, the sample includes a buffer solution. The testing surface may be optimized for Raman scattering, such as a surface of a SERS substrate. In various implementations, nanostructures (e.g., metal nanostructures) are exposed on the testing surface. The nanostructures, in various cases, are plasmonically active. In various cases, the testing surface is hydrophobic. For instance, a hydrophobic coating (e.g., containing the nanostructures and / or a hydrophobic coating material) is disposed on the testing surface. In some cases, the nanostructures are functionalized with one or more types of hydrophobic groups. According to some cases, a stimulus is applied to the testing surface before and / or during the scattering of the light from the sample disposed on the testing surface. In some cases, the stimulus includes a force applied to a substrate on which the testing surface is disposed, a chemical stimulus (e.g., application of an acidic or basic solution to the testing surface), or a combination thereof. In some cases, the scattered light is detected by a spectrometer present in the same location (e.g., same room, same building, etc.) where the sample was obtained (i.e., at a point-of-sampling location).

[0090] At 504, the entity determines whether the sample contains an analyte by analyzing at least one spectrum representing the scattered light. The analyte, for instance, is a chemical, a molecule, a cell, a material, or any combination thereof. In some cases, the analyte includes at least one of a nitrosamine, a food additive, a flavor, a nutrient, a pesticide, an herbicide, a drug, an antibiotic, a chemical, a pharmaceutical, a protein, a peptide, a lipid, a nucleic acid, an amino acid, a carbohydrate, a bacteria, a virus, a parasite, an essential oil, a plant extract, a tetrahydrocannabinol, a cannabinoid, a polymer, a plastic, an explosive material, a hydrocarbon, crude oil, or a petroleum product. In particular cases, the analyte includes one or more bacteria.

[0091] According to some cases, the at least one spectrum is analyzed using a predictive model. In particular cases, the predictive model includes one or more ML models that are pretrained to identify relevant features of the at least one spectrum that indicate the presence (or absence) of the analyte in the at least one spectrum.

[0092] At 506, the entity performs an action based on whether the sample contains the analyte. For example, the entity may output an indication of whether the sample contains the analyte. In some cases, the entity outputs a recommendation to administer, to a subject from which the sample was obtained, a treatment associated with the presence or absence of the analyte. For instance, if the entity determines that a sample obtained from a subject includes a pathogenic bacterium, the entity may recommend administration of, or administer, an antibiotic that targets the pathogenic bacterium.

[0093] FIG. 6 illustrates one or more devices 600 configured to perform various operations described herein. The device(s) 600 include one or more processor(s) 602. In some implementations, the processor(s) 602 includes a central processing unit (CPU), a graphics processing unit (GPU), both CPU and GPU, or other processing unit or component known in the art.

[0094] The processor(s) 602 is operably connected to memory 604. In various implementations, the memory 604 is volatile (such as random access memory (RAM)), non-volatile (such as read only memory (ROM), flash memory, etc.) or some combination of the two. The memory 604 stores instructions that, when executed by the processor(s) 602, causes the processor(s) 602 to perform various operations. In various examples, the memory 604 stores methods, threads, processes, applications, objects, modules, any other sort of executable instruction, or a combination thereof. In some cases, the memory 604 stores files, databases, or a combination thereof. In some examples, the memory 604 includes, but is not limited to, RAM, ROM, electrically erasable programmable read-only memory (EEPROM), flash memory, or any other memory technology. In some examples, the memory 604 includes one or more of CD-ROMs, digital versatile discs (DVDs), content-addressable memory (CAM), or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to store the desired information and which can be accessed by the processor(s) 602.

[0095] The processor(s) 602 is operably connected to one or more input devices 606 and one or more output devices 608. Collectively, the input device(s) 606 and the output device(s) 608 function as an interface between at least one user and the device(s) 600. The input device(s) 606 is configured to receive an input from a user and includes at least one of a keypad, a cursor control, a touch-sensitive display, a voice input device (e.g., a microphone), a haptic feedback device (e.g., a gyroscope), or any combination thereof. The output device(s) 608 includes at least one of a display, a speaker, a haptic output device, a printer, or any combination thereof. In various examples, the processor(s) 602 causes a display among the input device(s) 606 to visually output various data described herein. In some implementations, the input device(s) 606 includes one or more touch sensors, the output device(s) 608 includes a display screen, and the touch sensor(s) are integrated with the display screen.

[0096] In various implementations, the processor(s) 602 is operably connected to one or more transceivers 610 that transmit and / or receive data over one or more communication networks. For example, the transceiver(s) 610 includes a network interface card (NIC), a network adapter, a local area network (LAN) adapter, or a physical, virtual, or logical address to connect to the various external devices and / or systems. In various examples, the transceiver(s) 610 includes any sort of wireless transceivers capable of engaging in wireless communication (e.g., radio frequency (RF) communication). For example, the communication network(s) includes one or more wireless networks that include a 3rd Generation Partnership Project (3GPP) network, such as a Long Term Evolution (LTE) radio access network (RAN) (e.g., over one or more LTE bands), a New Radio (NR) RAN (e.g., over one or more NR bands), or a combination thereof. In some cases, the transceiver(s) 610 includes other wireless modems, such as a modem for engaging in WIFI®, WIGIG®, WIMAX®, BLUETOOTH®, or infrared communication over the communication network(s).

[0097] In some cases, the device(s) 600 includes an on-board power source 612 configured to output power to one or more components of the device(s) 600. For instance, the power source 612 includes a battery (e.g., a rechargeable battery), a capacitor, or other type of energy storage device. In some cases, the power source 612 includes an energygenerating device, such as a dynamo, a solar panel, a thermoelectric element (e.g., Peltier device), or any combination thereof. In some examples, at least a portion of the device(s) 600 includes a portable device.

[0098] In some implementations, the device(s) 600 include a spectrometer (e.g., the spectrometer 128). For example, the input device(s) 600 may include at least one light source 614 configured to output excitation light and / or at least one detector 616 configured to receive scattered light (e.g., including Raman scattered light from a sample disposedon a testing surface, such as the testing surface 124). In various cases, the memory 604 stores a spectra generator 618. Upon executing instructions in the spectra generator 618, the processor(s) 602 may cause the device(s) 600 to generate spectra based on the scattered light detected by the detector(s) 616.

[0099] In various implementations, the memory 604 further includes instructions for executing an analyzer 620. In various cases, the processor(s) 602, when executing the analyzer 620, cause the device(s) 600 to perform various functions of the analyzer 140 and / or analyzer 404 described herein. For example, the device(s) 600 may perform functions of a trainer 622 configured to optimize parameters of a predictive model 624 based on training data. In some cases, the memory 604 also stores the predictive model 624. Using the predictive model 624, the device(s) 600, for instance, may be configured to detect the presence or absence of one or more analytes-of-interest in the sample disposed on the testing surface, based on the spectra generated by the spectra generator 618.

[0100] FIG. 7A illustrates examples of cartridges that can be utilized for analyte testing. For example, the upper box of FIG. 7A illustrates a cartridge 702 suitable for testing a single sample, and the lower box of FIG. 7A illustrates a cartridge 704 suitable for testing multiple samples.

[0101] The cartridge 702 includes a single well with a depth D and a width S, in which a sample can be disposed. In various cases, D is in a range of 0.1 millimeter (mm) to 4 mm. In some examples, S is in a range of 1 mm to 8 mm. The cartridge 702 includes a housing that includes a top plate 706 and a bottom plate 708 that are adhered (e.g., glued, melted, or the like) together. A hole in the top plate 706 defines the well. In various cases, a gasket 710, a testing substrate 712, and a padding layer 714 are disposed between the top plate 706 and the bottom plate 708. The gasket 710, for instance, is an o-ring containing a compressible material, such as a synthetic rubber, silicone, or other compressible polymer. In various cases, the gasket 710 is disposed within a groove on an underside of the top plate 706 and can be adhered to the top plate 706 prior to assembly. When assembled, the gasket 710 prevents a fluid sample from spreading outside of the well. The testing substrate 712, for instance, has a same or similar structure to the testing substrate 120 described above. A portion of a surface of the testing substrate 712 exposed by the well may be a testing surface. The padding layer 714, for instance, includes a foam. In various cases, the padding layer 714 is configured to absorb forces within the cartridge 702 to prevent damage to the testing substrate 712. In various cases, the cartridge 702 can receive a sample at a point-of-sampling location, such as within a food preparation environment and / or within a clinical environment.

[0102] The cartridge 704 includes a housing with a top plate 716 and a bottom plate 718. The top plate 716 is configured to be attached (e.g., adhered) to the bottom plate 718 during assembly. The top plate 716, for instance, defines multiple wells of the cartridge 704. Each well has the depth D and the width S. Moreover, the overall cartridge 704 has a width W, a length L, and a height T. In some cases, W is in a range of 20 to 100 mm, L is in a range of 40 to 150 mm, and T is in a range of 5 to 20 mm. In various cases, a gasket 720 and a testing substrate 722 are configured to be disposed between the top plate 716 and the bottom plate 718 when the cartridge 704 is assembled. The gasket 720 prevents cross-contamination of fluids between the wells of the cartridge 704, for instance. In various cases, the gasket 720 includes a compressible material. Various testing surfaces of the testing substrate 722 are exposed by the wells and the holes in the gasket 720. In various cases, a spectrometer is configured to obtain spectra of different samples disposed on different testing surfaces in different wells of the cartridge 704. In various cases, the testing substrate 722 has the same or similar structure to the testing substrate 120 described above. In various cases, thecartridge 704 can receive multiple samples at a point-of-sampling location, such as within a food preparation environment and / or within a clinical environment.

[0103] FIG. 7B illustrates another example of a cartridge 724 that can be utilized for testing. The cartridge 724 includes a top plate 726 and a bottom plate 728. The top plate 726 is configured to be attached to the bottom plate 726 during assembly. The top plate 726 includes a well that exposes a testing surface of a testing substrate 730. The testing substrate 730 is disposed between the top plate 726 and the bottom plate 728 when the cartridge 724 is assembled. In various cases, the testing substrate 730 has the same or similar structure to the testing substrate 120 described above.

[0104] Each cartridge 702, 704, or 724 in various implementations, may be designed for detection of chemical and biological samples. In some cases, each cartridge 702, 704, or 724 can include a slotted design for precise placement and orientation into the reader / detection instrument. The cartridge 702, 704, or 724, for instance can be designed and optimized to be analyzed through laser and optical interface. In some cases, each cartridge 702, 704, or 724 is designed to house proprietary, surface-enhanced chips which are coated with a proprietary coating to enhance the acquisition of Raman spectra. Each cartridge 702, 704, or 724, in some cases, may have an anti-tamper design that can render each cartridge 702, 704, or 724 and internal chip useless if attempts are made to manipulate the each cartridge 702, 704, or 724 outside of standard operating procedures. Each cartridge 702, 704, or 724 may be designed, for instance, with a removable seal which protects the internal chip from contamination prior to use. Each cartridge 702, 704, or 724 may include a barcode, quick-response (QR) code, or radio-frequency identification (RFID) tag that encodes a unique identifier associated with the cartridge. For instance, the QR code may be printed on an exterior surface of the cartridge 702, 704, or 724.FIRST EXPERIMENTAL EXAMPLE

[0105] FIG. 8 illustrates the efficacy of a SERS-based technique for detecting the presence and concentration of a virus in a sample. Saliva samples with four different SARS-CoV-2 concentrations were obtained. Each sample was analyzed using a real-time PGR (RT-PCR)-based technique (left box of each quadrant) and a SERS-based technique (right box of each quadrant) as described herein. Both techniques could accurately identify that the virus was not present in a control sample (upper left quadrant), and accurately identified the presence of the virus in a sample with 102copies of the virus (lower left quadrant), in a sample with 105copies of the virus (upper right quadrant), and in a sample with 107copies of the virus (lower right quadrant).SECOND EXPERIMENTAL EXAMPLE

[0106] The present Experimental Example relates to techniques including a system for detecting multiple bacteria in a human biological samples and / or in food materials. The system includes a pathogen sensor with a plasmonically active sensing region. The Experimental Example describes various methods for label-free detection to perform multiplexing for identification and differentiation of specific infections based on applying an ML-based algorithm on surface enhanced Raman spectra acquired on a nano fractal SERS substrate. Techniques described herein can be used to detect, in a label-free fashion, multiple bacteria using a nanofractal SERS substrate and ML-based algorithm in food materials. The Experimental Example also describes methods for creating bacterial identification Raman libraries to use as references for bacteria identification.

[0107] Creation of Raman Spectral Library for Bacteria Such as E. coli, Klebsiella, and Pseudomonas aeruginosa. FIG. 9 illustrates an example scheme for preparation of a bacterial library for the detection of bacteria in a sample. This scheme includes the isolation of class activators specific to each bacterium and annotation using a predetermined algorithm, as well as a supervised random sampling by ML to enhance the classification of E. coli, virus, and cell lysates

[0108] The production of pure bacterial cultures for E. coli, Klebsiella, and Pseudomonas aeruginosa were initiated, as illustrated in FIG. 9. The procedure of this Example involves depositing ten microliters of varying bacterial concentrations onto a nanofractal SERS substrate, followed by the collection of spectra for each bacterial species. Class activators specific to each bacterium are isolated and annotated using a computer algorithm. Subsequently, a supervised random sampling machine learning approach is employed to enhance the classification of E. coli, virus, and cell lysates. As the investigation expanded to encompass Klebsiella and Pseudomonas aeruginosa within the analysis group, the algorithm demonstrated its capability to accurately classify these species.

[0109] Detection of E. coli in urine of patients with urinary tract infection (UTI). FIG. 10 illustrates an example of the detection of E. coli in urine of patients with UTIs. The data points on the graph are denoted by markers, representing five serial dilutions (1 : 10) of a UTI-positive patient's urine samples, with one spot for the undiluted sample. Additionally, some markers represent five additional UTI-positive patients' samples. Ten microliters of urine from both UTI patients and healthy subjects were directly applied to a nanofractal substrate, and Raman spectra were subsequently collected. The E. coli class activators obtained were compared with an existing library that encompasses E. coli, Klebsiella, Pseudomonas aeruginosa, SARSCoV-2 virus in Vera cells, and Vera cell lysates. Through matching, it was observed that in the absence of E. coli in the urine, a distinct cluster for E. coli emerged in supervised modeling using UMAP clustering. Both E. coli clusters were notably different from the virus cluster, as illustrated in FIG. 10. The data points on the graph are denoted by markers representing five serial dilutions (1 :10) of a UTI-positive patient's urine samples, with one spot for the undiluted sample. Additionally, some markers illustrated in FIG. 9 represent five additional UTI-positive patients' samples. Remarkably, positive samples and their dilutions formed separate clusters that aligned with the existing data of E. coli present in the library.

[0110] Assessment of UTI Bacterial Library for Multiplexing. FIG. 11 illustrates an example pipeline for better resolution of multiplexed bacterial detection. FIG. 11 includes a data analysis performed with Raman spectra obtained from tabletop Raman spectrometer, a data analysis performed with Raman spectra obtained from portal Raman spectrometer, and a data analysis using both Raman spectra. Mixtures of pure cultures of E. coli, Klebsiella, and Pseudomonas aeruginosa were prepared on three different dates, and Raman spectra were acquired using two distinct machines— an on-table Raman spectrometer (Renishaw) and a custom-made portable Raman spectrometer. To enhance benchmarking, a newly refined algorithm was implemented, offering improved resolution. Utilizing three- component supervised nonlinear machine learning models (SVMs), we successfully distinguished all three bacteria within the mixture of E. coli, Klebsiella, and Pseudomonas aeruginosa. In FIG. 11 , the measurements conducted on the tabletop machine effectively resolved E. coli, Klebsiella, and Pseudomonas aeruginosa. Notably, the resolution of all three bacteria remained consistent when the analysis was replicated using the portable machine, as demonstrated in FIG. 11. Combining and analyzing data from both machines in tandem revealed consistent trends and a uniform level of differentiation, as illustrated in FIG. 11.

[0111] Assessment of UTI Bacterial Library for Multiplexing in UTI patients. FIG. 12 illustrates an example of a UMAP-based pipeline for better resolution of bacteria in clinical UTI samples. Urine from UTI patients were spotted on the nanofractal substrate and measured using table-top Raman Spectroscope. All UTI samples were culture positive for infection of E. coli, Klebsiella, and Pseudomonas aeruginosa. To enhance resolution, a newly refined UMAP supervised machine learning algorithm was implemented using already created bacterial library as label. Utilizing three- component supervised nonlinear machine learning models, the results successfully distinguished all three bacteria within the urine of UTI patients (see FIG. 12).

[0112] The proposed system, outlined in this Experimental Example, holds substantial industrial applicability across diverse sectors. The system's relevance in medical diagnostics is evident, offering a reliable label-free method for detecting multiple bacteria in human biological samples. Its specific application in identifying organ- specific infections, exemplified by its use in UTIs, underscores its potential impact on swift and accurate disease diagnosis.

[0113] Moreover, the applicability extends to the food industry, where the system provides a valuable tool for detecting bacteria in food materials without the need for labeling. This aligns with regulatory standards and addresses critical aspects of food safety and quality control.

[0114] The portability of the system can be a significant advantage for on-site diagnostics, making it applicable in diverse industrial settings. This feature can be particularly valuable in scenarios where immediate and decentralized detection of bacteria is preferred. Furthermore, the methods for creating a bacterial identification Raman library described herein represent a valuable contribution, serving as a reference for bacteria identification. This library has utility across various industries where bacterial identification is integral to quality assurance.EXAMPLE CLAUSES

[0115] The following clauses provide various examples of the present disclosure:1. A method, including: receiving, on a testing surface, a sample; outputting excitation light to the testing surface; detecting, from the testing surface, scattered light including Raman scattering of the sample on the testing surface; generating at least one spectrum representing an intensity of the scattered light with respect to wavelengths of photons in the scattered light; and determining whether the sample includes at least one analyte by analyzing the at least one spectrum.2. The method of clause 1 , wherein the sample includes a sample obtained from a subject.3. The method of clause 2, wherein the sample includes blood, plasma, urine, saliva, or mucus.4. The method of any of clauses 1 to 3, wherein the sample includes a food sample.5. The method of any of clauses 1 to 4, wherein the sample includes residue from a surface.6. The method of any of clauses 1 to 5, wherein the sample includes a buffer solution.7. The method of any of clauses 1 to 6, wherein a testing substrate including the testing surface includes: nanostructures disposed on a base substrate.8. The method of clause 7, wherein the base substrate includes at least one of silicon, germanium, glass, quartz, a polymer, or a gel.9. The method of clause 7 or 8, wherein an electrically conductive layer including at least one of a metal or graphene is disposed between the base substrate and the nanostructures.10. The method of clause 9, wherein the metal includes at least one of gold, silver, copper, platinum, or aluminum.11 . The method of any of clauses 7 to 10, wherein the nanostructures include nanofractals and / or dendrites.12. The method of any of clauses / to 11, wherein the nanostructures include at least one of gold, silver, copper, platinum, or aluminum.13. The method of any of clauses 7 to 12, wherein the nanostructures include plasmonically active nanostructures.14. The method of any of clauses 7 to 13, wherein the nanostructures include hydrophobic groups.15. The method of clause 14, wherein the hydrophobic groups include at least one of an alkyl group, an alkane group, an alkene group, an alkyne group, a fluorinated group, a phenyl group, an ester group, or a nonpolar amino acid.16. The method of clause 14 or 15, wherein a portion of the Raman scattering from the hydrophobic groups has a different peak wavelength than a portion of the Raman scattering from the at least one analyte.17. The method of any of clauses 7 to 16, further including: modifying a structural conformation of the nanostructures by applying a stimulus to a testing substrate including the testing surface.18. The method of clause 17, wherein the stimulus includes a chemical stimulus that changes a pH of an environment including the sample disposed on the testing surface.19. The method of clause 17 or 18, wherein the stimulus includes a compressive force, a tensile force, a shear force, or a normal force.20. The method of any of clauses 7 to 19, wherein the testing substrate further includes a hydrophobic coating.21. The method of any of clauses 1 to 20, wherein the testing surface is hydrophobic.22. The method of any of clauses 1 to 21, wherein outputting the excitation light to the testing surface includes illuminating the testing surface with a laser emitting photons having at least one wavelength in a range of about 400 nanometers (nm) to about 800 nm.23. The method of any of clauses 1 to 22, wherein outputting the excitation light and detecting the scattered light is performed by a spectrometer at a point-of-sampling location.24. The method of any of clauses 1 to 23, wherein the analyte includes at least one of a nitrosamine, a food additive, a flavor, a nutrient, a pesticide, an herbicide, a drug, an antibiotic, a chemical, a pharmaceutical, a protein, a peptide, a lipid, a nucleic acid, an amino acid, a carbohydrate, a bacteria, a virus, a parasite, an essential oil, a plant extract, a tetrahydrocannabinol, a cannabinoid, a polymer, a plastic, an explosive material, a hydrocarbon, crude oil, or a petroleum product.25. The method of any of clauses 1 to 24, wherein the analyte includes one or more bacteria.26. The method of clause 25, wherein the one or more bacteria include at least one Bacillus spp. bacterium, at least one Campylobacter spp. bacterium, at least one Clostridium spp. bacterium, at least one Escherichia spp. bacterium, at least one Klebsiella spp. bacterium, at least one Listeria spp. bacterium, at least one Mycobacterium spp. bacterium, at least one Salmonella spp. bacterium, at least one Shigella spp. bacterium, at least one Staphylococcus spp. bacterium, at least one Pseudomonas spp. bacterium, or at least one Vibrio spp. bacterium.27. The method of clause 25, wherein the one or more bacteria include E. coli, K. pneumoniae, and P. aeruginosa.28. The method of any of clauses 1 to 27, wherein determining whether the sample contains the analyte by analyzing the at least one spectrum includes: removing, from the at least one spectrum, a baseline spectrum associated with the testing surface; and in response to removing, from the at least one spectrum, the baseline spectrum associated with the testing surface, determining whether the at least one spectrum includes at least one characteristic associate with the analyte.29. The method of any of clauses 1 to 28, wherein determining whether the sample contains the analyte by analyzing the at least one spectrum includes: inputting the at least one spectrum into a predictive model configured to identify predictive attributes of the spectrum that are associated with the analyte.30. The method of clause 29, wherein the predictive model includes at least one machine learning (ML) model.31 . The method of clause 30, wherein the at least one ML model includes at least one of a transformer, an artificial neural network, a support vector machine (SVM), a nearest neighbor model, a regression analysis model, a clustering models, a gradient boosting model, or a random forest.32. The method of clause 30 or 31, wherein the at least one ML model includes an SVM and a uniform manifold approximation and projection (UMAP) clustering model.33. The method of any of clauses 30 to 32, further including: training the at least one ML model by optimizing parameters of the at least one ML using training data.34. The method of clause 33, wherein the training data includes a library of spectra obtained from training samples, a portion of the training samples including the at least one analyte, a portion of the training samples omitting the at least one analyte.35. The method of any of clauses 1 to 34, wherein determining whether the sample contains the at least one analyte by analyzing the at least one spectrum includes: determining whether the sample contains a first analyte by analyzing the at least one spectrum; and determining whether the sample contains a second analyte by analyzing the at least one spectrum.36. The method of clause 35, wherein the first analyte includes a first bacteria, and wherein the second analyte includes a second bacteria.37. The method of any of clauses 1 to 36, further including: outputting an indication of whether the sample contains the at least one analyte.38. The method of any of clauses 1 to 37, further including: administering, to a subject, a treatment for a condition associated with the at least one analyte, wherein the sample is obtained from the subject.39. The method of clause 38, wherein the treatment includes an antibiotic.40. A testing substrate for surface enhanced Raman spectroscopy (SERS), the testing substrate including: a base substrate; an electrically conductive layer disposed on the base substrate; and nanostructures disposed on the electrically conductive layer, wherein a surface of the testing substrate is hydrophobic.41 . The testing substrate of clause 40, wherein the base substrate includes at least one of silicon, germanium, glass, quartz, a polymer, or a gel.42. The testing substrate of clause 40 or 41, wherein the base substrate is flexible.43. The testing substrate of clause 42, wherein the base substrate includes polyimide.44. The testing substrate of any of clauses 40 to 43, wherein the electrically conductive layer includes at least one of a metal or graphene.45. The testing substrate of clause 44 wherein the metal includes at least one of gold, silver, copper, platinum, or aluminum.46. The testing substrate of any of clauses 40 to 45, wherein the nanostructures include nanofractals and / or dendrites.47. The testing substrate of any of clauses 40 to 46, wherein the nanostructures include at least one of gold, silver, copper, platinum, or aluminum.48. The testing substrate of any of clauses 40 to 47, wherein the nanostructures include plasmonically active nanostructures.49. The testing substrate of any of clauses 40 to 48, wherein the nanostructures include hydrophobic groups.50. The testing substrate of clause 49, wherein the hydrophobic groups include at least one of an alkyl group, an alkane group, an alkene group, an alkyne group, a fluorinated group, a phenyl group, an ester group, or a nonpolar amino acid.51 . The testing substrate of any of clauses 40 to 50, further including: a hydrophobic coating disposed on the testing surface.52. The testing substrate of clause 51 , wherein the hydrophobic coating includes at least one of a silicone, polydimethylsiloxane (PDMS), polytetrafluoroethylene (PTFE), a fluoropolymer, fluorinated silane, calcium carbonate, titanium oxide nanoparticles, a manganese oxide polystyrene nanocomposite, a zinc oxide polystyrene nanocomposite, a silica nanocoating, carbon nanotubes, or the nanostructures.53. A cartridge including: the testing substrate of any of clauses 40 to 52; and a housing at least partially enclosing the testing substrate, at least one well of the housing exposing the testing surface.54. The cartridge of clause 53, wherein the housing includes a radio-frequency identification (RFID) tag.55. The cartridge of clause 53 or 54, wherein a QR code is printed on the housing.56. A kit, including: the cartridge of any of clauses 53 to 55.57. The kit of clause 56, further including a vessel configured to contain a sample.58. The kit of clause 57, further including: a swab configured to collect residue from a surface and to deposit the residue in the vessel, the sample including the residue.59. The kit of clause 58, further including: a cap configured to be removably coupled with the vessel and to be coupled to the swab.60. A system including: the cartridge of any of clauses 53 to 59; and a spectrometer including: at least one light source configured to emit excitation light to the surface; at least one detector configured to detect scattered light from a sample disposed on the surface; and at least one processor configured to generate at least one spectrum based on the scattered light61 . The system of clause 60, wherein the at least one processor is further configured to determine whether an analyte is present in the sample by analyzing the at least one spectrum.62. The system of clause 60 or 61, wherein the spectrometer is a portable device.CONCLUSION

[0116] All publications, patents, and patent applications mentioned in this specification are herein incorporated by reference in their entirety to the same extent as if each individual publication, patent, or patent application was specifically and individually indicated to be incorporated by reference in its entirety. In the event of a conflict between a term herein and a term in an incorporated reference, the term herein controls.

[0117] The features disclosed in the foregoing description, or the following claims, or the accompanying drawings, expressed in their specific forms or in terms of a means for performing the disclosed function, or a method or process for attaining the disclosed result, as appropriate, may, separately, or in any combination of such features, be used for realizing implementations of the disclosure in diverse forms thereof.

[0118] As will be understood by one of ordinary skill in the art, each implementation disclosed herein can comprise, consist essentially of or consist of its particular stated element, step, or component. Thus, the terms "include” or "including” should be interpreted to recite: "comprise, consist of, or consist essentially of.” The transition term "comprise” or "comprises” means has, but is not limited to, and allows for the inclusion of unspecified elements, steps, ingredients, or components, even in major amounts. The transitional phrase "consisting of” excludes any element, step, ingredient or component not specified. The transition phrase "consisting essentially of' limits the scope of the implementation to the specified elements, steps, ingredients or components and to those that do not materially affect the implementation. As used herein, the term "based on” is equivalent to "based at least partly on,” unless otherwise specified.

[0119] Unless otherwise indicated, all numbers expressing quantities, properties, conditions, and so forth used in the specification and claims are to be understood as being modified in all instances by the term "about.” Accordingly, unless indicated to the contrary, the numerical parameters set forth in the specification and attached claims are approximations that may vary depending upon the desired properties sought to be obtained by the present disclosure. At the very least, and not as an attempt to limit the application of the doctrine of equivalents to the scope of the claims, each numerical parameter should at least be construed in light of the number of reported significant digits and by applying ordinary rounding techniques. When further clarity is required, the term "about” has the meaning reasonably ascribed to it by a person skilled in the art when used in conjunction with a stated numerical value or range, i.e., denoting somewhat more or somewhat less than the stated value or range, to within a range of ±20% of the stated value; ±19% of the stated value; ±18% of the stated value; ±17% of the stated value; ±16% of the stated value; ±15% of the stated value; ±14% of the stated value; ±13% of the stated value; ±12% of the stated value; ±11% of the stated value; ±10% of the stated value; ±9% of the stated value; ±8% of the stated value; ±7% of the stated value; ±6% of the stated value; ±5% of the stated value; ±4% of the stated value; ±3% of the stated value; ±2% of the stated value; or ±1 % of the stated value.

[0120] Notwithstanding that the numerical ranges and parameters setting forth the broad scope of the disclosure are approximations, the numerical values set forth in the specific examples are reported as precisely as possible. Any numerical value, however, inherently contains certain errors necessarily resulting from the standard deviation found in their respective testing measurements.

[0121] The terms "a,” "an,” "the,” and similar referents used in the context of describing implementations (especially in the context of the following claims) are to be construed to cover both the singular and the plural, unless otherwiseindicated herein or clearly contradicted by context. Recitation of ranges of values herein is merely intended to serve as a shorthand method of referring individually to each separate value falling within the range. Unless otherwise indicated herein, each individual value is incorporated into the specification as if it were individually recited herein. All methods described herein can be performed in any suitable order unless otherwise indicated herein or otherwise clearly contradicted by context. The use of any and all examples, or exemplary language (e.g., "such as”) provided herein is intended merely to better illuminate implementations of the disclosure and does not pose a limitation on the scope of the disclosure. No language in the specification should be construed as indicating any non-claimed element essential to the practice of implementations of the disclosure.

[0122] Groupings of alternative elements or implementations disclosed herein are not to be construed as limitations. Each group member may be referred to and claimed individually or in any combination with other members of the group or other elements found herein. It is anticipated that one or more members of a group may be included in, or deleted from, a group for reasons of convenience and / or patentability. When any such inclusion or deletion occurs, the specification is deemed to contain the group as modified thus fulfilling the written description of all Markush groups used in the appended claims.

[0123] Certain implementations are described herein, including the best mode known to the inventors for carrying out implementations of the disclosure. Of course, variations on these described implementations will become apparent to those of ordinary skill in the art upon reading the foregoing description. The inventors expect skilled artisans to employ such variations as appropriate, and the inventors intend for implementations to be practiced otherwise than specifically described herein. Accordingly, the scope of this disclosure includes all modifications and equivalents of the subject matter recited in the claims appended hereto as permitted by applicable law. Moreover, any combination of the above-described elements in all possible variations thereof is encompassed by implementations of the disclosure unless otherwise indicated herein or otherwise clearly contradicted by context.

Claims

CLAIMSWhat is claimed is:

1. A method, comprising: receiving, on a testing surface, a sample; outputting excitation light to the testing surface; detecting, from the testing surface, scattered light comprising Raman scattering of the sample on the testing surface; generating at least one spectrum representing an intensity of the scattered light with respect to wavelengths of photons in the scattered light; and determining whether the sample comprises at least one analyte by analyzing the at least one spectrum.

2. The method of claim 1 , wherein the sample comprises a sample obtained from a subject.

3. The method of claim 2, wherein the sample comprises blood, plasma, urine, saliva, or mucus.

4. The method of claim 1 , wherein the sample comprises a food sample.

5. The method of claim 1 , wherein the sample comprises residue from a surface.

6. The method of claim 1 , wherein the sample comprises a buffer solution.

7. The method of claim 1, wherein a testing substrate comprising the testing surface comprises: nanostructures disposed on a base substrate.

8. The method of claim 7, wherein the base substrate comprises at least one of silicon, germanium, glass, quartz, a polymer, or a gel.

9. The method of claim 7, wherein an electrically conductive layer comprising at least one of a metal or graphene is disposed between the base substrate and the nanostructures.

10. The method of claim 9, wherein the metal comprises at least one of gold, silver, copper, platinum, or aluminum.11 . The method of claim 7, wherein the nanostructures comprise nanofractals and / or dendrites.

12. The method of claim 7, wherein the nanostructures comprise at least one of gold, silver, copper, platinum, or aluminum.

13. The method of claim 7, wherein the nanostructures comprise plasmonically active nanostructures.

14. The method of claim 7, wherein the nanostructures comprise hydrophobic groups.

15. The method of claim 14, wherein the hydrophobic groups comprise at least one of an alkyl group, an alkane group, an alkene group, an alkyne group, a fluorinated group, a phenyl group, an ester group, or a nonpolar amino acid.

16. The method of claim 14, wherein a portion of the Raman scattering from the hydrophobic groups has a different peak wavelength than a portion of the Raman scattering from the at least one analyte.

17. The method of claim 7, further comprising: modifying a structural conformation of the nanostructures by applying a stimulus to a testing substrate comprising the testing surface.

18. The method of claim 17, wherein the stimulus comprises a chemical stimulus that changes a pH of an environment comprising the sample disposed on the testing surface.

19. The method of claim 17, wherein the stimulus comprises a compressive force, a tensile force, a shear force, or a normal force.

20. The method of claim 7, wherein the testing substrate further comprises a hydrophobic coating.21 . The method of claim 1 , wherein the testing surface is hydrophobic.

22. The method of claim 1, wherein outputting the excitation light to the testing surface comprises illuminating the testing surface with a laser emitting photons having at least one wavelength in a range of about 400 nanometers (nm) to about 800 nm.

23. The method of claim 1 wherein outputting the excitation light and detecting the scattered light is performed by a spectrometer at a point-of-sampling location.

24. The method of claim 1 , wherein the analyte comprises at least one of a nitrosamine, a food additive, a flavor, a nutrient, a pesticide, an herbicide, a drug, an antibiotic, a chemical, a pharmaceutical, a protein, a peptide, a lipid, a nucleic acid, an amino acid, a carbohydrate, a bacteria, a virus, a parasite, an essential oil, a plant extract, a tetrahydrocannabinol, a cannabinoid, a polymer, a plastic, an explosive material, a hydrocarbon, crude oil, or a petroleum product.

25. The method of claim 1 , wherein the analyte comprises one or more bacteria.

26. The method of claim 25, wherein the one or more bacteria comprise at least one Bacillus spp. bacterium, at least one Campylobacter spp. bacterium, at least one Clostridium spp. bacterium, at least one Escherichia spp. bacterium, at least one Klebsiella spp. bacterium, at least one Listeria spp. bacterium, at least one Mycobacterium spp. bacterium, at least one Salmonella spp. bacterium, at least one Shigella spp. bacterium, at least one Staphylococcus spp. bacterium, at least one Pseudomonas spp. bacterium, or at least one Vibrio spp. bacterium.

27. The method of claim 25, wherein the one or more bacteria comprise E. coli, K. pneumoniae, and P. aeruginosa.

28. The method of claim 1 , wherein determining whether the sample contains the analyte by analyzing the at least one spectrum comprises: removing, from the at least one spectrum, a baseline spectrum associated with the testing surface; and in response to removing, from the at least one spectrum, the baseline spectrum associated with the testing surface, determining whether the at least one spectrum comprises at least one characteristic associate with the analyte.

29. The method of claim 1 , wherein determining whether the sample contains the analyte by analyzing the at least one spectrum comprises: inputting the at least one spectrum into a predictive model configured to identify predictive attributes of the spectrum that are associated with the analyte.

30. The method of claim 29, wherein the predictive model comprises at least one machine learning (ML) model.31 . The method of claim 30, wherein the at least one ML model comprises at least one of a transformer, an artificial neural network, a support vector machine (SVM), a nearest neighbor model, a regression analysis model, a clustering models, a gradient boosting model, or a random forest.

32. The method of claim 30, wherein the at least one ML model comprises an SVM and a uniform manifold approximation and projection (UMAP) clustering model.

33. The method of claim 30, further comprising: training the at least one ML model by optimizing parameters of the at least one ML using training data.

34. The method of claim 33, wherein the training data comprises a library of spectra obtained from training samples, a portion of the training samples comprising the at least one analyte, a portion of the training samples omitting the at least one analyte.

35. The method of claim 1 , wherein determining whether the sample contains the at least one analyte by analyzing the at least one spectrum comprises: determining whether the sample contains a first analyte by analyzing the at least one spectrum; and determining whether the sample contains a second analyte by analyzing the at least one spectrum.

36. The method of claim 35, wherein the first analyte comprises a first bacteria, and wherein the second analyte comprises a second bacteria.

37. The method of claim 1 , further comprising: outputting an indication of whether the sample contains the at least one analyte.

38. The method of claim 1 , further comprising: administering, to a subject, a treatment for a condition associated with the at least one analyte, wherein the sample is obtained from the subject.

39. The method of claim 38, wherein the treatment comprises an antibiotic.

40. A testing substrate for surface enhanced Raman spectroscopy (SERS), the testing substrate comprising: a base substrate; an electrically conductive layer disposed on the base substrate; and nanostructures disposed on the electrically conductive layer, wherein a surface of the testing substrate is hydrophobic.41 . The testing substrate of claim 40, wherein the base substrate comprises at least one of silicon, germanium, glass, quartz, a polymer, or a gel.

42. The testing substrate of claim 40, wherein the base substrate is flexible.

43. The testing substrate of claim 42, wherein the base substrate comprises polyimide.

44. The testing substrate of claim 40, wherein the electrically conductive layer comprises at least one of a metal or graphene.

45. The testing substrate of claim 44 wherein the metal comprises at least one of gold, silver, copper, platinum, or aluminum.

46. The testing substrate of claim 40, wherein the nanostructures comprise nanofractals and / or dendrites.

47. The testing substrate of claim 40, wherein the nanostructures comprise at least one of gold, silver, copper, platinum, or aluminum.

48. The testing substrate of claim 40, wherein the nanostructures comprise pl asmonical ly active nanostructures.

49. The testing substrate of claim 40, wherein the nanostructures comprise hydrophobic groups.

50. The testing substrate of claim 49, wherein the hydrophobic groups comprise at least one of an alkyl group, an alkane group, an alkene group, an alkyne group, a fluorinated group, a phenyl group, an ester group, or a nonpolar amino acid.51 . The testing substrate of claim 40, further comprising: a hydrophobic coating disposed on the testing surface.

52. The testing substrate of claim 51 , wherein the hydrophobic coating comprises at least one of a silicone, polydimethylsiloxane (PDMS), polytetrafluoroethylene (PTFE), a fluoropolymer, fluorinated silane, calcium carbonate, titanium oxide nanoparticles, a manganese oxide polystyrene nanocomposite, a zinc oxide polystyrene nanocomposite, a silica nanocoating, carbon nanotubes, or the nanostructures.

53. A cartridge comprising: the testing substrate of claim 40; and a housing at least partially enclosing the testing substrate, at least one well of the housing exposing the testing surface.

54. The cartridge of claim 53, wherein the housing comprises a radio-frequency identification (RFID) tag.

55. The cartridge of claim 53, wherein a QR code is printed on the housing.

56. A kit, comprising: the cartridge of claim 53.

57. The kit of claim 56, further comprising a vessel configured to contain a sample.

58. The kit of claim 57, further comprising: a swab configured to collect residue from a surface and to deposit the residue in the vessel, the sample comprising the residue.

59. The kit of claim 58, further comprising: a cap configured to be removably coupled with the vessel and to be coupled to the swab.

60. A system comprising: the cartridge of claim 53; and a spectrometer comprising: at least one light source configured to emit excitation light to the surface; at least one detector configured to detect scattered light from a sample disposed on the surface; and at least one processor configured to generate at least one spectrum based on the scattered light61 . The system of claim 60, wherein the at least one processor is further configured to determine whether an analyte is present in the sample by analyzing the at least one spectrum.

62. The system of claim 60, wherein the spectrometer is a portable device.

Images