Nanoparticle compositions and methods for biological measurements

Abstract

This disclosure provides nanoparticle compositions, and use thereof for isolating and measuring proteins, protein degrader action, and cells. The disclosed nanoparticles can be identified by barcodes that reveal the identity of and post translational modifications to the bound protein using image processing techniques described herein.

Description

[0001]Docket No.: 381200.00102 NANOPARTICLE COMPOSITIONS AND METHODS FOR BIOLOGICAL MEASUREMENTS CROSS-REFERENCE TO RELATED APPLICATIONS This application claims priority under 35 U.S.C. §119(e) to U.S. Provisional Patent Application No.63 / 667,456, filed July 3, 2024. The foregoing application is incorporated by reference herein in its entirety. REFERENCE TO AN ELECTRONIC SEQUENCE LISTING The contents of the electronic sequence listing (SeqListing_381200.00102.xml; Size 3,471 bytes; and Date of Creation: July 2, 2025) is herein incorporated by reference in its entirety. FIELD OF THE INVENTION The invention relates to nanoparticles capable of isolating and measuring protein or gene expression and the use of such nanoparticles. In particular, the invention relates to identifying barcodes present on nanoparticle constructs and performing analytic or sequencing techniques to measure and identify proteins or genes. BACKGROUND OF THE INVENTION Proteomics, the large-scale study of the proteome, has now emerged as an important field for delineating the structure, function and composition of proteins. Methods to interrogate proteins from biological samples have been developed that are capable of detecting proteins at single molecule levels. Current state of the art methods utilize either micro-arrays, beads or highly scaled nano-arrays. Some methods also utilize the formation of a protein corona around nanoparticles that are mixed with the biological sample to absorb proteins around their surface. The proteins around the surface are subsequently analyzed using mass spectrometry. These methods, however, cannot be used to measure intracellular protein concentration due to contamination of the surface of the particles with proteins outside the cells or the need to lyse cells to detect their intracellular protein composition and concentrations, which effectively eliminates spatial information of the proteins. This makes it difficult to detect low abundance proteins. Measurements of intracellular protein concentrations are important components in detecting cellular homeostatic and dynamic concentration of proteins and their Docket No.: 381200.00102 turnover rate in-line during in vitro experiments. Furthermore, these measurements aid in determining the functional state of cells for diagnostics as well as during drug screening campaigns for therapies that function by actively modulating cellular components or functions (e.g., protein degrader therapies, gene-therapies and cell-based therapies). Additionally, current methods use either lipid nanoparticles or iron oxide nanoparticles (IONPs) and have limited visual labelling capabilities, making downstream processing using either mass spectroscopy or sequencing necessary to decipher the type of protein and their abundance. Therefore, there is an urgent need for novel nanoparticle constructs and methods of identifying these constructs to measure intracellular proteins. SUMMARY OF THE INVENTION In one aspect, this disclosure provides a nanoparticle for measuring protein concentration. In some embodiments, the nanoparticle comprises: a membrane separating an interior space of the nanoparticle from an outer space, wherein the membrane comprises ionizable lipids, solid lipids, nanostructured lipid and / or polymers; and a probe encapsulated within the interior space wherein the probe is selected from the group consisting of magnetic nanoparticles, lipids, nanostructured lipid carriers (NLC), solid lipid carriers (SLN), polymers, aptamers, proteins or combinations thereof. In some embodiments, the membrane comprises cationic lipids, anionic lipids, ionizable lipids and / or polymers: wherein the cationic lipid is selected from the group consisting of 1,2-dioleoyl-3-trimethylammonium-propane (DOTAP), N-(1-(2,3- dioleyloxy)propyl)-N,N,N-trimethylammonium chloride (DOTMA), didodecyldimethylammonium bromide (DDAB), DC-cholesterol, 16:0 1,2-dipalmitoyl-3- dimethylammonium-propane (16:0 DAP) or a combination thereof; wherein the anionic lipids are selected from the group consisting of glycerophospholipids, lysophospholipids, phosphatidic acid, sterols, or combinations thereof; wherein the ionizable lipids are selected from the group consisting of SM-102, ALC-0315, C12-200, cKK-E12, and DLiN-MC3-DMA, oleic acid, or combinations thereof; wherein the polymers are selected from the group consisting of polylactide (PLA), polylactide-co-glycolide (PLGA), poly(ε -caprolactone) (PCL), or combinations thereof; and / or wherein the composition of the lipid, polymer, or lipid- polymer hybrid is tailored to allow selective internalization or tethering onto adherent, semi- adherent, and / or suspension cells by active and / or passive means. Docket No.: 381200.00102 In some embodiments, the membrane has a diameter ranging from about 50 nm to about 1000 nm. In some embodiments, the probe has a diameter ranging from about 5 nm to about 800 nm. In some embodiments, the membrane has at least one chemical modification comprising pegylation, methylation, thiolation, or a combination thereof. In some embodiments, the morphology of the membrane comprises spherical, pyramidal, cubical, pentagonal, or hexagonal geometries In some embodiments, the nanoparticle is localized to specific regions within the cell comprising cellular membranes, nuclei, cytoplasm, endomembrane system, and mitochondria. In some embodiments, the probe comprises a barcode selected from the group consisting of morphological features, dyes, colorimetric agents, and nucleic acid molecules comprising a polynucleotide encoding an aptamer, or an antibody. In some embodiments, the probe is dissolvable by an enzyme selected from lipases, amylases, hydrolases, laccases, and ureases. In some embodiments, the probe is dissolvable by an organic solvent selected from toluene, xylene, and cyclohexane. In some embodiments, the probe is further encapsulated with an organic and / or inorganic biocompatible coating; wherein the organic coating comprises lipids, zwitterionic polymers, and surfactants; wherein the inorganic coating comprises silica; and / or wherein the organic and / or inorganic biocompatible coating is either covalently or electrostatically attached to the probe. In some embodiments, the probe comprises a cysteine rich motif on its binding moiety rendering the moiety pH sensitive. In some embodiments, the dye is a Förster resonance energy transfer (FRET) dye comprising an organic fluorophore selected from cyanine, xanthene, naphthalene, and coumarin dyes. In some embodiments, the zwitterionic polymer comprises polydopamine, betaine, phosphorylcholine or a combination thereof. Docket No.: 381200.00102 In some embodiments, the surfactants comprise a fatty acid ester of glycerol or ethoxylated triglyceride. In some embodiments, the nanoparticle comprises two or more probes that are encapsulated within the membrane, and wherein the two or more probes are capable of binding to a protein of interest. In some embodiments, a first probe comprises a FRET acceptor and a second probe comprises a FRET donor wherein the second probe performs an energy transfer to the first probe when the first and second probes bind to the same protein of interest at different binding sites to enable detection by fluorescent imaging. In some embodiments, the nanoparticle is encapsulated within a microemulsion using a microfluidic device, such as a T-junction device. In some embodiments, the microemulsion containing nanoparticles is introduced to a microemulsion containing cells to perform high throughput screening. In some embodiments, the probe comprises a nucleic acid molecule comprising a polynucleotide encoding an aptamer to detect a post-translational modification. In another aspect, this disclosure also provides a method of manufacturing the nanoparticle as described herein. In some embodiments, the method comprises impingement jet mixing, homogenization, sonication or microfluidic mixing. In some embodiments, the manufacturing process involves heating a lipid phase that comprises a solid lipid and a liquid lipid to a temperature above the melting point of the solid lipid, and heating the aqueous phase comprising a surfactant to 90°C to 95°C; adding the probe to either the lipid phase or aqueous phase and mixing the aqueous solution to the lipid phase; emulsifying the lipid phase and the aqueous phase to form an emulsion using a high speed homogenizer with a mixing speed ranging from 10,000 rpm to 20,000 rpm for a period of time; ultrasonicating the emulsion using a probe sonicator using about 80% amplitude for about 10 minutes; and cooling the emulsion to about room temperature. In another aspect, this disclosure also provides a method of uniquely labelling a probe with an oligo barcode and associating the probe with a physical characteristic such as morphologies and color comprising: imaging the probes located in a compartment such that the probe has a distinguishable physical characteristic; ligating the probe with the oligo barcode; associating the probe with the oligo barcode and cataloging the barcoded probe; pooling the probe of each compartment together and splitting the probes into separate Docket No.: 381200.00102 compartments; imaging a new set of compartmentalized probes and associating the new set of compartmentalized probes with a oligo barcode; ligating a new set of oligo barcodes to the probes; and repeating the process until all the probes have a unique barcode. In another aspect, this disclosure also provides a method of imaging and cataloging the nanoparticle comprising: collecting a z-stack of 3-dimensional images of cells transfected with the nanoparticle, wherein the z-stack is deconvoluted; segmenting individual nanoparticles in each channel using a neural network model based on the StarDist framework; converting the segmented nanoparticles into a 3-dimensional mesh comprising spatial, spectral, and morphological features; passing extracted features through a classification pipeline invariant to rotation and minor scale shifts, wherein the classification pipeline associates each nanoparticle with a cataloged nanoparticle based on barcode feature similarity; assigning unique identifiers to the 3-dimensional mesh and corresponding barcode and storing the unique identifiers in a database; converting the 3-dimensional mesh into an embedding vector and storing the embedding vector in the database; comparing the embedding vector of the nanoparticle to the cataloged nanoparticle by computing a cosine distance; and matching the 3-dimensional mesh of the nanoparticle to the 3-dimensional mesh of the cataloged nanoparticle to determine morphological similarity. In another aspect, this disclosure provides a method of designing an aptamer sequence from a protein structure, comprising: identifying a homologous sequence to the protein structure using a sequence similarity search; converting the protein sequence into an embedding vector; comparing the embedding vector of the protein structure to an embedding vector of a protein with known RNA-binding affinity using a cosine distance; querying a database storing protein-RNA interactions to identify a validated RNA sequence that binds to the protein with known RNA-binding affinity; converting the validated RNA sequence into an embedding vector and comparing the validated RNA sequence embedding vector to an embedding vector of RNA with known protein binding interactions using a cosine distance; iteratively mutating the embedding vector of the validated RNA sequence using an evolutionary algorithm; and evaluating a protein sequence from a protein-RNA pair and an RNA sequence from a protein-RNA pair to determine the likelihood of a binding interaction. In another aspect, this disclosure provides a method of designing an aptamer sequence from a starting RNA sequence, comprising: identifying an RNA sequence with significant sequence similarity to the starting RNA sequence by performing a nucleotide sequence search; Docket No.: 381200.00102 converting the RNA sequence with significant sequence similarity into an embedding vector; comparing the embedding vector of the RNA sequence with significant sequence similarity to an embedding vector of an RNA sequence with known RNA-binding activity using a cosine distance; querying a database storing protein-RNA interactions to identify a validated protein sequence that binds to an RNA sequence with known RNA-binding activity; converting the validated protein sequence into an embedding vector and comparing the embedding vector of the validated protein sequence to an embedding vector of a protein sequence with known RNA- binding interactions using a cosine distance; evaluating a protein sequence from a protein-RNA pair and an RNA sequence from a protein-RNA pair to determine the likelihood of a binding interaction; and querying a database to identify a subcellular location, a gene, and a disease associated with the protein sequence from the protein-RNA pair. In another aspect, this disclosure provides a method of evaluating the protein sequence from the protein-RNA pair to determine the likelihood of a binding interaction, comprising: training a feed forward neural network classifier on a positive and a negative example of an RNA-protein interaction; scaling and normalizing an RNA embedding vector, a protein embedding vector, and a concatenated RNA-protein embedding vector; inputting the scaled and normalized RNA embedding vector, the scaled and normalized protein embedding vector, and the scaled and normalized concatenated RNA-protein embedding vector into the feed forward neural network classifier; and assigning a prediction score and / or a label to the protein sequence from the protein-RNA pair and the RNA sequence from the protein-RNA pair indicating their interaction potential using the feed forward neural network classifier. In another aspect, this disclosure provides a method of measuring a concentration of a protein in a cell culture after the action of a protein degrader, comprising: delivering the protein degrader to a cell; lysing the cell and introducing the nanoparticle containing the first probe into a cell lysate solution; isolating the first probe bound to the protein of interest (POI) using a magnetic field and discarding the cell lysate solution; introducing a solution containing the nanoparticle containing the second probe to the first probe and allowing the second probe to bind to the protein of interest on the first probe; and determining the quantity and identity of the second probe by the barcode associated therewith by an image analysis workflow. In some embodiments, the concentration of the protein of interest is measured using fluorescent imaging. Docket No.: 381200.00102 In some embodiments, the identity and concentration of the protein is determined by dissolving the second probe having the aptamer attached to the POI on the first probe and eluting the bound aptamers; followed by sequencing of the bound aptamers. In another aspect, this disclosure provides a method of measuring a concentration of an intracellular protein, comprising: introducing the nanoparticle containing the first probe into the cell to allow the first probe to bind to the intracellular protein; immobilizing the first probe bound to the intracellular protein using a magnetic field; introducing the nanoparticle containing the second probe into the cell to allow the second probe to bind to the intracellular protein on the first probe; determining the quantity and identity of the second probe by the barcode associated therewith by an image analysis. In some embodiments, the concentration of the intracellular protein is measured using the fluorescent imaging to determine abundance or absence of the intracellular protein. In some embodiments, the identity and abundance or absence of the intracellular protein is determined by dissolving the second probe having the aptamer and eluting the aptamer; and sequencing the eluted aptamer. In another aspect, this disclosure provides a method of isolating tethered cells comprising: adding a first probe comprising a magnetic nanoparticle with a binding moiety that is identifiable by the morphology, size, shape and / or color of the nanoparticle; the first probe tethered to a T cell receptor-complex on the surface of a T cell; isolating the T cell using a magnetic field; identifying a barcode on the probe; and dissolving the probe to selectively retrieve the T cell The foregoing summary is not intended to define every aspect of the disclosure, and additional aspects are described in other sections, such as the following detailed description. The entire document is intended to be related as a unified disclosure, and it should be understood that all combinations of features described herein are contemplated, even if the combinations of features are not found together in the same sentence, or paragraph, or section of this document. Other features and advantages of the invention will become apparent from the following detailed description. It should be understood, however, that the detailed description and the specific examples, while indicating specific embodiments of the disclosure, are given by way of illustration only, because various changes and modifications within the spirit and scope of the disclosure will become apparent to those skilled in the art from this detailed description. Docket No.: 381200.00102 BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 shows the structure of the delivery membrane encapsulating the probes Figure 2A shows the structure of SM-102 lipid nanoparticle membrane (~70nm) encapsulating iron oxide nanoparticle (IONPs) probes. Figure 2B shows the structure of nanostructured lipid carrier membranes with iron oxide nanoparticles probes encapsulated (~100-250nm). Figures 3A and 3B show the structure of the probes comprising Förster resonance energy transfer (FRET) acceptor and FRET donor dyes with pH responsive barcoded oligo labels conjugated to the probes surface. Figures 4A and 4B illustrates the nanoparticles being internalized by their target cell, promoting endosomal escape by lipolysis, and entering the cytoplasm. Figure 5 shows the nanoparticle comprising a nanostructured lipid carrier encapsulating a boron-dipyrromethene (BODIPY) dye can be internalized by induced pluripotent stem cells (iPSCs) and imaged after 2 hours, Figure 5A, and 20 hours, Figure 5B. Figures 6A and 6C show the nanoparticles comprising a nanostructured lipid carrier can encapsulate various amounts of a BODIPY dye and can be detected when delivered to iPSCs. Figure 6B shows the nanoparticles comprising a nanostructured lipid carrier can encapsulate rhodamine dye and be detected when delivered to iPSCs. Figure 7A shows two different nanoparticles comprising nanostructured lipid carriers and containing different fluorescent dyes accumulating in iPSCs. Figures 7B and 7C show nanoparticles comprising nanostructured lipid carriers encapsulating both rhodamine and BODIPY dyes are identifiable when delivered to Michigan Cancer Foundation-7 (MCF-7) cells. Figure 8A shows a fluorescent image of nanoparticles comprising nanostructured lipid carriers, dyes, proteins, and encapsulated iron oxide nanoparticles dried on a glass substrate. Figure 8B shows a fluorescent image of nanoparticles comprising nanostructured lipid carriers, dyes, proteins, and encapsulated iron oxide nanoparticles after incubation with iPSCs for 2 hours. Figure 9A shows how aptamers conjugated to the surface of a primary probe are activated following changes in intracellular pH followed by the capture of proteins of interest. Docket No.: 381200.00102 Figure 9B shows how secondary probes are introduced and can bind to proteins bound to the primary probe to emit a FRET signal that is detectable by microscopy: either in-situ or after cell lysis. Figure 10A shows how a probe can be degraded and a protein magnetically isolated in the presence of an applied magnetic field. Figure 10B shows how an oligo barcode can be released from the protein by denaturing the POI followed by sequencing to identify and quantify POIs. Figure 11 shows the method of using particle tracking to identify proteins at various time points whereby a nanoparticle comprising an aptamer is added to a cell and recognizes a protein. Next, a second probe containing an aptamer is added that binds to the same protein as the first probe. The second binding event results in a FRET signal that is used together with other barcodes to identify the bound protein. Figure 12 shows a scheme for assigning unique oligo barcodes to probes and associating them with other image-based barcode features. Figure 12A shows the imaged split pool method to make unique barcoded probes, probes that are unique in morphology, size and spectral labelling are imaged in separate sets and image processed to render a unique ID based on their characteristics. Figure 12B shows how each set is ligated with a first unique oligo barcode after imaging. Figure 12C depicts the imaged uniquely ligated probes are pooled together and Figure 12D shows the pool of probes is split again and ligated with a second set of oligos and consequently imaged to associate them with their first set of ligated oligos. Figure 13 shows the continuation of the imaged split and pool method where the probes are further imaged and ligated until all probes have a unique oligo barcode ligated to them. Figure 14 depicts how proteins within fixed or frozen tissues are quantified while preserving their spatial identity. Figure 14A depicts a fixed or frozen tissue slice, preferably 5-10 microns in thickness that is immobilized on a glass slide and imaged; Figure 14B shows the first probe added to the tissue and allowed to incubate, where the probe diffuses into the tissue and binds to the proteins of interest; this is followed by washing the tissue to remove un- bound probes. Figure 15 depicts the use of a uniform magnetic field to immobilize the first probes at their place of binding within the tissue. Figure 15A depicts the probes after washing the tissue; the probes are bound to the tissue and immobilized at their place of binding; a uniform magnetic Docket No.: 381200.00102 field is then applied to prevent washout when a tissue lysis buffer is added to separate the first probes and the bound proteins on their surface from the tissue. Figure 15B depicts the first probes with their bound proteins from the tissue at the same relative spatial position in relation to the tissue due to the application of a uniform magnetic field. Figure 16 depicts the process for the detection of bound tissue proteins to the first probes. Figure 16A depicts the second probes containing the unique barcode assigned using imaged split-pool method added to the first probes. Figure 16B depicts the second probes bound to the first probes after incubation and the subsequent imaging and processing of the second probes to associate them with their oligo barcode; the second probes are then degraded and the bound oligo on the first probes are released by denaturing the protein or degrading the first probe; sequencing reveals the identity of the attached proteins and their spatial location. Figure 17 shows a scheme for high throughput screening using a T-junction microfluidic device to form a droplet emulsion comprising the nanoparticle and one or more cells. Figure 18 shows the method of using artificial intelligence and / or machine learning (AI / ML) to predict aptamer sequences that bind to proteins by predicting the interactions of primary and secondary probes with a protein of interest. Figures 19A, 19B, and 19C show how ML can be used to identify and classify nanoparticles based on spatial, spectral, and morphological barcodes using confocal 3- dimensional (3D) imaging. Figure 19A shows a confocal image, Figure 19 shows ML-based instance segmentation of distinct nanoparticles based on morphology, and Figure 19C shows a 3D rendered mesh and identification generation of various nanoparticles detected and labeled using instance segmentation. DETAILED DESCRIPTION OF THE INVENTION The present disclosure provides nanoparticle compositions comprising an outer membrane that encapsulates an inner probe that can comprise one or more distinct barcodes configured to detect and quantify intracellular protein concentrations. The disclosed nanoparticle is capable of transfecting cells and facilitating the localization of a first probe to a target protein, followed by the localization of a second or third probe to the same protein. The identity of the protein may then be determined based on a corresponding nanoparticle catalog. The disclosed nanoparticle compositions are modular in nature and are designed to facilitate Docket No.: 381200.00102 specific binding to target proteins. Methods of using the nanoparticle compositions as disclosed are also provided. Such methods enhance assay throughput, reduce sample volume requirements, and accelerate the screening of libraries of available aptamers. Accordingly, the disclosed nanoparticles and methods of use represent improved strategies for the identification and quantification of proteins using aptamer-based detection systems. Nanoparticle Compositions In one aspect, this disclosure provides a nanoparticle for measuring protein concentration. In some embodiments, the nanoparticle comprises: a membrane separating an interior space of the nanoparticle from an outer space, wherein the membrane comprises ionizable lipids, solid lipids, nanostructured lipids and / or polymers; and a probe encapsulated within the interior space wherein the probe is selected from the group consisting of magnetic nanoparticles, lipids, nanostructured lipid carriers (NLC), solid lipid carriers (SLN), polymers, aptamers, proteins or combinations thereof. A. Membrane In some embodiments, the membrane comprises a cationic lipid, an anionic lipid, an ionizable lipid, and / or a polymer. As used herein, the term “cationic lipid” refers to a lipid molecule that bears a net positive charge at physiological pH. Cationic lipids typically contain one or more amine or quaternary ammonium groups that confer a positive charge under physiological or near- physiological conditions. Cationic lipids are capable of forming complexes with negatively charged molecules such as nucleic acids, and are commonly used in lipid-based delivery systems, including liposomes and lipid nanoparticles, to facilitate cellular uptake and intracellular delivery of therapeutic agents. Exemplary cationic lipids include, but are not limited to, DOTMA (N-[1-(2,3-dioleyloxy)propyl]-N,N,N-trimethylammonium chloride), DOTAP (1,2-dioleoyl-3-trimethylammonium-propane), and DDA (dimethyldioctadecylammonium). As used herein, the term “anionic lipids” refers to lipid molecules that possess a net negative charge under physiological or near-physiological conditions (e.g., pH 6.5–8.0). Anionic lipids typically comprise a hydrophilic head group bearing a negatively charged functional group, such as a phosphate, carboxylate, or sulfate moiety. Exemplary anionic lipids include, but are not limited to, phosphatidylserine (PS), phosphatidic acid (PA), Docket No.: 381200.00102 phosphatidylglycerol (PG), cardiolipin (CL), and synthetic derivatives thereof. Anionic lipids may be naturally derived or synthetically produced, and may be used alone or in combination with other lipid components to form lipid bilayers, liposomes, micelles, or other lipid-based delivery systems. In some embodiments, the cationic lipid comprises 1,2-dioleoyl-3- trimethylammonium-propane (DOTAP), N-(1-(2,3-dioleyloxy)propyl)-N,N,N- trimethylammonium chloride (DOTMA), didodecyldimethylammonium bromide (DDAB), DC-cholesterol, 16:0 1,2-dipalmitoyl-3-dimethylammonium-propane (16:0 DAP), 1,2- Dioleoyl-3-dimethylammonium-propane (DODAP), 1,2-dioleyloxy-3-dimethylaminopropane (DODMA), 1,2-dimyristoyl-3-trimethylammonium-propane (DMTAP), stearylamine, sphingosine, cholesteryl-spermine or a combination thereof. In some embodiments, the anionic lipids comprise glycerophospholipids, lysophospholipids, phosphatidic acid, sterols, or combinations thereof. In some embodiments, the glycerophospholipids comprise phosphatidylcholine, phosphatidylethanolamine, phosphatidylserine, phosphatidylinositol, phosphatidylglycerol, cardiolipin, or a combination thereof. In some embodiments, the lysophospholipids comprise lysophosphatidylserine, lysophosphatidylglycerol, lysophosphatidylinositol, lysophosphatidic acid, lysocardiolipin, or a combination thereof. In some embodiments, the sterols comprise cholesterol, desmosterol, lathosterol, pregnenolone, lanosterol, zymosterol, cholestanol, or a combination thereof. As used herein, the term “ionizable lipids” refers to lipid molecules that contain one or more ionizable functional groups capable of acquiring or losing a proton depending on the pH of the surrounding environment. Ionizable lipids are typically neutral at physiological pH but become positively or negatively charged under acidic or basic conditions, respectively. Such lipids are often used in lipid nanoparticles (LNPs) or other delivery systems to facilitate the encapsulation, stability, and cellular delivery of nucleic acids or other therapeutic agents by enabling pH-dependent interactions with biological membranes and promoting endosomal escape. The term encompasses both naturally occurring and synthetically modified lipids designed to exhibit pH-sensitive ionization behavior. Docket No.: 381200.00102 In some embodiments, the ionizable lipids are selected from the group consisting of SM-102, ALC-0315, C12-200, cKK-E12, and DLiN-MC3-DMA, oleic acid, or a combination thereof. As used herein, the term “polymer” includes linear and branched polymer structures and also encompasses crosslinked polymers as well as copolymers (which may or may not be crosslinked), thus including block copolymers, alternating copolymers, random copolymers, and the like. The polymer may be naturally occurring or obtained from synthetic sources. In certain embodiments, the polymer may be added to the composite or composition thereof to implement supplemental, desired properties to the composite In some embodiments, the polymers are selected from the group consisting of poly(lactic acid) (PLA), poly(glycolic acid) (PGA), poly(lactic-co-glycolic acid) (PLGA), poly(ε-caprolactone) (PCL), poly(L-lactide-co- ε-caprolactone) (PLCL), poly(ethylene glycol)-b-PLGA (PEG-PLGA), polyethylenimine (PEI), chitosan, gelatin, hyaluronic acid, dextran, or a combination thereof. As used herein, the term “lipid-polymer hybrid” refers to a nanoscale construct that comprises both lipid and polymer components integrated into a single architecture. The lipid portion typically consists of one or more naturally occurring or synthetic lipids, which may contribute to biocompatibility, membrane fusion, or endosomal escape, while the polymer component may provide structural stability, controlled release properties, or surface functionalization. Lipid-polymer hybrids combine the advantageous features of lipids and polymers, thereby enhancing delivery efficiency, stability, and bioavailability of encapsulated agents. The term encompasses a wide variety of architectures, including core-shell nanoparticles, vesicular structures, and interpenetrating networks, wherein lipid and polymer domains are either covalently linked or physically associated to form a stable composite. In some embodiments, the composition of the lipid, polymer, or lipid-polymer hybrid is engineered to allow selective cellular internalization or cellular tethering onto adherent, semi-adherent, or suspension cells. As used herein, the term “cellular internalization” refers to the process by which cells uptake external substances, particles, or assemblies through either passive or active transport mechanisms. Passive internalization occurs without the expenditure of cellular energy and typically involves the diffusion or spontaneous partitioning of molecules across a lipid bilayer, driven by concentration gradients or membrane affinity. Active internalization requires energy Docket No.: 381200.00102 dependent processes and involves endolytic pathways such as clathrin-mediated endocytosis, caveolin-mediated endocytosis, macropinocytosis, and phagocytosis. These mechanisms facilitate the selective and regulated uptake of materials into intracellular compartments through membrane invagination and vesicle formation. The term encompasses all modes of internalization by which materials traverse the cell membrane and enter the intracellular environment. As used herein, the term “cellular tethering” refers to the process by which particles, molecules, or biological entities adhere to the outer surface of the cell membrane without undergoing internalization into the intracellular environment. Cellular tethering is mediated through noncovalent interactions such as electrostatic interactions, hydrogen bonding, hydrophobic interactions, or ligand-receptor binding that facilitate proximity to the plasma membrane while maintaining the extracellular localization of the tethered material. The term encompasses both transient and sustained membrane-associated interactions that do not result in endocytosis or membrane penetration. As used herein, the term “adherent cell” refers to mammalian cells that require attachment to a solid or semi-solid substrate for survival, proliferation, or proper physiological function in vitro. These cells grow as monolayers or colonies on culture treated plastic, glass, or extracellular matrix coated substrates. Adherent cells are commonly maintained in static culture vessels such as flasks, plates, or dishes and require enzymatic or mechanical dissociation for passaging and harvesting. As used herein, the term “semi-adherent cells” refers to mammalian cells that exhibit partial or variable attachment to culture surfaces under standard in vitro conditions. These cells may loosely adhere to a substrate while a portion of the population remains in suspension, or they may transition between adherent and non-adherent states depending on factors such as cell density, passage number, or media composition. The term encompasses cell types that do not form stable monolayers yet are not fully suspension adapted. As used herein, the term “suspension cell” refers to mammalian cells that are capable of proliferating and surviving in liquid culture without the need for attachment to a solid substrate. Suspensions cells grow freely dispersed or are loosely associated aggregates within the culture medium and are typically maintained in agitated flasks, spinner vessels, or bioreactors under controlled conditions. The term encompasses both naturally non-adherent cell types and adherent cells that have been adapted or engineered for suspension growth. Docket No.: 381200.00102 In some embodiments, the cell internalizes the membrane by active transport including phagocytosis, macropinocytosis, pinocytosis, clathrin-mediated endocytosis, caveolin- mediated endocytosis, flotillin-mediated internalization, RhoA-mediated internalization, or receptor-mediated transcytosis. In some embodiments, the cell internalizes the membrane by passive transport including simple diffusion or facilitated diffusion. In some embodiments the suspension cells are monocytes, T cells, B cells, natural killer cells, plasma cells, dendritic cells, hematopoietic stem cells, myeloid progenitor cells, eosinophils, basophils, or neutrophils. In some embodiments, the semi-adherent cells are microglia, monocyte-derived dendritic cells, immature dendritic cells, oligodendrocyte precursor cells, activated T cells, activated B cells, perivascular macrophages, or alveolar macrophages. In some embodiments, the adherent cells are induced pluripotent stem cells, fibroblasts, epithelial cells, endothelial cells, mesenchymal stem cells, neurons, astrocytes, oligodendrocytes, Schwann cells, myoblasts, chondrocytes, osteoblasts, keratinocytes, cardiomyocytes, hepatocytes, pancreatic beta cells, adipocytes, melanocytes, bone marrow stromal cells, corneal epithelial cells, lung alveolar epithelial cells, carcinomas, sarcomas, melanomas, glioblastomas, colorectal carcinomas, renal cell carcinomas, thyroid carcinomas, or invasive ductal carcinomas. In some embodiments, the membrane has a diameter ranging from about 50 nm to about 1000 nm (e.g., 50 nm, 100 nm, 150 nm, 200 nm, 250 nm, 300 nm, 350 nm, 400 nm, 450 nm, 500 nm, 550 nm, 600 nm, 650 nm, 700 nm, 750 nm, 800 nm, 850 nm, 900 nm, 950 nm, 1000 nm). In some embodiments, the membrane has at least one chemical modification comprising pegylation, methylation, phosphorylation, amination, biotinylation, glycosylation, acetylation, fluorination, oxidation, halogenation, deuteration, sulfonation, carbamylation, amidation, esterification, nitration, azidation, alkylation, silylation, cyclopropanation, ubiquitination, acylation, and thiolation. Such chemical modifications may be introduced to alter the physicochemical properties, stability, biocompatibility, or targeting capability of the membrane. For example, pegylation may be employed to reduce immunogenicity and prolong circulation time in vivo, while glycosylation or biotinylation may facilitate receptor-mediated Docket No.: 381200.00102 targeting or affinity purification. The modifications may be applied individually or in combination and may occur at one or more locations on the membrane surface, depending on the desired functional attributes of the composition. In some embodiments, the morphology of the membrane comprises spherical, pyramidal, cubical, pentagonal, or hexagonal geometries. In some embodiments, the morphology facilitates internalization or tethering to a cell. In some embodiments, the nanoparticle is localized to specific regions within the cell comprising cellular membranes, nuclei, cytoplasm, endomembrane system, and mitochondria. In some embodiments, the nanoparticle is localized to specific sub regions and / or biological entities of cellular membranes comprising the phospholipid bilayer, cholesterol, glycolipids, peripheral membrane proteins, glycoproteins, integral membrane proteins lipid rafts, ion channels, caveolae, clathrin-coated pits, cell adhesion molecules, tight junction proteins, receptors, gap junctions, and desmosomes. In some embodiments the nanoparticle is localized to specific sub regions and / or biological entities of the nucleus comprising the nuclear envelope, outer nuclear membrane, inner nuclear membrane, nuclear pores, nuclear pore complexes, nuclear lamina, nucleoplasm, chromatin, DNA, histones, nucleosomes, nucleolus, transcription factors, and spliceosomes. In some embodiments the nanoparticle is localized to specific sub regions and / or biological entities of the cytoplasm comprising proteosomes, autophagosomes, and lipid droplets. In some embodiments the nanoparticle is localized to specific sub regions and / or biological entities of the endomembrane system comprising rough endoplasmic reticulum, smooth endoplasmic reticulum, ribosomes, Golgi apparatus, Golgi vesicles, transport vesicles, secretory vesicles, endosomes, endolysosomes, lysosomes, and peroxisomes. In some embodiments the nanoparticle is localized to specific sub regions and / or biological entities of the mitochondria comprising, outer mitochondrial membrane, inner mitochondria membrane, cristae, intermembrane space, mitochondrial matrix, mitochondrial DNA, ATP synthase complexes, and electron transport chain components. B. Probe Docket No.: 381200.00102 In some embodiments, the probe is selected from the group consisting of magnetic nanoparticles, lipids, nanostructured lipid carriers (NLC), solid lipid carriers (SLC), polymers, aptamers, proteins or a combination thereof. In some embodiments, the magnetic nanoparticles are iron oxide nanoparticles (IONPs) that are synthesized by a method comprising reduction, thermal decomposition, hydrothermal, microemulsion sol-gel, or combustion. The reduction method comprises reducing iron (III) salts in the presence of a reducing agent. Thermal decomposition comprises decomposition of organometallic precursors in high boiling point organic solvents at a temperature of about 200 to about 300°C (e.g., 200°C, 205°C, 210°C, 215°C, 220°C, 225°C, 230°C, 235°C, 240°C, 250°C, 255°C, 260°C, 265°C, 270°C, 275°C, 280°C, 285°C, 290°C, 295°C, 300°C) under an inert gas. Hydrothermal synthesis comprises precipitating iron salts in water using a seal autoclave at about 180 to about 220°C (e.g., 180°C, 185°C, 190°C, 195°C, 200°C, 205°C, 210°C, 215°C, 220°C). Microemulsion method comprises forming magnetic nanoparticles in water-in-oil or oil-in-water reverse micelles using surfactants. Sol-gel methods comprise subjecting iron alkoxide precursors to hydrolysis and condensation. Combustion comprises combusting iron precursors in oxygen rich flames. In some embodiments, the reducing agents comprise sodium borohydride, hydrazine hydrate, sodium hypophosphite, sodium citrate, sodium sulfite, or sodium thiosulfate. In some embodiments, the high boiling point solvents comprise benzyl ether, 1- octadecene, phenyl ether, di-n-octyl ether, 1-phenyloctane, diethylene glycol, triethylene glycol, ethylene glycol, dimethyl sulfoxide, and water. In some embodiments, the iron precursors comprise iron (II) chloride, iron (III) chloride, iron (III) nitrate, iron (II) sulfate, iron (III) sulfate, iron (II) acetate, iron (III) acetate, iron (III) perchlorate, iron (II) acetylacetonate, iron (III) acetylacetonate, iron (III) oleate complex, iron (III) stearate, iron (II) oleylamine complex, iron (II) laurate, iron (II) oxide, iron (III) oxide, iron (III) isopropoxide, iron (III) ethoxide, iron pentacarbonyl, ferric ammonium sulfate, or ferric citrate. In some embodiments, the inert gas comprises nitrogen or argon. In some embodiments, the IONP is doped with one or more transition metals to modulate magnetic or structural properties. The dopant may be selected from the group consisting of zinc, cobalt, manganese, nickel, or combinations thereof. Docket No.: 381200.00102 In some embodiments, the doped IONPs exhibit superparamagnetism. The superparamagnetic iron oxide nanoparticles have a diameter ranging from about 5 to about 25 nm (e.g., 5 nm, 6 nm, 7 nm, 8 nm, 9 nm, 10 nm, 11 nm, 12 nm, 13 nm, 14 nm, 15 nm, 16 nm, 17 nm, 18 nm, 19 nm, 20 nm, 21 nm, 22 nm, 23 nm, 24 nm, 25 nm). In some embodiments, the NLC comprises imperfect crystal NLCs, multiple oil-in-fat NLCs, or amorphous NLCs. The NLC is formed by mixing ionizable lipids, zwitterionic phospholipids, helpers, and stabilizing agents. As used herein, the term “nanostructured lipid carriers” or “NLCs” refers to submicron- sized lipid-based nanoparticles composed of a mixture of solid lipids and liquid lipids (oils), stabilized by surfactants, and designed to encapsulate and deliver active pharmaceutical ingredients. NLCs typically have a particle size in the range of about 50 to 1000 nanometers and form a non-crystalline or partially crystalline lipid matrix that allows for improved drug loading capacity, reduced drug expulsion during storage, and controlled or sustained drug release. The structural matrix of NLCs is characterized by a disordered inner phase resulting from the combination of solid and liquid lipids, which distinguishes them from solid lipid nanoparticles (SLNs) and enhances their ability to accommodate poorly water-soluble or lipophilic compounds. As used herein, the term “solid lipid carriers” refers to biocompatible and biodegradable lipid-based delivery systems comprising one or more lipids that are solid at room temperature and at body temperature. The solid lipid component may include, but is not limited to, natural, semisynthetic, or synthetic triglycerides, fatty acids, fatty alcohols, waxes, or esters thereof. Solid lipid carriers may further comprise surfactants, emulsifiers, or co-lipids to stabilize the carrier and enhance drug loading, encapsulation efficiency, and controlled release of a therapeutic or diagnostic agent. Solid lipid carriers include, for example, solid lipid nanoparticles (SLNs) and nanostructured lipid carriers (NLCs), and may be formulated to improve the solubility, stability, bioavailability, and targeted delivery of active pharmaceutical ingredients. In some embodiments, the NLC is prepared by mixing an aqueous phase with a lipid phase. The lipid phase comprises one or more solid lipids and one or more liquid lipids. The solid lipids may include, but are not limited to, triglycerides, fatty acids, fatty alcohols, waxes, or esters thereof that remain solid at ambient and physiological temperatures. The liquid lipids may include medium-chain triglycerides, oils, or other pharmaceutically acceptable lipophilic Docket No.: 381200.00102 compounds that remain in a liquid state under the same conditions. The aqueous phase may contain water and optionally one or more surfactants, stabilizers, or co-solvents to facilitate emulsification and stabilization of the lipid components. The lipid phase is typically melted and dispersed into the aqueous phase under high shear or homogenization conditions to form a nano- or microemulsion, which solidifies upon cooling to form the solid lipid carrier system. This process may optionally include sonication, extrusion, or other size-reduction techniques to achieve the desired particle size and distribution. In some embodiments, the lipid phase comprises solid lipids and liquid lipids. In some embodiments, the probe comprises a barcode selected from the group consisting of morphological features, dyes, colorimetric agents, and nucleic acid molecules comprising a polynucleotide. As used herein, the term “barcode” refers to an identifier associated with a probe, wherein the barcode comprises a series of nucleotides, oligonucleotides, or other detectable elements that uniquely identify the probe or a set of probes. The barcode may be physically or chemically attached to the probe and is designed to enable probe tracking, decoding, or multiplexing during analytical procedures such as sequencing, hybridization, or imaging. The barcode may be read by optical, biochemical, or electronic means, and may optionally include error-correction elements or indexing sequences. As used herein, the term “morphological feature” refers to any measurement or observable structural characteristics of the probe that defines the probe's external or internal physical form. Morphological features include, but are not limited to, probe size, shape, aspect ratio, surface roughness, porosity, and degree of deformation. These features may be assessed using imaging techniques such as confocal imaging or scattering techniques such as transmission electron microscopy, scanning electron microscopy, atomic force microscopy, or dynamic light scattering. Morphological features influence functional properties of the probe including cellular internalization, cellular tethering, biodistribution, endosomal escape, or degradation release kinetics. The term encompasses both intrinsic design parameters and environment-induced structural changes that affect nanoparticle performance. In some embodiments, the barcode is a morphological feature comprising diameter, circumference, shape, aspect ratio, surface roughness, surface curvature, crystallinity, core- shell structures, shell thickness, density, porosity, aggregation state, zeta potential, surface defects, grain boundaries, surface ligand distribution, electron density contrast, molecular Docket No.: 381200.00102 weight, chain conformation, tacticity, degree of branching, phase separation behavior, lamellar structures, fibrillar structures, micellar structures, self-assembly, surface topology, fracture patterns, brittleness, helical pitch, helical twist, major groove dimensions, minor groove dimensions, strandedness, secondary structures, tertiary structures, length, degree of supercoiling, condensation state, base pair stacking density, melting profile, folding architecture, bilayer thickness, lipid droplet size, curvature, domain segregation, packing parameter morphology, shape anisotropy, fission / fusion behavior, and membrane defects. In some embodiments, the probe has a diameter ranging from about 5 to about 800 nm (e.g., 5 nm, 30 nm, 55 nm, 80 nm, 105 nm, 130 nm, 155 nm, 180 nm, 205 nm, 230 nm, 255 nm, 280 nm, 305 nm, 330 nm, 355 nm, 380 nm, 405 nm, 430 nm, 455 nm, 480 nm, 505 nm, 530 nm, 555 nm, 580 nm, 605 nm, 630 nm, 655 nm, 680 nm, 705 nm, 730 nm, 755 nm, 780 nm, 800 nm ). In some embodiments, the barcode is a dye and / or colorimetric agent comprising, Nile Red, BODIPY 493 / 503, FM4-64, Laurdan, Dil, DiO, DiD, Oil Red O, Sudan III, Sudan IV, LipidTOX, R18, TopFluor Cholesterol, merocyanine 540, DAPI, Hoechst, ethidium bromide, SYBR Green I, SYBR Green II, GelRed, GelGreen, propidium iodide, Acridine Orange, YOYO-1, TOTO-1, EvaGreen, Thiazole Orange, PicoGreen, RiboGreen, Methylene Blue, Neutral Red, rhodamine B isothiocynate, fluorescein isothiocyanate, Texas Red, NHS- fluorescein, Coumarin-6, pyrene, dansyl chloride, tetramethylrhodamine, Bromophenol Blue, Crystal Violet, Congo Red, Coomassie Brillant Blue, Phenol Red, Trypan Blue, Alizarin Red S, and Reactive Blue 2. In some embodiments, the barcode is a nucleic acid molecule comprising a polynucleotides comprising aptamers, RNA, tRNA, rRNA, snRNA, snoRNA, miRNA, siRNA, piRNA, lncRNA, gRNA, circRNA, antisense RNA, DNA, mtDNA, pDNA, satellite DNA, minicircle DNA, peptide nucleic acids, locked nucleic acids, thiolated nucleic acids, biotinylated nucleic acids, fluorescently labeled nucleic acid, xeno nucleic acids, G-quadruplex DNA, and triple-helical DNA. In some embodiments, the probe comprises an aptamer, an oligonucleotide barcode, and a 3’ chemical modification. The aptamer may be between about 30 and about 60 nucleotides in length (e.g., 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, or 60 nucleotides). The oligonucleotide barcode may be between about 6 and about 12 nucleotides in length (e.g., 6, 7, 8, 9, 10, 11, or 12 Docket No.: 381200.00102 nucleotides). In some embodiments, the 3’ chemical modification may comprise one or more of the following: a 3’-phosphate group, 3’-hydroxyl group, 3’-inverted thymidine, 3’- biotinylation, 3’-fluorescent labeling, 3’-amino modification, 3’-poly(A) tailing, 3’- phosphorothioate linkage, 3’-thiolation, 3’-locked nucleic acids (LNAs), or 3’-pegylation. In some embodiments, the 3’-inverted nucleotide region comprises between about 2 and about 20 nucleotides (e.g., 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, or 20 nucleotides). Such 3’ modifications may be introduced to improve stability, biocompatibility, conjugation efficiency, or targeting capability of the probe. For example, the incorporation of locked nucleic acids may enhance nuclease resistance and structural stability, while a 3’-inverted thymidine may inhibit exonuclease degradation and confer conformational rigidity, thereby increasing the intracellular half-life of the probe. These modifications may be employed individually or in combination, depending on the intended functional characteristics of the probe or composition. In some embodiments, the probe is dissolvable by enzymes comprising dehydrogenases, oxidases, reductases, peroxidases, laccases, monooxygenases, dioxygenases, hydroxylases, catalases, nitric oxide synthases, cytochrome p450 enzymes, kinases, aminotransferases, methyltransferases, acetyltransferases, glycosyltransferases, phosphotransferases, sulfotransferases, nucleotidyltransferases, prenyltransferases, carboxyltransferases, lipases, phospholipases, esterases, proteases, amylases, cellulases, lactase, maltase, sucrase, nucleases, urease, phosphatases, glycosidases, chitinase, decarboxylases, aldolases, synthases, deaminases, dehydratases, fumarase, histidine ammonia- lyase, carbonic anhydrase, acetyl-CoA synthase, glutaminase, malate dehydrogenase, enolase, aldolase, glutathione peroxidase, urease, superoxide dismutase, or thioredoxin reductase. In some embodiments, the probe is dissolvable by organic solvents comprising hexane, toluene, xylene, petroleum ether, chloroform, dichloromethane, diethyl ether, benzene, cyclohexane, ethyl acetate, acetone, isopropanol, methanol, ethanol, tetrahydrofuran, dimethylformamide, dimethyl sulfoxide, 1,4-dioxane, N-methyl-2-pyrrolidone, formamide, phenol, phenol-chloroform, isoamyl alcohol, or acetonitrile In some embodiments, the probe is further encapsulated with an organic and / or inorganic biocompatible coating; wherein the organic coating comprises lipids, zwitterionic polymers, and surfactants; wherein the inorganic coating comprises silica; and / or wherein the Docket No.: 381200.00102 organic and / or inorganic biocompatible coating is either covalently or electrostatically attached to the probe. As used herein, the term “zwitterionic polymer” refers to a polymer that comprises repeating units bearing both positively and negatively charged functional groups within the same monomer or side chain, such that the overall polymer is electrically neutral at physiological pH. Zwitterionic polymers may include, but are not limited to, polymers incorporating betaine, phosphorylcholine, sulfobetaine, or carboxybetaine moieties. These polymers exhibit high hydrophilicity, antifouling characteristics, and resistance to non-specific protein adsorption due to the formation of a tightly bound hydration layer. The zwitterionic polymer may be linear, branched, or crosslinked, and may be synthesized as a homopolymer or incorporated as part of a copolymer system. In some embodiments, the organic biodegradable coating is a zwitterionic polymer comprising polydomaine, poly(sulfobetaine methacrylate), poly(carboxybetaine methacrylate), poly(2-methacryloyloxyethyl phosphorylcholine), poly(sulfopropylbetaine), poly(sulfobetaine), poly(vinyl betaine), poly(methacryloylaminosulfobetaine), or poly(sulfobetaine methacrylamide). As used herein, the term “surfactant” refers to a surface-active agent comprising amphiphilic molecules that contain hydrophilic and hydrophobic domains, enabling them to adsorb at interfaces between dissimilar phases. Surfactants function to reduce interfacial tension, stabilize emulsions or dispersion, and facilitate the formation of colloidal structures. The term encompasses a broad range of ionic and nonionic compounds including synthetic, natural, and polymeric variants. In some embodiments the surfactants comprise fatty acid ester of glycerol or ethoxylated triglyceride including glyceryl monostearate, glyceryl monooleate, glyceryl monolaurate, glyceryl monoacetate, glyceryl disterate, glyceryl tristearate, PEG-6 caprylic glycerides, PEG-7 glyceryl cocoate, PEG-8 glyceryl cocoate, PEG-40 hydrogenated castor oil, PEG-100 stearate, PEG-20 glyceryl stearate, or polyoxyethylene gylceryl tallowate. In some embodiments, the surfactants comprise Poloxamer 188, Tween 20, Tween 40, Tween 60, Tween 80, Span 80, Brij 35, polyethylene glycol, TritonX-100, Pluronic F-68, sodium dodecyl sulfate, sodium cholate, sodium deoxycholate, dioctyl sulfosuccinate sodium salt, linear alkylbenzene sulfonate, sodium stearate, sodium lauryl ether sulfate, cetyltrimethylammonium bromide, dodecyltrimethylammonium bromide, Docket No.: 381200.00102 dimethyldioctadecylammonium bromide, dioctadecyldimethylammonium bromide, benzalkonium chloride, poly(vinyl alcohol), lecithin, rhamnolipids, sophrolipids, or saponins. In some embodiments, the inorganic biocompatible coating comprises silica. The silica coating may be solid or mesoporous and may be synthesized from the group of synthetic methods consisting of Stöber method, microemulsion method, and seeded polymerization templated synthesis, flame pyrolysis, and precipitation method. In some embodiments, the silica coating is functionalized by grafting silanes containing various functional groups to the surface of the silica coating comprising, alkoxy silanes, amino silanes, epoxy silanes, vinyl silanes, acrylic silanes, isocyanate silanes, carboxyl silanes, aldehyde silanes, azide silanes, fluoro silanes, benzoyl silanes, PEG silanes, phosphonate silanes, sulfonic acid silanes, hydrosilanes, alkyne silanes, alkyl silanes, ureido silanes, methacrylate silanes, or mercapto silanes. For example, azide functionalized silanes may enable bioorthogonal click chemistry with alkyne containing probes for site specific attachment, while PEG silanes may confer antifouling properties and reduce non-specific binding by forming a sterically stabilized shell on the silane surface. In some embodiments, the organic and / or inorganic biocompatible coating is either electrostatically or covalently attached to the probe. In some embodiments, the degradation kinetics of the probe can be controlled by changing the thickness of the biodegradable coating, wherein thicker coatings may provide prolonged stability while thinner coatings may enable faster degradation. The degradation rate may further be modulated by selecting coating materials with different hydrolytic or enzymatic susceptibilities or by incorporating structural modifications such as crosslinking density, porosity, or surface area to volume ratio. For example, silica coatings may undergo pH dependent dissolution under acidic conditions such as those found in endosomal and lysosomal compartments while lipid-based coatings may degrade in response to enzymatic activity, oxidation, or membrane fusion events. In some embodiments, the aptamer conjugated to the probe is pH sensitive due to the presence of cysteine rich motifs, intercalated motifs, triplex-forming sequences, histidine rich tags, acridine modified nucleotides, imidazole modified nucleotides, protonable aptamer loops, or polyC hairpins. For example, cysteine rich motifs may undergo protonation induced thiol reactivity or disulfide rearrangement under mildly acidic conditions, while triplex forming Docket No.: 381200.00102 sequences may stabilize by cytosine protonation in the major groove, enabling pH-dependent hybridization and conditional activation of the probe in acidic environments. In some embodiments, the probe comprises a Förster resonance energy transfer dye comprising an organic fluorophore selected from cyanine, xanthene, naphthalene, and coumarin dyes. The organic fluorophore comprises In some embodiments, the Förster resonance energy transfer dyes comprise Alexa Fluor 350, Alexa Fluor 405, Alexa Fluor 448, Alexa Fluor 532, Alexa Fluor 546, Alexa Fluor 555, Alexa Fluor 561, Alexa Fluor 568, Alexa Fluor 594, Alexa Fluor 647, Cyanine3 (Cy3), Cyanine5 (Cy5), Cyanine7 (Cy7), 5-((2-Aminoethyl)amino)naphthalene-1-sulfonic acid (EDANS), 5(6)-Carboxyfluorescein (FAM) N-hydroxysuccinimide (NHS) ester, FAM sulfo- NHS ester, FAM vinylsulfone, FAM azide, FAM alkyne, FAM dibenzocyclooctyne (DBCO), FAM thiol, FAM maleimide, FAM hydrazide, FAM polyethylene glycol 4 (PEG4) alkyne, FAM amine, FAM dichlorotriazine, 5(6)-Carboxytetramethylrhodamine (TAMRA) NHS ester, TAMRA sulfo-NHS ester, TAMRA vinylsulfone, TAMRA azide, TAMRA alkyne, TAMRA DBCO, TAMRA thiol, TAMRA maleimide, TAMRA hydrazide, TAMRA polyethylene glycol 4 (PEG4) alkyne, TAMRA amine, TAMRA dichlorotriazine, CF405S, CD488A, CF594, CF633, DyLight 405, DyLight 488, DyLight550, DyLight 594, DyLight 633, Atto 425, Atto 488, Atto 532, Atto 550, Atto 565, Atto 594, or 5-carboxy-x-rhodamine (5- ROX). The Förster resonance energy transfer dye pair is selected to provide optimized spectral overlap and minimal bleed-through for multiplexed detection In some embodiments, two or more probes are encapsulated within the membrane to bind to a protein of interest through one or more binding mechanisms selected from the group consisting of allosteric binding, competitive binding, non-competitive binding, cooperative binding, and non-cooperative binding. For example, a first probe binds to an allosteric site of the protein inducing a conformational change that modulates the binding affinity of a second probe. Conversely, the first probe binds non-cooperatively to the protein and does not alter the affinity or binding kinetics of the second probe. In some embodiments, the probes may comprise chemically or structurally, distinct entities, including small molecules, peptides, aptamers, or antibodies and may engage the protein through electrostatic, hydrophobic, hydrogen binding, van der Waals, or covalent interactions to enable selective, multiplexed, or synergistic modulation of the protein. Docket No.: 381200.00102 In some embodiments, a first probe comprises a Förster resonance energy transfer acceptor and a second probe comprises a Förster resonance energy transfer donor whereby the second probe performs an energy transfer to the first probe when binding to the same protein of interest to enable detection by fluorescent imaging. As used herein, the term “microemulsion” refers to a thermodynamically stable, optically transparent or translucent colloidal dispersion comprising at least one immiscible liquid phase dispersed within another and stabilized by one or more surfactants. In certain embodiments, the microemulsion comprises a continuous phase and dispersed phase selected from oil-in-water, water-in-oil, or bicontinuous phase depending on the composition and concentration of the surfactants. In some embodiments, the microemulsion may encapsulate nanoparticles and may be formulated using surfactants selected from ionic, nonionic, zwitterionic, or polymeric species to provide colloidal stability and tunable interfacial properties. In some embodiments, the nanoparticle is encapsulated within a microemulsion using a microfluidic device. The microfluidic device comprises one or more fluidic geometries including T-junction channels, flow focusing channels, co-flow channels, cross-flow channels, electrowetting-on-dielectric, pneumatic valves, droplet merging junctions, or droplet splitting forks. In certain embodiments, the microfluidic device comprises a T-junction wherein two immiscible fluid streams intersect orthogonally to produce monodisperse droplets encapsulating the nanoparticle formulation. In other embodiments, a flow-focusing microfluidic configuration is used wherein a central stream containing the nanoparticle is sheared by two lateral carrier fluid streams to produce uniformly sized microemulsion droplets. The use of microfluidic systems enables precise control over droplet size and reproducible nanoparticle loading under constant flow conditions. In some embodiments, the microemulsion containing nanoparticles is introduced to a microemulsion containing cells to perform high throughput screening. The microemulsions may be formed and manipulated with a microfluidic pattern comprising one or more channel geometries described herein. In certain embodiments, discrete aqueous droplets encapsulating nanoparticles are fused or co-encapsulated with droplets containing individual suspension or adherent cells. The structure permits controlled and parallelized interrogation of nanoparticle internalization, localization, and measurements at a single cell resolution in a manner which is reproducible. Docket No.: 381200.00102 In some embodiments, the probe comprises a nucleic acid molecule comprising a polynucleotide encoding an aptamer to detect post-translational modifications. The aptamer may be a single-stranded RNA oligonucleotide that folds into a 3-dimensional conformation capable of recognizing a specific chemical modification on a protein. These chemical modifications comprise phosphorylation, methylation, acetylation, ubiquitination, glycosylation, SUMOylation, NEDDylation, palmitoylation, myristylation, prenylation, hydroxylation, nitrosylation, formylation, or citrullination. The aptamer may bind its post- translation modification by electrostatic, hydrogen bonding, hydrophobic interactions, conformational specific binding, structural rearrangements, or charge neutralization sensing and may further be conjugated to a signal-generating moiety or integrated into a Förster resonance energy transfer-based system for detection and quantification of modification specific binding events. Manufacturing Nanoparticle Compositions In another aspect, this disclosure also provides a method of manufacturing the nanoparticle as described herein. In some embodiments, the method comprises impingement jet mixing, homogenization, sonication or microfluidic mixing. In some embodiments, the manufacturing process involves heating a lipid phase that comprises a solid lipid and a liquid lipid to a temperature above the melting point of the solid lipid, and heating the aqueous phase comprising a surfactant to 90°C to 95°C (e.g., 90°C, 91°C, 92°C, 93°C, 94°C, 95°C); adding the probe to either the lipid phase or aqueous phase and mixing the aqueous solution to the lipid phase; emulsifying the lipid phase and the aqueous phase to form an emulsion using a high speed homogenizer with a mixing speed ranging from 10,000 rpm to 20,000 rpm (e.g., 10,000 rpm 10,500 rpm, 11,000 rpm, 11,500 rpm, 12,000 rpm, 12,500 rpm, 13,000 rpm, 13,500 rpm, 14,000 rpm, 14,500 rpm, 15,000 rpm, 15,500 rpm, 16,000 rpm, 16,500 rpm, 17,000 rpm, 17,500 rpm, 18,000 rpm, 18,500 rpm, 19,000 rpm, 19,500 rpm, 20,000 rpm) for a period of time; ultrasonicating the emulsion using a probe sonicator using about 80% amplitude for about 10 minutes; and cooling the emulsion to about room temperature. In some embodiments, the solid lipid comprises glyceryl monostearate and the liquid lipid comprises oleic acid. In some embodiments, the length of the lipid chain varies for the solid and / or liquid lipid. In some embodiments, the liquid lipids further comprise α-linolenic acid, stearidonic acid, eicosapentaenoic acid, cervonic acid, linoleic acid, linolelaidic acid, γ- Docket No.: 381200.00102 linolenic acid, di-homo-γ-linolenic acid, arachidonic acid, docosatetraenoic acid, palmitoleic acid, vaccenic acid, paullinic acid, elaidic acid, gondoic acid, erucic acid, nervonic acid, and mead acid. In some embodiments, the solid lipid further comprises monolaurin and glyceryl hydroxystearate. In some embodiments glyceryl monostearate comprises 50% to 75% (e.g., 50%, 51%, 52%, 53%, 54%, 55%, 56%, 57%, 58%, 59%, 60%, 61%, 62%, 63%, 64%, 65%, 66%, 67%, 68%, 69%, 70%, 71%, 72%, 73%, 74%, 75%) and oleic acid comprises 25% to 50% (e.g., 25%, 26%, 27%, 28%, 29%, 30%, 31%, 32%, 33%, 34%, 35%, 36%, 37%, 38%, 39%, 40%, 41%, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, 50%) of the lipid structure. In some embodiments, the hydrophobic dye is a BODIP dye. In some embodiments, the BODIPY dye comprises 0.1% to 1% (e.g., 0.1%, 0.15%, 0.20%, 0.25%, 0.30%, 0.35%, 0.40%, 0.45%, 0.50%, 0.55%, 0.60%, 0.65%, 0.70%, 0.75%, 0.80%, 0.85%, 0.90%, 0.95%, 1.0%) of the lipid structure. In some embodiments, the lipid phase is heated to 70°C to 75°C (e.g., 70°C, 71°C, 73°C, 74°C, 75°C). In some embodiments, the rhodamine phalloidin solution is added to the heated lipid phase and thoroughly mixed to ensure an even distribution of dye in the solution. In some embodiments, the membrane is formed by mixing ionizable lipids, zwitterionic phospholipids, helpers, and stabilizing agents. In some embodiments the ionizable lipids comprises SM-102, the zwitterionic phospholipids comprises 1,2-Distearoyl-sn-glycero-3- phosphocholin (DSPC), the helper comprises cholesterol, and the stabilizing agent comprises polyethylene glycol (PEG). In some embodiments SM-102 comprises 40% to 50% (e.g., 40%, 41%, 42%, 43%, 44%, 45%, 46%, 47%, 48%, 49%, or 50%), DSPC comprises 7.5% to 10% (e.g., 7.5%, 8.0%, 8.5%, 9.0%, 9.5%, or 10%), cholesterol comprises 35.5% to 38.5% (e.g., 35.5%, 36.0%, 36.5%, 37%, 37.5%, 38%, or 38.5%), and PEG comprises 1.5% to 4.5% (e.g., 1.5%, 2.0%, 2.5%, 3.0%, 3.5%, 4.0%, or 4.5%) of the lipid composition. In some embodiments, the formed nanoparticles have their pH adjusted to 7.4 using sodium hydroxide or hydrochloric acid. In some embodiments, the formed nanoparticles are washed using dialysis for about 24 hours. In some embodiments 1 ml of nanoparticles are placed in a dialysis chamber and washed with about 1 L of deionized water for about 24 hours. In some embodiments, the formed nanoparticles are characterized using analytical techniques to determine particle size, zeta potential and polydisperson index. In some embodiments these analytical techniques include dynamic light scattering and ultraviolet-visible spectroscopy. In some embodiments the washed nanoparticles are stored in the dark at about 4°C to prevent Docket No.: 381200.00102 degradation and / or photobleaching of the encapsulated dye. In some embodiments, the nanoparticles encapsulating IONPs are magnetically separated using an applied magnetic field. In some embodiments, the washed nanoparticles are delivered to a cell. In some embodiments an effective amount of the nanoparticle of any of the embodiments described herein is delivered to a cell for about 2 hours to 24 hours (e.g., 2 hour, 3 hours, 4 hours, 5, hours, 6 hours, 7 hours, 8 hours, 9 hours, 10, hours, 11 hours, 12 hours, 13 hours, 14 hours, 15 hours, 16 hours, 17 hours, 18 hours, 19, hours, 20 hours, 21 hours, 22 hours, 23 hours, 24 hours) and the cells are washed with a buffer solution 3 times. In some embodiments, the nanoparticle solution is diluted by a factor of about 100 to about 1000 (e.g., 100, 150, 200, 250, 300, 350, 400, 450, 500, 550, 600, 650, 700, 750, 800, 850, 900, 950, 1000). In some embodiments the buffer is phosphate buffered saline. In some embodiments, the degradation kinetics of the lipid nanoparticle are modulated by the ratio of liquid lipid to solid lipid. In some embodiments, the degradation kinetics of the lipid nanoparticle are modulated by the identity of the liquid lipid and solid lipid. In some embodiments, this allows various lipid nanoparticles to differentially degrade. In another aspect, this disclosure also provides a method of uniquely labelling a probe with an oligo barcode and associating the probe with a physical characteristic such as morphologies and color comprising: imaging the probes located in a compartment such that the probe has a distinguishable physical characteristic; ligating the probe with the oligo barcode; associating the probe with the oligo barcode and cataloging the barcoded probe; pooling the probe of each compartment together and splitting the probes into separate compartments; imaging a new set of compartmentalized probes and associating the new set of compartmentalized probes with a oligo barcode; ligating a new set of oligo barcodes to the probes; and repeating the process until all the probes have a unique barcode. Measurements using Nanoparticle Compositions In another aspect, this disclosure also provides a method of imaging and cataloging the nanoparticle comprising: collecting a z-stack of 3-dimensional images of cells transfected with the nanoparticle, wherein the z-stack is deconvoluted; segmenting individual nanoparticles in each channel using a neural network model based on the StarDist framework; converting the segmented nanoparticles into a 3-dimensional mesh comprising spatial, spectral, and morphological features; passing extracted features through a classification pipeline invariant to rotation and minor scale shifts, wherein the classification pipeline associates each Docket No.: 381200.00102 nanoparticle with a cataloged nanoparticle based on barcode feature similarity; assigning unique identifiers to the 3-dimensional mesh and corresponding barcode and storing the unique identifiers in a database; converting the 3-dimensional mesh into an embedding vector and storing the embedding vector in the database; comparing the embedding vector of the nanoparticle to the cataloged nanoparticle by computing a cosine distance; and matching the 3-dimensional mesh of the nanoparticle to the 3-dimensional mesh of the cataloged nanoparticle to determine morphological similarity. As used herein, the term “deconvolution” refers to a computational process applied to raw image data in which optical distortions, background noise and out-of-focus signal contributions are mathematically corrected or minimized to improve spatial resolution, contrast, and signal fidelity. In certain embodiments, deconvolution is performed on z-stacks or volumetric data sets acquired from optical microscopy using algorithms such as iterative constrained deconvolution, maximum likelihood estimation, Richardson-Lucy deconvolution, or deep-learning-based restoration. In some embodiments, the point spread function of the imaging system is experimentally determined or modeled to reconstruct the distribution of fluorescence or signal intensity from within the sample. Deconvoluted images therefore exhibit enhanced resolution of subcellular structures, improved boundary segmentation, and accurate quantification of features. As used herein, the term “StarDist” refers to a deep learning architecture that models object shapes as star-convex polygons or polyhedra and is well-suited for segmentation overlapping or irregularly shaped nanoparticles in cellular environments. The StatDist model comprises a convolution neural network backbone, such as a U-net architecture, trained to predict both the object probability and radial distances from the object center to its boundary. This representation enables efficient and accurate instance segmentation of isotropic and anisotropic features, making it valuable for identifying individual nanoparticles with varied morphologies in complex biological environments. In some embodiments, the model is trained on 2-dimensional and 3-dimensional imaging data that match the imaging modality used in the nanoparticle analysis including fluorescence, brightfield, or hyperspectral microscopy. The model incorporates domain adaptation or transfer learning to improve generalizability across datasets acquired with different optical protocols. The segmented output created by StarDist comprises spatial masks for each nanoparticle. Docket No.: 381200.00102 In some embodiments, the z-stack of 3-dimensional images of cells transfected with the nanoparticle are collected using advanced optical imaging systems selected from the group consisting of laser scanning confocal microscopy, spinning disk confocal microscopy light sheet fluorescence microscopy, structured illumination fluorescence microscopy, stimulated emission depletion microscopy, total internal reflection fluorescence microscopy, or correlative microscopy techniques. In some embodiments, the extracted spatial features comprise positional, relational, and contextual information that define the location, distribution, and organization of the nanoparticle within 2- or 3-dimensional imaging volumes. Spatial features further comprise the centroid coordinates of the nanoparticle, the relative distance between the nanoparticle and a defined subcellular structure, or the angular orientation of the nanoparticle with respect to cellular axes. In some embodiments spatial features further comprise inter-nanoparticle distances, nearest neighbor connectivity, spatial density estimates, or pair-correlation function, which characterize the local or global spatial arrangement of nanoparticle populations. In certain embodiments, spatial features also comprise context-aware measures such as the localization index to quantify the nanoparticles preferential accumulation in defined areas, or radial distribution functions relative to cellular or nuclear boundaries. In some respects, spatial relationships are encoded in spatial graphs or adjacency matrices where nanoparticles are treated as nodes and spatial proximities as weighted edges. Spatial features further comprise spatial entropy measures to distinguish patterns of organized verse stochastic nanoparticle localization. In some embodiments, the extracted spectral features are statistical or intensity-based measurements associated with emission or excitation profiles of fluorescent labels used during imaging. In some embodiments, the spectral features comprise intensity measurements for each wavelength channel of a confocal, widefield, or multiphoton microscope, including but not limited to the mean, median, minimum, maximum, and standard deviation and integrated intensity values. In some embodiments, spectral features further comprise wavelength-specific characteristics including spectral centroid, spectral width, emission skewness, and kurtosis across channels. In additional embodiments, radiometric features are calculated by comparing intensity ratios between two or more wavelength channels to reveal probe responsiveness. In some embodiments, spectral bleed-through, autofluorescence, or background signal are computationally subtracted or normalized. In certain embodiments, variations due to Docket No.: 381200.00102 photobleaching, spectral unmixing artifacts, or microscope configuration differences are addressed using calibration standards or data normalization procedures. In some embodiments, the extracted morphological features comprise geometric, shape descriptors, and radial distribution metric features that define the spatial structure and form of the nanoparticle in 3-dimensions. In some embodiments, geometric features include absolute or relative measures such as volume, area, surface area, perimeter, width, height, length, convex hull area, convex hull volume, Feret diameter, maximum inscribed circle radius compactness, or roundness. In some embodiments, shape descriptors provide dimensionless or comparative metrics of particle form and complexity including compactness, elongation, sphericity, circularity, aspect ratio, shortest radii, longest radii, short-to-long ratio, and inertia ratio. Additional descriptors may include topological or structural complexity indicators such as Euler characteristics, convexity, solidity, and oriented bounding box ratio. In some embodiments, radial distribution metrics are used to describe how mass or intensity is spatially distributed from the center of the nanoparticle, and may include 25thpercentile radii, 50thpercentile radii, 75th percentile radii, radius variance, and radius entropy. These descriptors allow the classification of nanoparticles based on symmetry, anisotropy, or deviation from a reference shape. In some embodiments, rotation-invariant and scale-invariant morphological features are emphasized to enable robust feature extraction across imaging platforms, magnifications, or orientations. The morphological features may be normalized to account for variability in optical resolution or segmentation quality. As used herein, the term “classification pipeline” refers to a computational sequence or workflow comprising one or more algorithms or modules that process input data to assign the data to a predefined category or class based on extracted features. In some embodiments, the classification pipeline comprises preprocessing steps such as normalization, feature selection, dimensionality reduction, and data augmentation. In some embodiments, the classification pipeline includes supervised learning models or unsupervised learning techniques to identify class boundaries. In some embodiments, the classification pipeline is invariant to rotation, translation, scale, or illumination conditions. In some embodiments, the pipeline receives a set of spatial, spectral, and / or morphological features and outputs a class label or probability distribution indicating similarity to one or more reference classes. In some embodiments, confidence scores, embeddings, or decision trees generated by Docket No.: 381200.00102 the classification pipeline are stored or used to guide downstream actions, such as sorting, image prioritization, or experimental planning. In some embodiments, the embedding vector provides a compact numerical representation of the nanoparticle’s identity to facilitate efficient retrieval, similarity matching, and cross-referencing. In some embodiments, the similarity matching comprises computing a similarity or distance metric between the nanoparticle embedding vector and the embedding vectors stored in a catalog to identify the most similar candidates. In certain embodiments, the similarity or distance metric is selected from the group consisting of cosine distance, Euclidean distance, Mahalanobis distance, Manhattan (L1) distance, Chebyshev distance, correlation distance, Jaccard similarity, Hamming distance, or a learned metric derived from a neural network-based similarity model. In some embodiments, the similarity computation is weighted or regularized to emphasize specific vector components corresponding to spatial, spectral, or morphological features of high diagnostic value. In some embodiments, the candidate’s 3-dimensional mesh from the catalog is compared to the nanoparticles 3-dimensional mesh to assess structural similarity and shape congruence. In some embodiments, the geometric mesh comparison is performed using a shape alignment method comprising normalizing the shape and size of the nanoparticle mesh, preliminary alignment using a point cloud registration algorithm, performing a rotation search, calculating a geometric error metric between each rotation alignment, identifying and selecting the best match, and creating a single transformation matrix that is applied to the nanoparticle mesh to produce a final aligned mesh form the catalog that closely matches the position, scale, and orientation from the nanoparticle mesh. In another aspect, this disclosure provides a method of designing an aptamer sequence from a protein structure, comprising: identifying a homologous sequence to the protein structure using a sequence similarity search; converting the protein sequence into an embedding vector; comparing the embedding vector of the protein structure to an embedding vector of a protein with known RNA-binding affinity using a cosine distance; querying a database storing protein-RNA interactions to identify a validated RNA sequence that binds to the protein with known RNA-binding affinity; converting the validated RNA sequence into an embedding vector and comparing the validated RNA sequence embedding vector to an embedding vector of RNA with known protein binding interactions using a cosine distance; iteratively mutating the embedding vector of the validated RNA sequence using an Docket No.: 381200.00102 evolutionary algorithm; and evaluating a protein sequence from a protein-RNA pair and an RNA sequence from a protein-RNA pair to determine the likelihood of a binding interaction. As used herein, the term “homologous sequence” refers to a nucleic acid or amino acid sequence that shares a threshold level of sequence similarity or identity with a reference sequence, typically as determined by alignment algorithms such as BLAST, Clustal Omega, or Needleman-Wunsch. In some embodiments the homologous sequence comprises at least 70%, 80%, 90%, 95% or 98% (e.g., 70%, 71%, 72%, 73%, 74%, 75%, 76%, 77%, 78%, 79%, 80%, 81%, 82%, 83%, 84%, 85%, 86%, 87%, 88%, 89%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%, 100%) sequence identity over a region of defined length relative to the reference sequence. In some embodiments, the homologous sequence preserves functional or structural conserved domains, motifs, or secondary structure elements. In some embodiments, the homologous sequence is derived from an orthologous or paralogous gene in a different organism or is synthetically generated to mimic conserved regions of interest. In some embodiments, the validated RNA sequence embedding vector is iteratively mutated using a genetic algorithm to explore a sequence space that preserves binding affinity while optimizing other desired parameters such as binding specificity, structural stability, or chemical accessibility. In some embodiments, the genetic algorithm comprises initializing a population of embedding vectors derived from a validated RNA sequence, applying stochastic mutation and crossover operations to generate variant sequences, evaluating the fitness of each candidate sequence based on predicted binding affinity to the target protein or structural compatibility with known RNA-protein interaction motifs, and selecting high scoring candidates to propagate to subsequent generations. In some embodiments, fitness is computed using a machine-learned affinity prediction model or a structural similarity score derived from molecular docking or co-folding simulations. In some embodiments, constraints are imposed during optimization to maintain conserved structural motifs or thermodynamically favorable secondary structures. The genetic algorithm continues iterating until convergence criteria are met, such as reaching a target binding threshold, minimizing a loss function, or stabilizing the population density. In another aspect, this disclosure provides a method of designing an aptamer sequence from a starting RNA sequence, comprising: identifying an RNA sequence with significant sequence similarity to the starting RNA sequence by performing a nucleotide sequence search; Docket No.: 381200.00102 converting the RNA sequence with significant sequence similarity into an embedding vector; comparing the embedding vector of the RNA sequence with significant sequence similarity to an embedding vector of an RNA sequence with known RNA-binding activity using a cosine distance; querying a database storing protein-RNA interactions to identify a validated protein sequence that binds to an RNA sequence with known RNA-binding activity; converting the validated protein sequence into an embedding vector and comparing the embedding vector of the validated protein sequence to an embedding vector of a protein sequence with known RNA- binding interactions using a cosine distance; evaluating a protein sequence from a protein-RNA pair and an RNA sequence from a protein-RNA pair to determine the likelihood of a binding interaction; and querying a database to identify a subcellular location, a gene, and a disease associated with the protein sequence from the protein-RNA pair. In some embodiments, a database comprising experimentally validated or computationally predicted protein-RNA interaction data is queried to identify a protein sequence that binds to the RNA sequence exhibiting known RNA-binding activity. The database comprises resources such as RBPDB, ATtract, NPInter, POSTAR, or ENCODE, which contain curated interaction supported by CLIP-seq, RIP-seq, or other high-throughput binding assays. The query may involve aligning the RNA sequences against interaction entries using a nucleotide or structure-aware similarity algorithm, followed by retrieval of associated protein binding partners. In some embodiments, the database may also annotate binding affinity, RNA-binding domains, binding motifs, or interaction context. The identified protein sequence serves as a reference for further validation, modeling, or aptamer design. In some embodiments, once a protein-RNA interaction is predicted, the system queries integrated biological databases to retrieve metadata associated with the protein component in the pair. The subcellular location may be determined through reference to experimentally derived databases comprising Human Protein Atlas, Uniprot, or COMPARTMENTS which provide localization tags such as nuclear, cytoplasmic, membrane-bound, or mitochondrial. Gene identity may be assigned by mapping the protein to its canonical gene using Ensembl, HGNC, or NCBI Gene. In some embodiments, disease associations are retrieved from OMIM, DisGeNET, ClinVar, or COSMIC, indicating whether the protein is implicated in genetic disorders, cancer, neurodegeneration, or infectious disease pathways. This information can be used to prioritize aptamer designs for specific applications. Docket No.: 381200.00102 In another aspect, this disclosure provides a method of evaluating the protein sequence from the protein-RNA pair to determine the likelihood of a binding interaction, comprising: training a feed forward neural network classifier on a positive and a negative example of an RNA-protein interaction; scaling and normalizing an RNA embedding vector, a protein embedding vector, and a concatenated RNA-protein embedding vector; inputting the scaled and normalized RNA embedding vector, the scaled and normalized protein embedding vector, and the scaled and normalized concatenated RNA-protein embedding vector into the feed forward neural network classifier; and assigning a prediction score and / or a label to the protein sequence from the protein-RNA pair and the RNA sequence from the protein-RNA pair indicating their interaction potential using the feed forward neural network classifier. As used herein, the term “feedforward neural network” refers to an artificial neural network architecture in which information flows unidirectionally from an input layer through one or more hidden layers to an output layer without forming cycles or feedback loops. Each layer comprises a plurality of nodes, where each node performs a weighted summation of its inputs followed by a nonlinear activation function. The feedforward neural networks comprise, fully connected layers, wherein each node in a given layer is connected to every node in the subsequent layer. Fully connected layers comprise linear layers, bias-less dense layers, low- rank approximation layers, grouped fully connected layers, maked dense layers, sparse dense layers, orthogonal dense layers, mult-head attention layers, self-attention feedforward layers, hypernetwork-generated fully connected layers, time distributed sense layers, flatten layers, dropConnect layers, quantized sense layers, binary dense layers, tertiary dense layers, hashed dense layers, or product key memory fully connected layers. In some embodiments, the activation functions comprise rectified linear unit, leaky rectified linear unit, parametric rectified linear unit, sigmoid, hard sigmoid, swish, Mish, E-swish, funnel rectified linear unit, exponential linear unit, scaled exponential linear unit, hyperbolic tangent, hard hyperbolic tangent, maxout, softplus, softsign, binary step, binary sigmoid, ternary activation, quantized rectified linear unit, straight-through estimator, or Gaussian error linear unit. The feedforward neural network is trained by adjusting the weights of the connections between nodes using a supervised learning algorithm such as backpropagation along with an optimization method such as stochastic gradient descent, mini-batch gradient descent, momentum, Nesterov accelerated gradient, AdaGrad, RMSProp, Nadam, AMSGrad, conjugated gradient descent, L2 regularization, dropout training, or Adam. Docket No.: 381200.00102 In some embodiments, the feedforward neural network comprises single layer perceptrons, multi-layer perceptrons, convolution neural networks, residual networks, autoencoders, radial basis function networks, cascade correlation networks, functional link neural networks, or probabilistic neural networks. In some embodiments, the convolution neural networks comprise convolution layers, pooling layers, normalization layers, activation layers, upscaling layers, or fully connected layers to promote scale, rotational, and translational invariance. In some embodiments, the convolution layers further comprise standard 2D convolution layers, 1D convolution layers, 1x1 convolution layers, depthwise convolutions layers, pointwise convolution layers, dilated convolution layers, separable convolution layers, grouped convolution layers, strided convolution layers, or transpose convolution layers. In some embodiments, the pooling layers further comprise max pooling, average pooling, global max pooling, global average pooling, stochastic pooling, or mixed pooling. In some further embodiments, the normalization layers comprise batch normalization, batch renormalization, ghost batchnorm, layer normalization, RMS layer normalization, instance normalization, group normalization, switchable normalization, weight normalization, spectral normalization, filter response normalization, self-normalizing neural networks, adaptive layer normalization, conditional batchnorm, feature-wise linear modulation, batch-instance normalization, adaptive instance normalization, or pixel normalization. In some further embodiments, the upscaling layers further comprise, nearest neighbor, bilinear interpolation, bicubic interpolation, 1D interpolation, convTranspose1D, convtranspose2D, convTranspose3D, fractionally strided convolution, pixel shuffle, depth-to-space, space-to-depth, max unpooling, average unpooling, learnable interpolation, deformable upsampling, content-aware ReAssembly, spline interpolation, or Fourier upsampling. In some embodiments, the residual networks comprise convolution layers, identity and short cut connection layers, residual blocks, normalization layers, activations, pooling layers, fully connected layers, regulation layers, and downsampling mechanisms. In some embodiments, the convolution layers comprise those convolution layers mentioned herein. In some embodiments, the identify and shortcut layers further comprise identity layers, residual shortcut layers, shortcut with 1x1 convolution layers, zero-padding shortcut layers, highway network gates, learned skip connection layers, attention shortcut layers, DenseNet shortcut layers, feature aggregation shortcut layers, ResNeXt aggregated shortcut layers, squeeze-and- Docket No.: 381200.00102 excitation shortcut layers, or split attention shortcut layers. In some embodiments, the residual blocks further comprise basic residual blocks, bottleneck blocks, pre-activation blocks, full- pre-activation blocks, wide residual blocks, grouped residual blocks, dilated residual blocks, inverted residual blocks, fused MBConv blocks, squeeze-and-excitation residual blocks, ghost residual blocks, shuffle residual blocks, transformer residual blocks, perceiver residual blocks, conformer residual blocks, ResBlock in StyleGAN blocks, ResNet-GAN generator blocks, ResNet discriminator blocks, attention residual blocks, dual path blocks, multi-scale residual blocks, or split attention blocks. In some embodiments, the autoencoder comprises encoder layers, bottleneck layers, and decoder layers. In some embodiments, the autoencoder transforms input data into a structured latent or feature representation and comprises autoencoder encoders, convolution neural network-based encoders, transformer encoders, or structured encoders. In some embodiments, autoencoders encoders further comprise, vanilla autoencoder encoders, sparse autoencoder encoders, contractive autoencoder encoders, denoising autoencoder encoders, or variational autoencoder encoders. In some embodiments, convolution neural network-based encoders further comprise convolutional autoencoder encoders, DenseNet encoders, ResNet encoders, MobileNet encoders, EfficientNet encoders, UNet encoders, SegNet encoders, vision transformer patch encoders, or masked autoencoder encoders. In some embodiments, the transformer encoder comprises BERT encoders, vision transformer encoders, perceiver encoders, or transformer FFN encoders. In some embodiments, structured encoders further comprise GCN encoders, GraphSAGE encoder, or GAT encoders. In some embodiments, the bottleneck layers reduce the dimensionality of an input feature space into the layer and comprises standard bottleneck layers, convolutional neural network bottleneck blocks, transformer bottleneck layers, autoencoder specific bottleneck layers, or regularization bottleneck layers. In some embodiments, the standard bottleneck layers further comprise linear bottleneck layers, MLP bottleneck layers, projection bottleneck layers, feature bottleneck layers, or latent vector layers. In some embodiments, the convolutional neural network bottlenecks blocks further comprise 1x1 convolution bottleneck blocks, ResNet bottleneck blocks, MobileNetV2 bottleneck blocks, Squeeze layers, ghost or bottleneck blocks. In some embodiments, the transformer bottlenecks further comprise projection layers, masked autoencoder bottlenecks, or downsampling transformer blocks. In some embodiments, the autoencoder specific bottleneck layers further comprise code layers, Docket No.: 381200.00102 variational autoencoder bottleneck layers, and contractive bottleneck layers. In some embodiments, the regularization bottleneck layers further comprise information bottleneck layers, drop bottleneck layers, group bottleneck layers, or attention bottleneck layers. In some embodiments, the decoder layers reconstruct outputs from a latent or encoded feature space and comprises standard decoder layers, convolutional decoder layers, autoencoder specific decoders, transformer decoders, or specialized decoders. In some embodiments, the standard decoder layers further comprise linear layer decoders, MLP decoders, latent-to-output layers, or projection layers. In some embodiments, the convolutional decoder layers further comprise transpose convolution layers, sub-pixel decoders, bilinear / bicubic and convolutional decoders, or UNet decoder blocks. In some embodiments, the autoencoder specific decoder layers further comprise symmetric MLP decoders, sparse decoders, contractive decoders, denoising decoders, or variational decoders. In some embodiments, the transformer decoder layers further comprise feedforward decoder blocks, masked feedforward decoder layers, or latent-to-output MLP layers. In some embodiments, the specialized decoder layers further comprise embedding decoders, multi-head regression layers, gated feedforward decoder layers, attention-modulated feedforward network decoder layers, and normalization layers. In some embodiments, the autoencoder further comprises a discriminator to reinforce constraints on the latent space to improve a model’s ability to distinguish between true samples and fake samples generated by the encoder. In some embodiments, the discriminator comprises standard discriminators, discriminators based on loss functions, and latent-structured discriminators. In some embodiments, standard discriminators further comprise binary MLP discriminators, categorical MLP discriminators, multiclass discriminators, or one-class discriminators. In some embodiments, discriminators based on loss functions further comprise sigmoid cross-entropy discriminators, Wasserstein discriminators, least squares discriminators, or hinge loss discriminators. In some embodiments, latent-structure discriminators further comprise mixture discriminators, conditional discriminators, or auxiliary classifier discriminators. In some embodiments, the discriminators as described herein may further be altered to comprise deep MLP discriminators, dropout regulation discriminators, spectral normalization, gradient penalty, and batch normalization. In some embodiments, the radial basis function neural network comprises radial basis function as activation functions to compute a response based on the distance between the input Docket No.: 381200.00102 and a prototype center. In some embodiments, the radial basis functions are calculated by determining a center and a width and comprise Gaussian, multiquadratic, inverse multiquadratic, linear, cubic, thin plate spline, bump function, Cauchy, Matern, Wendland, polyharmonic splines, exponential, and inverse quadratic functions. In some embodiments, the center selection methods comprise random sampling, fixed grid, domain knowledge, k-means clustering, Gaussian mixture model, fuzzy C-means, spectral clustering, gradient-based optimization, expectation maximization algorithm, prototype learning, self-organizing maps, one center per training example, greedy selection, principal component analysis initialization, evolutionary algorithms, reinforced learning inspired methods, or active learning-based methods. In some embodiments, the method of determining width comprises single global method, heuristic global rule, mean distance between centers, local distance to nearest neighbor, average distance to k-nearest centers, distance to training samples, clustering variance, backpropagation, Bayesian optimization, genetic algorithms, swarm optimization, expectation maximum, adaptive bandwidth estimation, Mahalanobis distance, or data-drive multi-scale widths. In some embodiments, the distance used in the width algorithm is calculated by L1 normalization, L2 normalization, squared Euclidean distance, Mahalanobis distance, Chebyshev distance, cosine distance, Bray-Curtis distance, Minkowski distance, or Hamming distance. In some embodiments, the cascade-correlation neural network comprises cascade hidden units that are incrementally added during training by the guidance of an objective function to grow the network. The cascade hidden elements are initially connected to the input nodes but connect to all previously hidden units during training. In some embodiments, the objective functions comprise output error objective functions and candidate unit objective functions. In some embodiments, the output error objective functions comprise mean square error, cross-entropy loss, Huber loss, binary cross-entropy, or KL divergence. In some embodiments, the candidate unit objective function comprises correlative maximization, covariance objective, dot product with error, negative mean square error between error and candidate output, L1-regularization correlation, L2-regularized correlation, or information gain. In some embodiments, the functional link neural network comprises functional expansion layers that enrich the feature space using a set of basis functions. In some embodiments the functional expansions comprise polynomial series expansions, trigonometric Docket No.: 381200.00102 expansions, radial basis function expansions, tensor expansions, custom nonlinear function expansions, or orthogonal expansions. In some embodiments, the polynomial series expansions further comprise monomial expansions, Legendre polynomial expansions, Chebyshev polynomial expansions, Hermite polynomial expansions, Laguerre polynomial expansions, or Gegenbauer polynomial expansions. In some embodiments, the trigonometric expansions further comprise wavelet expansions, Fourier series expansions, or Walsh / Hadamard expansions. In some embodiments, the radial basis function further comprises Gaussian radial basis function expansions, multiquadratic expansions, inverse quadratic expansions, thin plate spline expansions. In some embodiments the tensor expansions further comprise cross-term expansions or higher-order polynomial expansions. In some embodiments, the custom nonlinear function expansions further comprise exponential expansions, logarithmic expansions, sigmoid / hyperbolic tangent expansions, rectified linear unit and leaky rectified linear unit expansions, or power law expansions. In some embodiments, the orthogonal expansions further comprise Gram-Schmidt orthogonalization or data-driven basis expansions. In some embodiments, probabilistic neural networks comprise pattern layers that contain one node per training sample and summation layers that estimate the likelihood of class membership for the input. In some embodiments, the pattern layers comprise Gaussian radial basis function layers, L1 radial basis function layers, multquadratic radial basis function layers, inverse multiquadratic radial basis function layers, polyharmonic spline layers, thin plate layers, Cauchy kernel layers, Leplacian kernel layers, triangular kernel layers, Epanechnikov kernel layers, sigmoidal activation pattern layers, wavelet-based pattern layers, custom learned distance pattern layers, Fourier layers, spectral similarity layers, or functional link pattern layers. In some embodiments, the summation layers comprise standard summation layers, normalized summation layers, weighted summation layers, SoftMax layers, max pooling layers, mean pooling layers, sum across features layers, hierarchical summation layers, attention-weighted summation layers, probabilistic aggregation layers, mixture density aggregator layers, or winner-take-all summation layers. A. Protein Concentration In another aspect, this disclosure provides a method of measuring a concentration of an intracellular protein, comprising: introducing the nanoparticle containing the first probe to allow the first probe to bind to the intracellular protein inside the cell; isolating the first probe bound to the intracellular protein using a magnetic field; fixing or lysing the cell; introducing Docket No.: 381200.00102 the nanoparticle containing the second probe into the cell to allow the second probe to bind to the intracellular protein on the first probe; determining the quantity and identity of the second probe by the barcode associated therewith by an image analysis. In some embodiments, the concentration of the intracellular protein is measured using the fluorescent imaging to determine abundance or absence of the intracellular protein. In some embodiments, the identity and abundance or absence of the intracellular protein is determined by dissolving the second probe having the aptamer and eluting the aptamer; and sequencing the eluted aptamer. In some embodiments, a third probe is used to determine the presence of a proteoform. The third probe may be designed to recognize a post-translational modification such as phosphorylation, ubiquitination, acetylation, glycosylation, or methylation or to distinguish between isoforms generated by alternative splicing or proteolytic processing. In some embodiments, the third probe comprises an aptamer, antibody, or peptide ligand that binds selectively to the modified variant region of the proteoform. In some embodiments, the third probe is tagged with a distinct optical barcode or fluorescent label to enable simultaneous and sequential detection alongside the first and second probes. In some embodiments, the third probe binds to the protein on the first probe after the second probe is dissolved, leaving behind its bound moiety on the POI. In some embodiments the moiety of the second probe on the POI is removed before the third probe is introduced. In some embodiments, binding of the third probe is analyzed by image-based colocalization with the first and second probe, Förster resonance energy transfer analysis, or fluorescence intensity profiling to determine the relative abundance of spatial distribution of the proteoform within the cell. In some embodiments, the presence of the proteoform is further validated by eluting the third probe and the bound POI and performing sequencing or mass spectrometry on the associated aptamer, peptide label and / or protein. Isolating Tethered Cells using Nanoparticle Compositions In another aspect, this disclosure provides a method of isolating tethered cells comprising: adding a first probe comprising a magnetic nanoparticle with a binding moiety that is identifiable by the morphology, size, shape and / or color of the nanoparticle; the first probe tethered to a T cell receptor-complex on the surface of a T cell; isolating the T cell using a magnetic field; identifying a barcode on the probe; and dissolving the probe to selectively retrieve the T cell Docket No.: 381200.00102 In some embodiments, the probe comprises an aptamer, antibody, or peptide ligand that selectively recognizes a cell surface protein that remains localized at the cell membrane such as a target cell or a tumor cell and does not induce internalization upon ligand binding. In some embodiments, the recognized cell surface receptors comprise cluster of differentiation markers such as CD47, CD24, CD31, beta-2 macroglobulin (B2M), programmed death-ligand 1 (PD- L1), Hepatitis A virus cellular receptor 2 (HAVCR2), or major histocompatibility complex class I (MHC I). These membrane-associated proteins are selected for their ability to support prolonged surface retention of the tethered probe. In some embodiments, cells that are tethered to the probes are subjected to sorting processes to separate target activated T cells from a heterogeneous population. In some embodiments, cell sorting techniques comprise magnetophoresis, fluorescence activated cell sorting, or immunoaffinity chromatography. Magnetophoresis may be performed using magnetic nanoparticles conjugated to the probes, and by applying an external magnetic field to isolate labeled cells. Fluorescence activated cell sorting may be employed when the probes or associated nanoparticles are conjugated to fluorophores that permit high-throughput optical detection and sorting. In some embodiments, immunoaffinity chromatography is used by passing the cell suspension through a column functionalized with ligands specific to the surface-bound probes, thereby enabling selective retention and elution of target cells. In some embodiments, the probe includes a barcode that enables downstream identification. The barcode may be read by image-based analysis- fluorescence imaging, or sequencing. In some embodiments, the probes are designed to be differentially dissolvable or cleavable, such that the tether between the nanoparticle and the cell may be selectively broken to recover the labeled cells after sorting or imaging. In certain embodiments, the cells are magnetically isolated using iron oxide embedded nanoparticles functionalized with binding agents and subjected to a magnetic field to selectively capture or deplete a specified subpopulation. Additional Definitions To aid in understanding the detailed description of the compositions and methods according to the present disclosure, a number of express definitions are provided to facilitate unambiguous interpretation of the various aspects of the disclosure. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains. Docket No.: 381200.00102 The term “nucleic acid,” may refer to a polymer composed of a multiplicity of nucleotide units (ribonucleotide, deoxyribonucleotide, or related structural variants) linked via phosphodiester bonds, including but not limited to, DNA or RNA. The term encompasses sequences that include any of the known base analogs of DNA and RNA. Examples of nucleic acids include, and are not limited to, mRNA, miRNA, tRNA, rRNA, snRNA, siRNA, dsRNA, cDNA and DNA / RNA hybrids. Nucleic acids may be single stranded or double stranded, or may contain portions of both double stranded and single stranded sequences. The nucleic acid may be DNA, both genomic and cDNA, RNA, or a hybrid, where the nucleic acid may contain combinations of deoxyribo- and ribo-nucleotides, and combinations of bases including uracil (U), adenine (A), thymine (T), cytosine (C), guanine (G), and their derivative compounds. Nucleic acids may be obtained by chemical synthesis methods or by recombinant methods. The depiction of a single strand also defines the sequence of the complementary strand. Thus, a nucleic acid also encompasses the complementary strand of a depicted single strand. Many variants of a nucleic acid may be used for the same purpose as a given nucleic acid. Thus, a nucleic acid also encompasses substantially identical nucleic acids and complements thereof. The term “protein” refers to polymers of amino acids of any length. The polymer may be linear or branched, it may comprise modified amino acids, and it may be interrupted by non- amino acids. The term also encompasses an amino acid polymer that has been modified; for example, disulfide bond formation, glycosylation, lipidation, acetylation, phosphorylation, pegylation, or any other manipulation such a conjugation with a labelling component. As used herein, “amino acid” includes natural and / or unnatural or synthetic amino acids, including glycine and both the D or L optical isomers, and amino acid analogs and peptidomimetics. The term “peptide” may refer to peptide compounds containing two or more amino acids linked by the carboxyl group of one amino acid to the amino group of another, to form an amino acid sequence. Peptides may be purified and / or isolated from natural sources or prepared by recombinant or synthetic methods. The term “antibody” (Ab) is used in the broadest sense, and specifically may include any immunoglobulin, whether natural, or partly, or wholly synthetically produced, including, but not limited to monoclonal antibodies, polyclonal antibodies, multispecific antibodies (for example, bispecific antibodies and polyreactive antibodies), and antibody fragments. Thus, the term “antibody,” as used in any context within this specification, is meant to include, but not be limited to, any specific binding member, immunoglobulin class and / or isotype (e.g., IgG1, Docket No.: 381200.00102 IgG2a, IgG2b, IgG3, IgG4, IgM, IgA1, IgA2, IgD, and IgE), and biologically relevant fragment, or specific binding member thereof, including, but not limited to, Fab, F(ab’)2, scFv (single chain or related entity) and (scFv)2. The terms “cell” and "cells" refer to a eukaryotic cell characterized by a membrane bound nucleus and other membrane bound organelles that compartmentalize cellular functions. Preferably, the cells are eukaryotic cells that are derived from multicellular animals such as mammals and may be adherent, semi-adherent, or suspension cells. As used herein, “adherent cells” are cells that grow while attached to a surface in a liquid medium, “suspension cells” are cells nonadherent cells that grow in a liquid medium, “semi-adherent cells” are cells that exhibit characteristics of both adherent and suspension cells in that they can grow while attached to a surface but also detach and grow in suspension. The term “transfect” refers to the process of introducing nucleic acids into a cell. Preferably, the nucleic acids are delivered using the nanoparticle described herein. The term “aptamer” refers to single stranded nucleic acid molecules that can associate with targets, regardless of manner of target recognition. The term “biodegradable” refers to a material that can be broken down by biological means. The term “embedding vector” refers to a numerical representation of data sets occupying a multi-dimensional space that captures key features of the data set wherein similar data points will have vectors closer within this space. The distance between embedding vectors can be measured using a cosine distance. As used herein, “cosine distance” measures how dissimilar the orientation of two vectors is in a multi-dimensional space. The term “fitness function” refers to an objective function that guides an evolutionary process by evaluating how close a candidate solution is to achieving a specific goal. The term “effective amount” refers to the amount of an active compound / agent that is required to confer a therapeutic effect on a treated subject. Effective doses will vary, as recognized by those skilled in the art, depending on the types of conditions treated, route of administration, excipient usage, and the possibility of co-usage with other therapeutic treatment. The terms “stable” and “stability” refer to a material that does not degrade at room temperature. Docket No.: 381200.00102 As used herein, the term “in vitro” refers to events occurring in an artificial environment, such as in a test tube, reaction vessel, or cell culture, rather than within a multicellular organism. As used herein, the term “in vivo” refers to events occurring within a multicellular organism, such as a non-human animal. It is noted that, as used in this specification and the appended claims, the singular forms “a,” “an,” and “the” include plural referents unless the context clearly dictates otherwise. The terms “including,” “comprising,” “containing,” and “having,” and variations thereof, are intended to encompass the listed items and equivalents thereof, as well as additional subject matter, unless otherwise specified. The phrases “in some embodiments,” “in various embodiments,” and similar expressions may be used repeatedly throughout the disclosure. Such phrases do not necessarily refer to the same embodiment, although they may, unless the context dictates otherwise. The terms “and / or” or “ / ” refer to any one of the listed items, any combination of the listed items, or all of the listed items. The term “substantially” does not exclude “completely.” For example, a composition described as “substantially free” of component Y may, in fact, be completely free of Y. Where appropriate, the term “substantially” may be omitted from the definition of the invention. As used herein, the terms “approximately” or “about,” when applied to one or more values of interest, refer to values that are similar to a stated reference value. In some embodiments, the term “approximately” or “about” refers to a range of values within 25%, 20%, 19%, 18%, 17%, 16%, 15%, 14%, 13%, 12%, 11%, 10%, 9%, 8%, 7%, 6%, 5%, 4%, 3%, 2%, 1%, or less in either direction (greater than or less than) of the stated reference value, unless otherwise indicated or evident from the context (except where such a range would exceed 100% of a possible value). Unless indicated otherwise herein, the term “about” is intended to include values proximate to the recited range that are equivalent in terms of the functionality of the individual ingredient, the composition, or the embodiment. It is to be understood that wherever numerical values or ranges are provided herein, all values and ranges encompassed thereby are intended to fall within the scope of the invention. Moreover, all values that fall within such ranges, including the endpoints, are contemplated by the present disclosure. Docket No.: 381200.00102 As used herein, the term “each,” when referring to a collection of items, identifies individual items within the collection, but does not necessarily refer to all items in the collection, unless explicitly stated or dictated by context. The use of examples or exemplary language (e.g., “such as”) is intended solely to illuminate aspects of the invention and should not be construed as limiting the scope of the invention unless explicitly recited in the claims. The term “exemplary” is intended to mean “by way of example” and does not imply preference or requirement. All methods described herein may be performed in any suitable order unless otherwise indicated or clearly contradicted by context. Where a method comprises multiple steps, such steps may be performed sequentially or simultaneously. Unless otherwise noted, any order of steps described is not intended to be limiting. Where a method comprises a combination of steps, all combinations or sub- combinations of those steps are intended to be encompassed by the disclosure, unless specifically excluded. Each publication, patent application, patent, and other reference cited herein is incorporated by reference in its entirety, to the extent not inconsistent with the present disclosure. These references are provided solely for their disclosure as of the filing date of the present application. Nothing herein should be construed as an admission that the present invention is not entitled to antedate such references by virtue of prior invention. Furthermore, the publication dates provided herein may differ from the actual dates of public availability and may require independent verification. It is understood that the examples and embodiments described herein are provided for illustrative purposes only, and that modifications and variations thereof will be apparent to those skilled in the art. Such modifications are considered to fall within the scope and spirit of the present disclosure and the appended claims. Examples The present disclosure is not to be limited in scope by the specific embodiments described herein. Indeed, various modifications of the invention, in addition to those described herein, will become apparent to those skilled in the art from the foregoing description and the accompanying figures. Such modifications are intended to fall within the scope of the appended claims. Docket No.: 381200.00102 EXAMPLE 1 Materials and Methods Used for Probe and Membrane Composition Membranes: SM-102 cationic ionizable lipid / DSPC / cholesterol / PEG-DMG2K (50 / 10 / 38.5 / 1.5 mol%) (Provided by Helix Biotech, USA) (Figure 1). Probes: Absolute Mag Methyl coated iron oxide nanoparticles (CD Bioparticles, USA) NLC probes prepared in-house comprising: solid lipid e.g., glyceryl monostearate (Spectrum Chemical, USA), liquid lipid e.g., oleic acid (Sigma-Aldrich India) (Figure 3) Surfactants: Poloxamer 188 (ThermoFisher,Belgium), TWEEN 80 (Thermofisher, USA) Dyes: Rhodamine Phalloidin (Sigma-Aldrich India), BODIPY 500 / 510 C1 C12 (Invitrogen USA ) Manufacturing method for membranes: Jet Impingement Mixer (Helix Biotech, USA) (iron oxide nanoparticles or nanostructured lipid carriers in aqueous phase and lipids with PEG, cholesterol, and DSPC in organic solvent phase). NLC preparation method: Lipids (solid and liquid) were combined in a 50:50 ratio by weight and heated together at 70-75°C (about 10°C above their melting temperature). Surfactants are dissolved in water (2% Poloxamer 188 and 2% Tween 80) and heated to 90- 95°C. Dyes are incorporated either into the lipid phase or the aqueous phase depending on whether they are hydrophobic or hydrophilic. Similarly, iron oxide nanoparticles are added to the lipid or aqueous phase depending on their wettability characteristics. The heated surfactant solution is then added slowly to the melted lipid mixture under homogenization using a high- speed homogenizer (Polytron PT 1200, Avantor Sciences, USA). The mixture is then allowed to homogenize at high speed (10,000 rpm to 20,000 rpm) for 10 minutes to form a coarse emulsion. Depending on the size of the NLC probes required the coarse mixture is ultrasonicated using a Labquake CV 33 (Sonics and Materials, USA) at 80% for 5-10 minutes. The mixture is then allowed to cool to room temperature to attain the NLC probes. The pH of the resulting solution is adjusted to 7.4 using dilute sodium hydroxide (NaOH) or hydrochloric acid (HCl) to maintain long-term stability. Purification of probes after conjugation to binding moieties: After the probes are conjugated to their binding moieties (physical adsorption: 20 μg of antibody or aptamer: 1 mL of probes, incubated for 20 hrs), they are dialyzed using a dialysis chamber in 1L of deionized Docket No.: 381200.00102 water for 24 hrs. A magnetic field is then applied to separate the magnetic probes from any residual antibody or aptamer in solution. Probe Characterization: Probes were characterized using a transmission electron microscope (TEM) (JEOL 1400+ with Gatan One View, JEOL, USA) (Figure 2) and by using dynamic light scattering (Zetasizer Nano, Malvern Instruments, USA). Imaging: Probe solutions were diluted between 100-1000 times their stock concentration and seeded into cells for at least 2hrs (Figure 4, Figure 5A and Figure 8B) and preferably for 12hrs (Figure 5B and Figure 8A). Cells containing the probes were fixed using 4% PFA. Probes were then imaged using a confocal microscope (SP8, Leica, USA) a STED module (3D STED module, Aberrior, USA) was also used to verify true particle location and compared to confocal images. Imaging is preferably performed at a 100 nm step size with a 100x oil objective having a numerical aperture (NA) of 1.4. Probes with different conjugated dyes were also mixed together and seeded into cells to demonstrate their visualization with different spectral filters (Figure 7 and Figure 11). Identity extraction of probes after confocal imaging using Machine Learning: A confocal Z-stack of the sample slide was acquired to capture the spatial distribution, spectral signature, and morphological shape of the embedded nanoparticles. The acquisition of the Z- stack enables downstream analysis of both labeling and 3D structure. For this example, imaging was performed using a Leica SP8 confocal microscope equipped with an HC PL APO CS2100× / 1.40 OIL objective, which provides high-resolution, volumetric imaging suitable for characterizing sub-micron particles. The acquired volume had voxel dimensions of 45.429 nm × 45.429 nm × 201.67 nm, with a total image size of 2008 × 2008 × 5 pixels. The imaging system captured 16-bit resolution images, providing a high dynamic range for accurate intensity quantification. Individual particles ("blobs") were segmented using a neural network–based model, such as StarDist, which separates them from the background and neighboring structures. In example experiments described in this patent, a custom model trained on synthetic data using the StarDist framework segments nanoparticles. For illustration, the first image (Figure 19A) represents the original 3D image stack acquired via confocal microscopy, while the second image (Figure 19B) shows the segmentation result, with each detected particle distinctly labeled. The first image stack (Figure 19A) displays BODIPY-labeled nanoparticles, which were imaged using confocal microscopy. The data was Docket No.: 381200.00102 tiled and resampled to achieve a resolution suitable for instance segmentation, with voxel dimensions of 44 nm x 44 nm x 131^nm. The nanoparticles were converted into a 3D mesh model (Figure 19C) to capture and preserve their structural geometry. These mesh representations serve as the basis for subsequent spatial, morphological, and comparative analyses in later stages. The resulting 3D models (Figure 19C) provide a standardized format for cross-phase comparison. Specifically, nanoparticles imaged during the outside cells are associated with those detected in inside cells by evaluating similarities in mesh structure and derived morphological descriptors. Each segmented nanoparticle was analyzed to extract key features from its unique spectral and morphological profile. These features establish correspondence between the particles imaged either inside or outside cells. Intensity statistics, including mean, median, minimum, maximum, and standard deviation, are computed for each wavelength channel to characterize the spectral barcode for example for probes used in Figure 6 and Figure 7. In parallel, a comprehensive set of morphological descriptors were extracted. These include: • Geometric features such as surface area, volume, width, and height • Shape descriptors such as compactness, elongation, shortest and longest radii, the short- to-long radius ratio, inertia ratio, Euler characteristic, and oriented bounding box (OBB) ratio Radial distribution metrics, including the 25th, 50th, and 75th percentile radii, radius variance, and radius entropy, were also calculated to capture shape complexity and internal structure. Each mesh object was assigned a unique identifier (Figure 19C) and stored in an experiment-specific database along with its associated spectral barcode, morphological descriptors, and relevant bound surface moieties. A mesh embedding was computed separately using a graph mesh encoder, which processes only the 3D morphological mesh to generate a Docket No.: 381200.00102 compact numerical representation of the nanoparticle's structure. This embedding served as a similarity-based key within the database, enabling efficient cross-referencing, matching, and retrieval of particles, particularly during downstream classification and tracking. Each nanoparticle was matched to its counterpart outside the cell using a combination of its spectral barcode, morphological descriptors, and geometrical structure derived from its 3D mesh. The spectral and morphological features narrow down candidate matches before a detailed geometrical comparison. Matching is done in two steps: 1. Feature Embedding Similarity Search A feature embedding was computed for each nanoparticle based on its mesh and morphological attributes. Cosine distance was then used to identify the most similar candidates from the in vitro catalog. 2. Geometric Mesh Comparison Each candidate's 3D mesh was compared against the query particle's mesh to assess structural and spectral similarity. This step confirmed the match by evaluating shape congruence and particle spectral characteristics between the particle and its potential cataloged counterparts. EXAMPLE 2 Using Nanoparticles Embedded in Membranes for Live Cell Experiments In this workflow, it is important to delineate proteins that are within the cell and proteins that are extracellular. Also, the spatial context of the proteins that are captured within these complex matrices is important (for example to decipher which T cells are active or the proteins that are produced in tumor cells that are eventually broken into peptides and presented on the MHC receptor surface of tumors). If the probes (either organic or inorganic) that are labeled with capture aptamers or antibodies are directly incubated with the cell line without encapsulating them within the membrane, they will immediately start to accumulate extracellular proteins around them (protein corona effect). If this happens, it becomes virtually impossible to define a spatial context to where these proteins are produced within the cell Docket No.: 381200.00102 culture as well as which cells are producing them. Since the entry of the probes into the live cells relies on active cell entry methods such as endocytosis, pinocytosis or phagocytosis, cells are cultured with the membranes typically for a few hours (Figure 5). The use of the membrane provides a protective layer to prevent direct absorption / binding of extracellular proteins onto the probes, thus preserving the spatial and intracellular context of proteins that bind to the probes. Once the probes that are protected by the membrane enter the cell, the natural process of lipolysis breaks down the membranes releasing the contained probes within the cell cytoplasm, allowing the probes to only bind to proteins / peptides within the cell. Through image analysis it is also possible to determine which probes enter the cells and at what time thereby allowing the interrogation of temporal protein production which is useful when monitoring drug action or response of a T cell to a target cell (e.g. activation of methylation factors inside the T cell) (Figure 11). To determine the identity of the probes (i.e., what aptamers or antibodies are attached to their surface), they can be labeled with fluorescent dyes. While a few fluorescent iron oxide nanoparticles (IONP’s) are commercially available only a dozen different color signatures can be obtained. Organic probes (e.g., nanostructured lipid carriers) were thus designed, in which several different lipids were used in their matrix, their imperfect crystal lattice enables obtaining theoretically an infinite number of identity signatures based on the unique shapes they can form. These features are used to identify what binding moiety they are conjugated with, which is known during the manufacturing and conjugation of the NLC’s. These features can be identified through machine learning methods as previously described. In addition, the NLC’s were rendered magnetic so they can be pulled down after the cells are lysed to preserve spatial information. There are two main methods to quantify the proteins / other binding partners to the surface of the probes. In one method, as shown in Figure 9, a donor probe or the primary probe within the membrane is incubated with live cells. The membranes containing the primary probes are actively taken into the cell and the membrane is dissolved by lipolysis within the cell. The probes are then released, which in turn bind to their respective binding partners present within the cell depending on their binding moiety. The primary probe that are used in this workflow can vary in size between 10 nm to 800nm, since their entry into the cell is accomplished by active cell ingestion, sizes can be varied with a larger range. Once a sufficient amount of probes enter the cells, two methods can Docket No.: 381200.00102 be used to quantify probe binding, namely, 1) using secondary probes to bind to the primary probes by directly allowing them to be ingested into the cell using active cell processes along with the primary probes; and 2) using a uniform magnetic field to capture the primary probes at their relative place within the cell after the experiment followed by cell lysis and washing. The secondary probes are then added and allowed to bind to the primary probes. After either of these two processes are completed, one workflow to quantify the bound moieties involves allowing the primary probe to bind to the secondary probe that contains a binding moiety that attaches to a complementary site of the protein bound to the primary probe, the type of binding moiety is identified by the shape, size and color signature of both probes (primary would have one binding moiety in this example and secondary would have a second binding moiety) as shown in Figure 9B. The proteins can thus be directly quantified by using FRET intensities and probe identities. In the second workflow as shown in Figure 10 the secondary probe has an oligo barcode that contains information on the type of binding moiety on its surface and a unique spatial barcode that is cataloged during the imaged split and pool process as shown in Figure 12 and Figure 13. This gives us spatial information (probe and its oligo barcode identity and its relative position in the image) as well as the identity of the binding moiety on the secondary probe. Once binding to the secondary probe is complete, the secondary probe is degraded using a lipolysis reagent exposing only the bound moieties attached to the target on the primary probe. This is followed by washing away all non-bound moieties of the secondary probe. This degradation rate of the secondary probe can be varied by changing the lipid composition (e.g., some lipids such as short chain triglycerides undergo quick degradation while longer chains take a longer time). Thus, several different types of probes can be made that can differentially degrade based on incubation time or concentration of the degrading agent allowing the ability to collect separate fractions of bound barcoded oligonucleotides. The moieties bound to the target via the secondary probe may be selectively released from the target, for example, by introducing a competitive binding agent or by denaturing the target under controlled conditions (e.g., thermal, chemical, or enzymatic treatment). Following release, the detached moieties can be subjected to sequencing or other analytical methods to determine both the identity (i.e., type) and spatial distribution of the binding moieties based on their oligo barcodes. This dual characterization enables quantitative assessment of the concentration of the target molecule within the sample. Such an approach provides an Docket No.: 381200.00102 additional layer of verification and enhances the accuracy of the detection and quantification process. EXAMPLE 3 Using Nanoparticles Directly without Membranes for Formalin-Fixed Paraffin- Embedded (FFPE) Tissues In the case of using the probes for dead tissue (FFPE or frozen), the possibility for extracellular proteins binding to the probes is minimal or non-existent. In this case, since there are no active transport processes, only passive transport of the probes into the cells is possible. To quantify intracellular proteins in this example probes that are small enough to enter through membrane pores are used (typically 5 to 30nm). The permeation of the probes into the tissue are enhanced using permeabilization reagents such as surfactants (Triton-X) or organic solvents such as methanol or acetone. The probes, which are preferably fluorescently labelled, are incubated with the tissue typically overnight (Figure 14). The probes have binding moieties attached to their surface and can be magnetically actuated. The primary probes diffuse into the cells in the tissue and bind to their target (Figure 14). Once incubation is complete, a uniform magnetic field is actuated, and the probes are held in place within the tissue at the site where they are bound to their target (Figure 15). Cell / tissue lysis and / or solubilization reagents such as CellLytic M or RIPA buffer are used to dissolve the cell / tissue, while the primary probes that are attached to proteins / mRNA or other binding partners of interest are left behind by being held down by a uniform magnetic field in their relative spatial positions corresponding to their spatial positions inside the cells / tissue. The secondary probe which is the NLC in this case is then added to the slide and allowed to bind to the primary probes that have bound targets from the tissue (Figure 16). The identity of the secondary probe can be deciphered from the shape, size and color which is associated with a unique Oligo barcode that contains information on the type of binding moiety and a unique spatial identifier derived during the imaged split-pool process (Figure 12 and Figure 13). Once the identity of the secondary probe is deciphered: the secondary probe is degraded using a lipolysis agent leaving only the bound moieties from the secondary probe on the surface of the primary probe. The quantification of the target the secondary probe was bound to can then be determined using sequencing. Similarly, quantification of the target can also be carried out using imaging techniques as shown in Figure 9B. EXAMPLE 4 Docket No.: 381200.00102 Machine Learning Prediction of RNA Aptamer to a Protein of Interest (Figure 18) Step 1: Select Target Protein The workflow begins by selecting a target protein of interest. For example, UniProt accession O43426 is chosen as the target protein, and its amino acid sequence is provided as input to the workflow. Select target: O43426 Sequence: MAFSKGFRIYHKLDPPPFSLIVETRHKEECLMFESGAVAVLSSAEKEAIKGTYSKVLD AYGLLGVLRLNLGDTMLHYLVLVTGCMSVGKIQESEVFRVTSTEFISLRIDSSDEDRI SEVRKVLNSGNFYFAWSASGISLDLSLNAHRSMQEQTTDNRFFWNQSLHLHLKHYG VNCDDWLLRLMCGGVEIRTIYAAHKQAKACLISRLSCERAGTRFNVRGTNDDGHVA NFVETEQVVYLDDSVSSFIQIRGSVPLFWEQPGLQVGSHRVRMSRGFEANAPAFDRH FRTLKNLYGKQIIVNLLGSKEGEHMLSKAFQSHLKASEHAADIQMVNFDYHQMVKG GKAEKLHSVLKPQVQKFLDYGFFYFNGSEVQRCQSGTVRTNCLDCLDRTNSVQAFL GLEMLAKQLEALGLAEKPQLVTRFQEVFRSMWSVNGDSISKIYAGTGALEGKAKLK DGARSVTRTIQNNFFDSSKQEAIDVLLLGNTLNSDLADKARALLTTGSLRVSEQTLQS ASSKVLKSMCENFYKYSKPKKIRVCVGTWNVNGGKQFRSIAFKNQTLTDWLLDAPK LAGIQEFQDKRSKPTDIFAIGFEEMVELNAGNIVSASTTNQKLWAVELQKTISRDNKY VLLASEQLVGVCLFVFIRPQHAPFIRDVAVDTVKTGMGGATGNKGAVAIRMLFHTTS LCFVCSHFAAGQSQVKERNEDFIEIARKLSFPMGRMLFSHDYVFWCGDFNYRIDLPN EEVKELIRQQNWDSLIAGDQLINQKNAGQVFRGFLEGKVTFAPTYKYDLFSDDYDTS EKCRTPAWTDRVLWRRRKWPFDRSAEDLDLLNASFQDESKILYTWTPGTLLHYGRA ELKTSDHRPVVALIDIDIFEVEAEERQNIYKEVIAVQGPPDGTVLVSIKSSLPENNFFD DALIDELLQQFASFGEVILIRFVEDKMWVTFLEGSSALNVLSLNGKELLNRTITIALKS PDWIKNLEEEMSLEKISIALPSSTSSTLLGEDAEVAADFDMEGDVDDYSAEVEELLPQ HLQPSSSSGLGTSPSSSPRTSPCQSPTISEGPVPSLPIRPSRAPSRTPGPPSAQSSPIDAQP ATPLPQKDPAQPLEPKRPPPPRPVAPPTRPAPPQRPPPPSGARSPAPTRKEFGGIGAPPS PGVARREMEAPKSPGTTRKDNIGRSQPSPQAGLAGPGPAGYSTARPTIPPRAGVISAP QSHARASAGRLTPESQSKTSETSKGSTFLPEPLKPQAAFPPQSSLPPPAQRLQEPLVPV AAPMPQSGPQPNLETPPQPPPRSRSSHSLPSEASSQPQVKTNGISDGKRESPLKIDPFED LSFNLLAVSKAQLSVQTSPVPTPDPKRLIQLPSATQSNVLSSVSCMPTMPPIPARSQSQ Docket No.: 381200.00102 ENMRSSPNPFITGLTRTNPFSDRTAAPGNPFRAKSEESEATSWFSKEEPVTISPFPSLQP LGHNKSRASSSLDGFKDSFDLQGQSTLKISNPKGWVTFEEEEDFGVKGKSKSACSDL LGNQPSSFSGSNLTLNDDWNKGTNVSFCVLPSRRPPPPPVPLLPPGTSPPVDPFTTLAS KASPTLDFTER (SEQ ID NO: 1) Step 2: Protein Sequence Similarity Search Proteins similar to the target are found by performing a sequence similarity search. A tool such as BLAST (Basic Local Alignment Search Tool) finds proteins with homologous sequences to the target protein, which often share similar functions or binding partners. Selecting proteins closely related to the target (e.g., those with high sequence identity or conserved domains) increases the likelihood of discovering proteins with known RNA-binding interactions. Step 3: Protein Cosine Similarity Search For each protein retrieved in the previous step, a numeric embedding vector is computed and compared to the embeddings of proteins with known RNA-binding activity. Proteins whose embeddings have a cosine distance below a threshold (e.g., 0.1) to known RNA-binding proteins are retained as likely candidates. Cosine similarity provides a context-aware measure of sequence relatedness, capturing functional and structural similarities beyond simple sequence identity. Using embedding-based filtering can significantly speed up the search and maintain accuracy in identifying relevant proteins. (Embeddings can be precomputed and stored for efficient comparison.) In this example, one similar protein (accession Q8CHC4) was identified as meeting the cosine similarity criterion, indicating it is closely related to the target and is known to bind RNA. Step 4: Protein-RNA Interaction Search Using the filtered protein candidates, a search for any experimentally validated RNA molecules is performed. Databases of known protein–RNA interactions are queried to retrieve RNAs that bind to these proteins. These RNAs are likely binding partners for the original target protein by homology of interaction (the interolog principle). Step 5: RNA Cosine Similarity Search Docket No.: 381200.00102 Next, each RNA obtained from the previous step is represented as an embedding vector and compared against a background of RNA sequences with known protein-binding interactions. Any RNA embedding within a cosine distance cutoff (e.g., <0.2) of known interacting RNAs is kept for further analysis. This step expands the pool of candidate RNAs to include those directly found by sequence alignment and those embeddings near known binders, potentially capturing more remote analogs. (RNAs at distance zero would represent previously validated interactions.) Step 6: Evolutionary Sequence Expansion An evolutionary algorithm is employed to expand and optimize the set of candidate RNA sequences. In this approach, new RNA sequences are generated by iterative mutation and selection, guided by a fitness function that rewards high cosine similarity to the reference set of known interacting RNAs. This step enables the exploration of novel RNA variants (for example, lengths of 20 to 200 nucleotides) that may bind the protein beyond the initially known sequences. Step 7: Interaction Prediction Each unique protein–RNA pair compiled through the above steps is then evaluated by a predictive model to determine the likelihood of a binding interaction. A feedforward neural network (FNN) classifier, trained on known protein–RNA interaction data, assigns a score or label to each candidate pair, indicating whether the pair is predicted to interact. In this workflow, the FNN integrates features from the protein and RNA sequences (or their embeddings) and outputs a binding prediction. Pairs with high confidence scores are carried forward as potential interactions for validation.

Claims

Docket No.: 381200.00102 CLAIMS What is claimed is:

1. A nanoparticle comprising: a membrane separating an interior space of the nanoparticle from an outer space, wherein the membrane comprises lipid nanoparticles, nanostructured lipids, solid lipids, and / or polymers; and a probe encapsulated within the interior space wherein the probe is selected from the group consisting of magnetic nanoparticles, lipids, nanostructured lipid carriers, solid lipid carriers, polymers, aptamers, proteins, or combinations thereof.

2. The nanoparticle of claim 1, wherein the membrane comprises a cationic lipid, an anionic lipid, an ionizable lipid, and / or a polymer; wherein the cationic lipid is selected from the group consisting of (1,2-dioleoyl-3- trimethylammonium-propane) (DOTAP), N-(1-(2,3-dioleyloxy)propyl)-N,N,N- trimethylammonium chloride (DOTMA), didodecyldimethylammonium bromide (DDAB), DC-cholesterol, 16:0 1,2-dipalmitoyl-3-dimethylammonium-propane (16:0 DAP), or a combination thereof; wherein the anionic lipids selected from the group consisting of glycerophospholipids, lysophospholipids, phosphatidic acid, sterols, or combinations thereof; wherein the ionizable lipid is selected from the group consisting of SM-102, ALC-0315, C12-200, cKK-E12, and DLiN-MC3-DMA, oleic acid, or combinations thereof; and / or wherein the polymer is selected from the group consisting of polylactide (PLA), polylactide-co-glycolide (PLGA), poly ε-caprolactone (PCL), or combinations thereof.

3. The nanoparticle of any one of the preceding claims, wherein the membrane has a diameter ranging from about 50 nm to about 1000 nm.

4. The nanoparticle of any one of the preceding claims, wherein the probe has a diameter ranging from about 5 nm to about 800 nm.Docket No.: 381200.00102 5. The nanoparticle of any one of the preceding claims, wherein the membrane has at least one chemical modification comprising pegylation, methylation, thiolation, or a combination thereof.

6. The nanoparticle of any one of the preceding claims, wherein the morphology of the membrane comprises spherical, pyramidal, cubical, pentagonal, or hexagonal geometries.

7. The nanoparticle of any one of the preceding claims, wherein the nanoparticle is localized to a specific region within the cell comprising the nucleus, cytosol, Golgi apparatus, or membrane.

8. The nanoparticle of any one of the preceding claims, wherein the probe comprises a barcode selected from the group consisting of morphological features, dyes, colorimetric agents, and nucleic acid molecules comprising a polynucleotide encoding an aptamer, or an antibody.

9. The nanoparticle of any one of the preceding claims, wherein the probe is dissolvable by an enzyme selected from lipases, amylases, hydrolases, laccases, and ureases.

10. The nanoparticle of any one of the preceding claims, wherein the probe is dissolvable by an organic solvent selected from toluene, xylene, and cyclohexane.

11. The nanoparticle of any one of the preceding claims, wherein the probe is further encapsulated with an organic and / or inorganic biocompatible coating; wherein the organic coating comprises lipids, zwitterionic polymers, and surfactants; wherein the inorganic coating comprises silica; and / or wherein the organic and / or inorganic biocompatible coating is either covalently or electrostatically attached to the probe.

12. The nanoparticle of any one of the preceding claims, wherein the probe binds to a protein and / or a drug of interest.Docket No.: 381200.00102 13. The nanoparticle of any one of the preceding claims, wherein the probe comprises a cysteine rich motif on its binding moiety rendering the moiety pH sensitive.

14. The nanoparticle of claim 8, wherein the dye is a Förster resonance energy transfer (FRET) dye comprising an organic fluorophore selected from cyanine, xanthene, naphthalene, and coumarin dyes.

15. The nanoparticle of claim 11, wherein the zwitterionic polymer comprises polydopamine, betaine, phosphorylcholine, or a combination thereof.

16. The nanoparticle of claim 11, wherein the surfactants comprise a fatty acid ester of glycerol or ethoxylated triglyceride.

17. The nanoparticle of claim 14, wherein the nanoparticle comprises two or more probes that are encapsulated within the membrane, and wherein the two or more probes are capable of binding to a protein of interest.

18. The nanoparticle of claim 17, wherein the two or more probes comprise: a first probe comprising a FRET acceptor, and a second probe comprising a FRET donor, wherein the second probe performs an energy transfer to the first probe when the first and second probes bind to the same protein of interest to enable detection by fluorescent imaging.

19. The nanoparticle of claim 18, wherein the nanoparticle is encapsulated within a microemulsion using a microfluidic device.

20. The nanoparticle of claim 19, wherein the microemulsion containing nanoparticles is introduced to a microemulsion containing cells to perform high throughput screening.

21. The nanoparticle of claim 19, wherein the probe comprises a nucleic acid molecule comprising a polynucleotide encoding an aptamer to detect a post-translational modification.Docket No.: 381200.00102 22. A method of manufacturing the nanoparticle of any one of the preceding claims, comprising impingement jet mixing, microfluidic mixing, homogenization, and / or sonication.

23. The method of claim 22, comprising: heating a lipid phase that comprises a solid lipid and a liquid lipid to a temperature above the melting point of the solid lipid, and heating the aqueous phase comprising a surfactant to 90°C to95°C; adding the probe to either the lipid phase or aqueous phase and mixing the aqueous solution to the lipid phase; emulsifying the lipid phase and the aqueous phase to form an emulsion using a high speed homogenizer with a mixing speed ranging from 10,000 rpm to 20,000 rpm for a period of time; ultrasonicating the emulsion using a probe sonicator using about 80% amplitude for about 10 minutes; and cooling the emulsion to about room temperature.

24. A method of uniquely labelling a probe with an oligo barcode and associating the probe with a physical characteristic comprising: imaging the probe located in a compartment such that the probe has a distinguishable physical characteristic; ligating the probe with the oligo barcode; associating the probe with the oligo barcode and cataloging the barcoded probe; pooling the probe of each compartment together and splitting the probes into separate compartments; imaging a new set of compartmentalized probes and associating the new set of compartmentalized probes with an oligo barcode; ligating a new set of oligo barcodes to the probes; and repeating the process until all the probes have a unique oligo barcode.

25. A method of imaging and cataloging the nanoparticle of claim 8, comprising: collecting a z-stack of 3-dimensional images of cells transfected with the nanoparticle, wherein the z-stack is deconvoluted; segmenting individual nanoparticles in each channel using a neural network model based on the StarDist framework;Docket No.: 381200.00102 converting the segmented nanoparticles into a 3-dimensional mesh comprising spatial, spectral, and morphological features; passing extracted features through a classification pipeline invariant to rotation and minor scale shifts, wherein the classification pipeline associates each nanoparticle with a cataloged nanoparticle based on barcode feature similarity; assigning unique identifiers to the 3-dimensional mesh and corresponding barcode and storing the unique identifiers in a database; converting the 3-dimensional mesh into an embedding vector and storing the embedding vector in the database; comparing the embedding vector of the nanoparticle to the cataloged nanoparticle by computing a cosine distance; and matching the 3-dimensional mesh of the nanoparticle to the 3-dimensional mesh of the cataloged nanoparticle to determine morphological similarity 26. A method of designing an aptamer sequence from a protein structure, comprising: identifying a homologous sequence to the protein structure using a sequence similarity search; converting the protein sequence into an embedding vector; comparing the embedding vector of the protein structure to an embedding vector of a protein with known RNA-binding affinity based on a cosine distance; querying a database storing protein-RNA interactions to identify a validated RNA sequence that binds to the protein with known RNA-binding affinity; converting the validated RNA sequence into an embedding vector and comparing the validated RNA sequence embedding vector to an embedding vector of RNA with known protein binding interactions based on a cosine distance; iteratively mutating the embedding vector of the validated RNA sequence using an evolutionary algorithm; and evaluating a protein sequence from a protein-RNA pair and an RNA sequence from a protein-RNA pair to determine the likelihood of a binding interaction.

27. A method of designing an aptamer sequence from a starting RNA sequence, comprising: identifying an RNA sequence with significant sequence similarity to the starting RNA sequence by performing a nucleotide sequence search;Docket No.: 381200.00102 converting the RNA sequence with significant sequence similarity into an embedding vector; comparing the embedding vector of the RNA sequence with significant sequence similarity to an embedding vector of an RNA sequence with known RNA-binding activity based on a cosine distance; querying a database storing protein-RNA interactions to identify a validated protein sequence that binds to an RNA sequence with known RNA-binding activity converting the validated protein sequence into an embedding vector and comparing the embedding vector of the validated protein sequence to an embedding vector of a protein sequence with known RNA-binding interactions based on a cosine distance; evaluating a protein sequence from a protein-RNA pair and an RNA sequence from a protein-RNA pair to determine the likelihood of a binding interaction; and querying a database to identify a subcellular location, a gene, and a disease associated with the protein sequence from the protein-RNA pair.

28. The method of claim 26 and 27, wherein the protein sequence from the protein-RNA pair is evaluated to determine the likelihood of a binding interaction, comprising: training a feed forward neural network classifier on a positive and a negative example of an RNA-protein interaction; scaling and normalizing an RNA embedding vector, a protein embedding vector, and a concatenated RNA-protein embedding vector; inputting the scaled and normalized RNA embedding vector, the scaled and normalized protein embedding vector, and the scaled and normalized concatenated RNA-protein embedding vector into the feed forward neural network classifier; and assigning a prediction score and / or a label to the protein sequence from the protein-RNA pair and the RNA sequence from the protein-RNA pair indicating their interaction potential using the feed forward neural network classifier.

29. A method of measuring a concentration of an intracellular protein, comprising: introducing the nanoparticle containing the first probe of claim 18 into a cell to allow the first probe to bind to the intracellular protein; isolating the first probe bound to the intracellular protein using a magnetic field;Docket No.: 381200.00102 introducing the nanoparticle containing the second probe into the cell to allow the second probe to bind to the intracellular protein on the first probe; and determining the quantity and identity of the second probe by the barcode associated therewith by an image analysis.

30. The method of claim 29, comprising degrading the second probe and introducing the nanoparticle containing a third probe for binding the intracellular protein to evaluate a proteoform and determine the presence of the proteoform.

31. The method of claim 29 and 30, comprising using the fluorescent imaging to measure the concentration of the intracellular protein.

32. The method of claim 29, comprising dissolving the second probe having the aptamer with oligo barcode and eluting the aptamer; and sequencing the eluted aptamer with oligo barcode to determine identity, spatial location and / or abundance of the intracellular protein.

33. A method of isolating tethered cells, comprising: adding a probe comprising (a) a magnetic nanoparticle and (b) a protein or an aptamer that is identifiable by the morphology, size, shape and / or color of the probe; tethering the probe to a T cell receptor-complex on the surface of a T cell; isolating the T cell using a magnetic field; identifying a barcode on the probe; and dissolving the probe to selectively retrieve the T cell.

Images