A system and method for determining the protein concentration of unknown protein samples based on automated multi-wavelength calibration.

Abstract

A UV-based imaging method for determining protein concentrations in unknown protein samples based on automated multi-wavelength calibration is disclosed. In various embodiments, a processor receives each of a set of standard wavelength data, such as data recorded by a detector, and a set of unknown wavelength data. Each of the set of standard wavelength data and the set of unknown wavelength data defines a series of absorbance-wavelength value pairs across a first wavelength range selected from a series of single-wavelength beams from the UV spectrum. The processor generates a multi-wavelength calibration model based on each of the first series of first absorbance-wavelength value pairs from the set of standard wavelength data. The processor implements the multi-wavelength calibration model to determine multiple protein concentration values ​​for each unknown protein sample in a given unknown protein sample.

Description

[0001] Cross-reference to related applications

[0002] This application claims priority to U.S. Application No. 62 / 883,320, filed August 6, 2019, entitled “Systems and methods for determining protein concentrations of unknown protein samples based on automated multi-wavelength calibration”, which is incorporated herein by reference in its entirety. Technical Field

[0003] This disclosure generally relates to ultraviolet (UV)-based imaging systems and methods related to the automation and engineering of therapeutic agents, and more specifically, to UV-based imaging systems and methods for determining the protein concentration of unknown protein samples based on automated multi-wavelength calibration. Background Technology

[0004] In the preparation and development of therapeutic pharmaceutical products, process monitoring and product quality attribute control are routinely performed. For example, techniques related to process monitoring and product quality attribute control are commonly used to ensure the efficacy of such therapeutic pharmaceutical products.

[0005] Current instruments (including those with fixed and variable path lengths, such as mechanical instruments) are designed around the application and are limited by their reliance on the inherently linear Beer-Lambert law. Specifically, because the linear range of absorbance is finite, such instruments may fail to accurately measure sample concentrations when absorbance exceeds this linear range.

[0006] The following formula illustrates the Beer-Lambert Law:

[0007] A=ε*C*l

[0008] In the above formula, A is the absorbance value, ε is the extinction coefficient, C is the concentration value, and l is the path length of the light beam through the sample. The absorbance value (A) is typically measured as the response factor of the light beam through the sample. The extinction coefficient (ε) measures the intensity of light absorbed by a substance for a given wavelength. Regarding therapeutic agent measurements, the extinction coefficient (ε) is generally an inherent property of proteins and does not change.

[0009] A concentration value (C) refers to the concentration of a given sample (e.g., a sample of a therapeutic product or other protein-based product). More generally, concentration is the abundance of a component divided by the total volume of the mixture. Concentration can refer to several different types, including mass concentration, molar concentration, number concentration, and / or volume concentration. Concentration can generally refer to any type of chemical mixture, but is most commonly used for solute or solvent solutions.

[0010] As mentioned above, the Beer-Lambert law formula assumes linearity, limiting current protein measurement instruments and techniques to linear functions of sample concentration, fixed extinction coefficient, and static optical path length. Applying the Beer-Lambert formula may fail to accurately measure the concentration-to-absorbance ratio of nonlinear samples, potentially leading to biases, false positives, and / or false negatives in process monitoring and product quality attribute control.

[0011] Due to the limitations of the Beer-Lambert system, preparing samples for measurement and controlling product quality attributes using a UV spectrometer with a fixed path length is typically both time-consuming and error-prone. For example, measuring product quality attributes may require continuous dilutions until the sample concentration falls within the dynamic range. This dilution can be prone to error, especially when measuring values ​​for several product samples, as different sample dilution factors may be required throughout the process, and concentration calculations using dilution factors can be error-prone. The dilution process applied to several product samples, and the errors typically introduced by this process, often result in discrepancies in the measurements of related product quality attributes. Such erroneous measurements can lead to an unacceptable number of false positives and false negatives during process monitoring and product quality attribute control.

[0012] Variable path length instruments have been developed to allow for the measurement of UV absorbance at variable path lengths (as opposed to traditional fixed path length UV absorbance measurements). In such instruments, the path length is defined as the distance between the tip of the optical fiber (fibrette) and the bottom of the sample container. The path length is dynamically controlled by moving the fiber up and down within the sample using integrated hardware. Variable path length instruments address the issue of directly measurable concentration ranges. In fixed path length instruments, the direct measurement range is narrow, thus always requiring sample dilution. However, with variable path length instruments, high protein concentrations can be directly measured without any dilution, as these instruments can scan the variable path length. However, the acquisition and maintenance of such instruments are typically expensive, and they are also inherently susceptible to measurement errors due to the mechanical movement and manipulation of the optical fiber. For example, a lack of routine calibration or maintenance of the fiber and / or slight variations in the mechanical components of the variable path length instrument often result in errors, where measurement results are offset due to the degree of mechanical failure, miscalibration, or mechanical variation. These errors typically lead to variations in sample measurements and make it impossible to obtain repeatable results and adequately calibrate within acceptable limits, including for purposes such as product measurement and quality control.

[0013] For the reasons mentioned above, there is a need for UV-based imaging systems and methods to determine the protein concentration of unknown protein samples based on automated multi-wavelength calibration. Summary of the Invention

[0014] This disclosure generally relates to the automation and testing of therapeutic agents, including the development, processing, monitoring, and / or product quality attribute control associated with such therapeutic agents. In various embodiments, UV-based imaging systems and methods are disclosed for determining the protein concentration of unknown protein samples based on automated multi-wavelength calibration. The determined protein concentration may be the concentration of an unknown sample of a therapeutic agent or product. Such protein concentration determination can be used for process monitoring and product quality attribute control, for example, during the manufacture of therapeutic agents or other protein-based products. Such protein concentration determination can also be used for process monitoring and product quality attribute control related to product quality control, and / or to ensure compliance with regulations or other efficacy of such products.

[0015] Therefore, as described herein, in various embodiments, a UV-based imaging system can be configured to determine the protein concentration of an unknown protein sample based on automatic multi-wavelength calibration. A UV-based imaging system may include a light source configured to project multi-wavelength beams. Additionally, a UV-based imaging system may include a monochromator configured to receive the multi-wavelength beams and project a series of single-wavelength beams of the UV spectrum based on the multi-wavelength beams.

[0016] UV-based imaging systems may further include a sample holder operable for receiving protein samples. In various embodiments, the protein sample may be (1) a standard protein sample selected from a set of standard protein samples, or (2) an unknown protein sample selected from a set of unknown protein samples. As described herein, the unknown protein sample has an unknown protein concentration. The sample holder is typically positioned to allow the series of single-wavelength beams to pass through at least a portion of the protein sample, regardless of whether the protein sample is a standard protein type or an unknown protein type.

[0017] UV-based imaging systems may further include detectors operable to detect the series of single-wavelength beams passing through at least a portion of a protein sample.

[0018] A UV-based imaging system may further include a memory storing program instructions and a processor communicatively coupled to the memory. The processor may be configured to execute the program instructions to receive a set of standard wavelength data, such as data recorded by a detector for each standard protein sample in the set of standard protein samples. For each standard protein sample, the set of standard wavelength data may include a first series of first absorbance-wavelength value pairs spanning a first wavelength range selected from the series of single-wavelength beams. Additionally, each first absorbance-wavelength value pair may include a UV absorbance value and a wavelength value.

[0019] In addition, the processor of the UV-based imaging system can be further configured to execute program instructions to generate a multi-wavelength calibration model based on each of the first series of first absorbance-wavelength value pairs of the set of standard wavelength data.

[0020] The processor of the UV-based imaging system can be further configured to execute program instructions to receive a set of unknown wavelength data, such as data recorded by the detector for each unknown protein sample in the set of unknown protein samples. For each unknown protein sample, the set of unknown wavelength data may include a second series of second absorbance-wavelength value pairs spanning a second wavelength range selected from the series of single-wavelength beams. Each second absorbance-wavelength value pair may include a UV absorbance value and a wavelength value.

[0021] Furthermore, the processor of the UV-based imaging system can be further configured to execute program instructions to calibrate the model using a multi-wavelength method and, for each unknown protein sample in the set of unknown protein samples, determine multiple protein concentration values. Each of these multiple protein concentration values ​​can correspond to a second wavelength range selected from the series of single-wavelength beams.

[0022] In another embodiment described herein, a UV-based imaging method is described for determining the protein concentration of an unknown protein sample based on automated multi-wavelength calibration. The UV-based imaging method includes receiving at a processor a set of standard wavelength data, such as data recorded by a detector for each of a set of standard protein samples. For each standard protein sample, the set of standard wavelength data may include a first series of first absorbance-wavelength pairs across a first wavelength range selected from a series of single-wavelength beams from the UV spectrum. Each first absorbance-wavelength pair may include a UV absorbance value and a wavelength value.

[0023] The UV-based imaging method may further include using a processor to generate a multi-wavelength calibration model based on each of the first series of first absorbance-wavelength value pairs of the set of standard wavelength data.

[0024] UV-based imaging methods may further include receiving at a processor a set of unknown wavelength data, such as data recorded by a detector for each of a set of unknown protein samples. For each unknown protein sample, the set of unknown wavelength data may include a second series of second absorbance-wavelength value pairs spanning a second wavelength range selected from the series of single-wavelength beams. Each second absorbance-wavelength value pair may include a UV absorbance value and a wavelength value.

[0025] UV-based imaging methods can further include determining multiple protein concentration values ​​for each unknown protein sample in the set of unknown protein samples using a multi-wavelength calibration model, such as that implemented on a processor. Each of these multiple protein concentration values ​​can correspond to a second wavelength range selected from the series of single-wavelength beams.

[0026] In another embodiment described herein, a tangible, non-transitory computer-readable medium is described storing instructions for determining the protein concentration of an unknown protein sample based on automated multi-wavelength calibration. When executed by one or more processors of a computing device, the computing device includes one or more memories, and these instructions cause the computing device to receive at the processor a set of standard wavelength data, such as data recorded by a detector for each of a set of standard protein samples. For each standard protein sample, the set of standard wavelength data may include a first series of first absorbance-wavelength value pairs across a first wavelength range selected from a series of single-wavelength beams from the UV spectrum. Each first absorbance-wavelength value pair may include a UV absorbance value and a wavelength value.

[0027] In addition, when executed by one or more processors, these instructions can further enable the computing device to generate a multi-wavelength calibration model based on each of the first series of first absorbance-wavelength value pairs of the set of standard wavelength data.

[0028] Furthermore, when executed by one or more processors, these instructions can further cause the computing device to receive at the processor a set of unknown wavelength data, such as data recorded by the detector for each of a set of unknown protein samples. Additionally, for each unknown protein sample, the set of unknown wavelength data may include a second series of second absorbance-wavelength value pairs spanning a second wavelength range selected from the series of single-wavelength beams. Each second absorbance-wavelength value pair may include a UV absorbance value and a wavelength value.

[0029] Furthermore, when executed by one or more processors, these instructions can cause the computing device to determine multiple protein concentration values ​​for each unknown protein sample in the set of unknown protein samples using a multi-wavelength calibration model, such as that implemented on the processor. Each of these multiple protein concentration values ​​can correspond to a second wavelength range selected from the series of single-wavelength beams.

[0030] In another embodiment described herein, an imaging apparatus for determining the protein concentration of an unknown protein sample based on automated multi-wavelength calibration is described. The imaging apparatus includes means for receiving a set of standard wavelength data for each of a set of standard protein samples. For each standard protein sample, the set of standard wavelength data includes a first series of first absorbance-wavelength pairs spanning a first wavelength range selected from a series of single-wavelength beams from the UV spectrum. Each first absorbance-wavelength pair includes a UV absorbance value and a wavelength value.

[0031] The imaging apparatus further includes means for generating a multi-wavelength calibration model for each of the first series of first absorbance-wavelength value pairs based on the set of standard wavelength data.

[0032] The imaging apparatus further includes means for receiving a set of unknown wavelength data for each of a set of unknown protein samples. For each unknown protein sample, the set of unknown wavelength data includes a second series of second absorbance-wavelength pairs spanning a second wavelength range selected from the series of single-wavelength beams. Each second absorbance-wavelength pair includes a UV absorbance value and a wavelength value.

[0033] The imaging apparatus further includes means for determining multiple protein concentration values ​​for each unknown protein sample in the set of unknown protein samples using a multi-wavelength calibration model. Each of these multiple protein concentration values ​​corresponds to a second wavelength range selected from the series of single-wavelength beams.

[0034] The UV-based imaging systems and methods disclosed herein overcome the limitations of currently known instruments. For example, as described herein, the UV-based imaging systems and methods of this disclosure describe automated workflows for process monitoring and product quality attribute control used in the preparation and development of pharmaceutical therapeutic products. Furthermore, the automated workflow enhances and improves upon currently known methods by providing a full scan of the UV spectrum rather than being limited to a single or dual wavelength. For example, instead of relying on a single UV wavelength (e.g., 280 nm) to measure sample protein concentration as is typical in conventional instruments, the UV-based imaging systems and methods of this disclosure measure protein concentration, allowing scanning over a wide range, including, for example, the UV range of 280–315 nm.

[0035] Furthermore, the UV-based imaging system and method of this disclosure overcomes the limitations of both fixed-path-length and multi-path-length UV spectrometers, including common problems related to the dynamic range of measured wavelengths and absorbance values, as well as mechanical challenges introduced by closer instruments. For example, current instruments rely on and are limited by the application of Beer-Lambert's law. To overcome the linear limitation of Beer-Lambert's law, the UV-based imaging system and method of this disclosure are configured to detect the absorbance value for each wavelength measured for a given sample, and then dynamically calibrate the absorbance response factor based on changes in the relevant extinction coefficient. As a result, applying the UV-based imaging system and method of this disclosure can improve fixed-path-length UV spectrometers and their associated data output, providing accurate measurements of protein concentration over a wide range, including, for example, from 0.1 to 126 mg per milliliter (mL).

[0036] Additionally, in various embodiments, the UV-based imaging systems and methods of this disclosure typically generate overlap of wavelength values ​​over a wide range of protein concentration measurements. In such embodiments, enhanced predictions (e.g., enhanced accuracy of concentration measurements) can be generated by applying the average, median, derivation, derivative, or otherwise compiled from the wavelength range to produce enhanced multi-sample measurements. For example, in some embodiments, a signal averaging can be applied to the overlapping concentration measurements, resulting in enhanced accuracy. For example, "enhanced" predictions may have higher accuracy (e.g., tighter confidence intervals or smaller standard errors) compared to single-wavelength-based reference predictions.

[0037] Furthermore, unlike variable UV path-length instruments that use mechanical light guides to scan with different path lengths, thereby extending the linear protein concentration range, the UV-based imaging system and method of this invention have no mechanical error in their natural extinction coefficient tuning and are applicable to any conventional UV spectrometer.

[0038] Furthermore, the UV-based imaging systems and methods disclosed herein provide improvements to computer functionality and / or other technologies, at least because they enhance the accuracy and precision of computing devices (such as spectrophotometers) used to measure information about protein-related products. That is, this disclosure describes improvements to the functionality of the computing device itself (e.g., a spectrophotometer or a device or system for measuring or monitoring protein-related products) because the application of the UV-based imaging techniques described herein increases the accuracy and precision of such instruments, devices, or systems. This leads not only to increased reliability of the underlying system but also to increased reliability of the system's end use, such as monitoring and / or measuring protein concentrations in protein-related products (such as therapeutic agents), as described herein. This is an improvement over existing technologies, at least because, as described herein, existing instruments, such as fixed-path-length instruments and mechanical, variable-based instruments, are limited, error-prone, and / or expensive to use. In contrast, the UV-based imaging systems and methods of this disclosure are compatible with any UV spectrometer with full-scan capability. This allows for the deployment of UV-based imaging systems and methods throughout analytical monitoring and / or testing networks. For similar reasons, the disclosed systems and methods extend the analytical capabilities of existing systems and instruments without requiring capital investment or large equipment footprint.

[0039] Similarly, the UV-based imaging systems and methods disclosed herein relate to improvements in other technologies or technical fields within the automation and engineering of therapeutic agents, at least because UV-based imaging systems and methods provide a fully automated workflow, enabling real-time measurement and monitoring with zero offline sample and data processing. Using this automated workflow, sample dilution and manual data processing can be eliminated or reduced, which reduces human error and improves the reliability of the entire process.

[0040] As described herein, in at least some embodiments, UV-based imaging systems and methods include the use and application of, or the use of, specific machines such as spectrophotometers and / or related components used by spectrophotometers or other such UV imaging devices.

[0041] Additionally, the UV-based imaging systems and methods disclosed herein include the transformation or restoration of a specific article to different states or states, for example, transforming or restoring a multi-wavelength beam to a multi-wavelength calibration model for predicting multiple protein concentration values ​​that can be used to monitor or measure the protein concentration of protein-related products. This monitoring and measurement can be used during the manufacturing process of protein-related products and / or can be used to determine whether protein-related products comply with known specifications or regulations related to protein-related products, such as those used in the manufacture of therapeutic products. For example, based on the measured protein concentration, the protein concentration of a protein-related product can be increased or decreased to fall within a specific range of protein concentration. For example, a batch of protein-related products can be accepted or rejected depending on whether the measured protein concentration falls within or outside a specific range.

[0042] Furthermore, the UV-based imaging systems and methods disclosed herein include specific features beyond those well-known, conventional, and routine activities in the art, or add unconventional steps that limit the claims to specific useful applications, such as predictive, augmentation, and measurement applications lacking in current instruments. For example, as described herein, conventional instruments are limited by reliance on Beer-Lambert's law, mechanical instruments (e.g., guide lights), or manipulation, which are overcome by the UV-based imaging systems and methods of this disclosure.

[0043] The advantages will become more apparent to those skilled in the art from the following description of the preferred embodiments shown and described by way of illustration. As will be appreciated, embodiments of the invention may have other and different embodiments, and their details may be modified in various aspects. Therefore, the drawings and description are to be regarded in an illustrative rather than restrictive manner. Attached Figure Description

[0044] The accompanying drawings described below depict various aspects of the systems and methods disclosed therein. It should be understood that each drawing depicts an embodiment of a specific aspect of the disclosed systems and methods, and each of the drawings is intended to correspond with its possible embodiments. Furthermore, wherever possible, the following description refers to reference numerals included in the following drawings, wherein features depicted in the plurality of drawings are indicated by consistent reference numerals.

[0045] The arrangement currently under discussion is illustrated in the accompanying drawings; however, it should be understood that embodiments of the invention are not limited to the precise arrangement and tools shown, wherein:

[0046] Figure 1 An exemplary UV-based imaging system, according to various embodiments disclosed herein, is shown for determining the protein concentration of an unknown protein sample based on automated multi-wavelength calibration.

[0047] Figure 2A and Figure 2B A flowchart of an exemplary UV-based imaging method according to various embodiments disclosed herein is shown, which is used to determine the protein concentration of an unknown protein sample based on automatic multi-wavelength calibration.

[0048] Figure 3A and Figure 3B A list of example computer programs, including pseudocode, is shown for implementation according to various embodiments disclosed herein. Figure 2A and Figure 2B A UV-based imaging method.

[0049] Figure 4A A graph representing a set of standard protein samples across a single wavelength beam is shown, according to various embodiments disclosed herein.

[0050] Figure 4B The diagram illustrates, according to various embodiments disclosed herein, a series of first absorbance-to-wavelength value pairs representing a first wavelength range selected from the UV spectrum.

[0051] Figure 5A A diagram illustrating a multi-wavelength calibration model consisting of a single-wavelength calibration model, according to various embodiments disclosed herein, is shown.

[0052] Figure 5B Various embodiments disclosed herein are shown, indicating Figure 5A Plots of slope and y-intercept for various single-wavelength calibration models.

[0053] Figure 6 The diagram illustrates various series of second absorbance versus wavelength pairs across a second wavelength range selected from the UV spectrum, according to the various embodiments disclosed herein.

[0054] Figure 7 Various embodiments disclosed herein are shown, illustrating the use of Figure 5A A graph showing protein concentration values ​​determined by a multi-wavelength calibration model.

[0055] Figure 8 Various embodiments disclosed herein are illustrated, depicting applications. Figure 5A A flowchart for using a multi-wavelength calibration model to monitor or measure the protein concentration of protein-related products.

[0056] The accompanying drawings depict preferred embodiments for illustrative purposes only. Alternative embodiments of the systems and methods shown herein may be employed without departing from the principles of the invention described herein. Detailed Implementation

[0057] Figure 1 An example UV-based imaging system 100 according to various embodiments disclosed herein is illustrated, which is used to determine the protein concentration of an unknown protein sample based on automated multi-wavelength calibration. In various embodiments, system 100 may include a UV imaging device 101, which may be a UV spectrophotometer or other such imaging device. UV imaging device 101 includes a light source 102 configured to project a multi-wavelength beam 104 onto a monochromator 106. In various embodiments, monochromator 106 may be a prism, a diffraction device, or any other device implementing a diffraction grating to split the multi-wavelength beam into a single-wavelength beam. Figure 1 In one embodiment, the monochromator 106 is configured to receive a multi-wavelength beam 104 and use the multi-wavelength beam 104 as a light source to project or otherwise provide a series of single-wavelength beams 108 of the UV spectrum.

[0058] In various embodiments, the series of single-wavelength beams 108 may include single-wavelength beams in the UV spectrum from 200 nanometers (nm) to 400 nanometers (nm). Alternatively, the single-wavelength beams 108 may include a large number of wavelengths (e.g., multiple wavelengths extended over a small span of the UV spectrum), wherein each interval of the single-wavelength beams 108 is a wavelength in the UV spectrum. For example, in some embodiments, each single-wavelength beam 108 in the UV spectrum may be separated by a single nanometer interval. However, different intervals and interval sizes are considered herein, such that a single nanometer interval is only one example embodiment.

[0059] The UV imaging apparatus 101 may further include a sample holder 110 operable to receive, hold, or position a protein sample relative to the UV imaging apparatus 101. The sample holder 110 may be configured to hold at least a portion of a flow cell, colorimetric tube, or other container for encapsulating the protein sample. The sample holder 110 is typically positioned within or as part of the UV imaging apparatus 101 to allow a series of single-wavelength beams of light to pass through at least a portion of the protein sample, regardless of whether the protein sample is a standard protein type or an unknown protein type. The sample holder may be formed of plastic, glass, fused silica, or other such materials that allow light transmission through the material and the protein sample.

[0060] In various embodiments, according to the various embodiments herein, the protein sample may be a standard protein sample (e.g., standard protein sample 140s) or an unknown protein sample (e.g., unknown protein sample 140u). As used herein, the term "unknown protein sample" refers to a protein sample having an unknown protein concentration. The protein sample may be liquid, semi-liquid, solid, or other measurable state. The protein in a given protein sample may be a proprietary protein of a therapeutic drug product.

[0061] As described in the various embodiments herein, the unknown protein sample has an unknown protein concentration, and the standard protein sample has a known protein concentration. Note that the unknown protein sample contains at least one protein of unknown concentration; of course, more than one protein may be present. The sample may include other protein-free components, such as buffers, salts, and other organic and inorganic molecules. The standard protein sample may be a reference standard or other known, pre-measured, or predetermined sample with a known, pre-measured, or predetermined dilution, concentration, or other known protein-based percentage or allocation.

[0062] In various embodiments, a protein sample may be (1) a standard protein sample selected from a set of standard protein samples (e.g., standard protein sample 140s), or (2) an unknown protein sample selected from a set of unknown protein samples (e.g., unknown protein sample 140u). As used herein, a set of protein samples (e.g., standard or unknown) includes one or more protein samples. For example, in one embodiment, a set of standard protein samples (e.g., standard protein sample 140s) may include a plurality of standard protein samples (e.g., standard protein sample 140s) having their respective known concentrations, and a set of unknown protein samples (e.g., unknown protein sample 140u) may include a single protein sample having an unknown protein concentration.

[0063] UV imaging apparatus 101 may further include detector 114 operable to detect a series of single-wavelength beams 112 passing through at least a portion of a protein sample, such as that held by sample holder 110. The single-wavelength beams 112 may include light of the same or different number, type, intensity, and / or size as the single-wavelength beams 108. Similarities or differences between the single-wavelength beams 108 and the multi-wavelength beams 112 are generally caused by the protein sample (and its associated protein concentration) located within sample holder 110, where such similarities or differences result in the generation and / or output 116 of wavelength data 118 (e.g., for standard or unknown samples) specific to that protein sample (based on its protein concentration), as recorded by detector 114.

[0064] Detector 114 may be or may include a photomultiplier tube, photodiode, photodiode array, charge-coupled device (CCD), or other photosensitive device or array configured to collect and / or record respective values ​​of light at various wavelengths. For example, in a CCD-based detector, a single-wavelength beam 112 may be focused onto a pixel array of detector 114, wherein the pixels of the array may measure the intensity of each wavelength (e.g., each color in the case of the visible region) and record it as one or more values. The CCD-based detector may then output 116 its measurements as wavelength data 118, which may together represent the spectrum or one or more values ​​of the intensity of each wavelength of light. More generally, wavelength data 118 measures the absorbance value of the protein sample being measured (e.g., the absorbance of the protein concentration of the protein sample currently in sample holder 110). That is, at least some of the single-wavelength beam 108 passing through sample holder 110 is at least partially absorbed, or varies according to the concentration of the protein sample, such that the single-wavelength beam 112 also changes similarly. For example, if a protein sample located in sample holder 110 absorbs light of a specific wavelength, that wavelength of light will be absent or partially absent in single-wavelength beam 112. Therefore, detector 114 typically measures the difference between single-wavelength beam 108 and single-wavelength beam 112 to determine absorbance or absorbance values ​​as one or more response factors.

[0065] In at least some embodiments, each of the light source 102, monochromator 106, sample holder 110, and detector 118 of the UV imaging device 101 can form at least a part of a spectrophotometer apparatus. For example, in at least one embodiment, the spectrophotometer can be a THERMO FISHER SCIENTIFIC TM EVOLUTION TM Spectrophotometer or other such imaging devices.

[0066] Detector 114 can be communicatively coupled to various components 121 to 128 of the UV imaging apparatus 101 via bus 123 (i.e., electronic communication bus). For example, detector 114 can be communicatively coupled to processor 124. Processor 124 can be a microprocessor or a central processing unit (CPU), such as based on... based on Or other such microprocessors. Processor 124 may be responsible for controlling various components communicatively coupled via bus 123. For example, processor 124 may control the output 116 of detector 114 as wavelength data 118, which, in one embodiment, is stored in memory 121. Additionally, processor 124 may receive commands or other instructions from input / output component 126. Input / output component 126 may interface with or otherwise connect to various input / output devices, such as a keyboard, mouse, or similar components. Such components may be used to access or otherwise manipulate or retrieve wavelength data 118 (e.g., in memory 121) as the output 116 of detector 116. Processor 124 may also be communicatively connected to display 128. Display 128 may be a screen in which processor 124 presents a representation of wavelength data 118, such as two-dimensional (2D), three-dimensional (3D), or other representations, on the display screen of display 128.

[0067] Processor 124 may be further communicatively connected to transceiver 122 via bus 123. Processor 124 may be communicatively coupled to computing device 151 via computing network 130 via transceiver 122. In the illustrated embodiment, computing device 151 may transmit 134 and receive 132 data (e.g., wavelength data 118) via computer network 130, via remote processor 154 and remote transceiver 152 of computing device 151. Remote processor 154 and remote transceiver 152 may be communicatively coupled to each other via bus 153 (e.g., electronic communication bus). Computing device 151 may further include remote memory 159 communicatively coupled via bus 153 to store data such as data received via remote transceiver 152 (e.g., wavelength data 118). The computing device 151 may further include a display 158 and an input / output unit 156 communicatively coupled via a bus 153 to facilitate input / output operations, such as receiving commands via a touchscreen, keyboard, etc., and displaying data, such as wavelength data 118, on the screen of the display 158. Thus, the computing device 151 includes a remote processor 154 and a remote memory 159 that can be used to store and / or process wavelength data 110 remote from the UV imaging device 101.

[0068] exist Figure 1 In embodiments, memory 121 and / or remote memory 159 may store program instructions to cause any one or both of processors 124 and / or 154 to execute the program instructions, thereby implementing as described herein. Figure 2A and Figure 2BThe described UV-based imaging method 200. Program instructions may be program code in a programming language such as Python, Java, C#, or other programming languages. In some embodiments, the program instructions may be client-server based, where a remote processor 154, acting as a client, communicates with a processor 124, acting as a server, via a computing network 130. In such embodiments, the remote processor 154 may request data to be transferred from the UV imaging device 101 to the computing device 151, such as wavelength data 118 (e.g., stored in memory 121 or as a new output 116 of detector 114). Wavelength data 118 may be requested by the remote processor 154 via an online application programming interface (API) (such as a RESTful API), where the processor 124 implements the API to receive the request from the remote processor 154 and responds by providing wavelength data 118 via the computer network 130. In other embodiments, the UV imaging device 101 may implement a push-based interface, where newly generated and / or output wavelength data 118 is transmitted to the computing device 151 via the computer network 130. In a further embodiment, a user of the UV imaging device 101 or computing device 151 may receive or retrieve wavelength data from an external storage device (not shown) (such as a disk or thumb drive) via input / output 126 or 156.

[0069] Figure 2A and Figure 2B A flowchart of an example UV-based imaging method 200 according to various embodiments disclosed herein is shown, which is used to determine the protein concentration of an unknown protein sample (e.g., an unknown protein sample 140u) based on automated multi-wavelength calibration. The UV-based imaging method 200 overcomes the limitations of the prior art, including reliance on the Beer-Lambert law and its assumptions and limitations regarding linearity. Figure 2A and Figure 2B and Figure 3A and Figure 3B Displayed together. Figure 3A and Figure 3B A list of example computer programs, including pseudocode, is shown for implementing various embodiments disclosed herein. Figure 2A and Figure 2B UV-based imaging methods 200. (For example, regarding...) Figure 3A and Figure 3BAs shown, to overcome the linear limitation of the Beer-Lambert law, the program code (as shown in pseudo-Python code) can apply a sample filter to determine the linear absorbance range for each wavelength and dynamically calibrate the absorbance response factor based on changes in one or more extinction coefficients. As a result, the UV imaging device 101 (e.g., an embodiment including the UV imaging device 101 as a fixed path length UV spectrometer) can accurately measure one or more protein concentrations over a wide range, such as protein concentrations in the range of 0.1 to 126 mg / mL.

[0070] As by Figure 2A The depicted UV-based imaging method 200 begins (202) at block 204, wherein a processor (e.g., processor 125 or remote processor 154) receives a set of standard wavelength data (e.g., wavelength data 118) as recorded by a detector (e.g., detector 114) for each of a set of standard protein samples (e.g., standard protein samples 140s). In some embodiments, for example, the set of standard wavelength data (e.g., wavelength data 118) may be received as comma-separated value (CSV) data. In other embodiments, the set of standard wavelength data may be received in a database format, Extensible Markup Language (XML), JavaScript Object Notation (JSON), or other data or text formats. Figure 3A and Figure 3B In the embodiment of the computer program list, for example, at code segment 5, wavelength data of standard protein samples (e.g., standard protein sample 140s) are read as CSV data into memory (e.g., memory 121 or remote memory 159) by a processor (e.g., processor 124 or remote processor 154), wherein each protein standard sample is assigned its known concentration. Figure 3A and Figure 3B The list of computer programs is written in pseudo-Python, with the pandas library "pd" used to read in wavelength data. This is shown in code snippets 1 and 5. Code snippet 1 also shows an import of the additional library "sklearn," which provides a linear regression training algorithm as an example to illustrate the training of a predictive model as described herein.

[0071] For each standard protein sample, the set of standard wavelength data (e.g., wavelength data 118) may include a first series of first absorbance-wavelength pairs across a first wavelength range selected from a series of single-wavelength beams from the UV spectrum. Each first absorbance-wavelength pair may include a UV absorbance value and a wavelength value. For example, Figure 4AFigure 400 illustrates a set of standard protein samples (e.g., standard protein sample 140s) 400a to 400m on the UV spectrum of a single wavelength beam 404, according to various embodiments disclosed herein. Specifically, Figure 400 includes a UV absorbance axis 402 (y-axis), which has a range of absorbance values ​​used as a response factor measurement. Generally, for a given wavelength, the larger the absorbance value, the less light is detected (e.g., by detector 114). Additionally, Figure 400 includes a wavelength axis 404 (x-axis), which has a range of light wavelengths spaced in nanometers (although other spacings are also considered herein).

[0072] exist Figure 4A In this embodiment, standard protein samples (e.g., standard protein sample 140s) with protein standard solutions in the range of 0.1-126 mg / ml were scanned in a 200-400 nm UV spectrum (with 1 nm intervals) using a spectrophotometer. The standard protein samples were contained in quartz colorimetric tubes with a path length of 1 cm. The wavelength data of the UV spectra were saved as CSV data. Thus, Figure 400 depicts the raw UV spectral data of the protein calibration standards (e.g., standard protein samples (e.g., standard protein sample 140s)) scanned in the wavelength range of 200 nm to 400 nm.

[0073] like Figure 4A As shown, Figure 400 represents 13 standard protein samples (e.g., standard protein sample 140s) 400a to 400m. For each standard protein sample, the same protein (e.g., protein molecule) is measured by protein concentration. The protein molecule can be a molecule, such as a proprietary protein or other proteins used to develop therapeutic products or other such pharmaceutical products.

[0074] like Figure 4AAs shown, each standard protein sample (e.g., standard protein sample 400a and standard protein sample 400m) is represented by a graph of multiple absorbance versus wavelength pairs in Figure 400. For example, for standard protein sample 400m, absorbance versus wavelength pair 400m305 is the measurement of standard protein sample 400m at a wavelength of 305 and an absorbance of approximately 2.5. Therefore, absorbance versus wavelength pair 400m305 is located at the corresponding UV absorbance axis 402 (y-axis) and the corresponding wavelength axis 404 (x-axis) in Figure 400. As a further example, for standard protein sample 400a, absorbance versus wavelength pair 400a230 is the measurement of standard protein sample 400a at a wavelength of 230 and an absorbance of approximately 1.0. Therefore, the absorbance to wavelength pair 400a230 is located at the corresponding UV absorbance axis 402 (y-axis) and the corresponding wavelength axis 404 (x-axis) within Figure 400. The residual absorbance to wavelength pairs for each standard protein sample from 400a to 400m are similarly plotted within Figure 400, each representing the UV absorbance value (along UV absorbance axis 402) scanned and measured at each wavelength (along wavelength axis 404) in the wavelength spectrum from 200 nm to 400 nm.

[0075] about Figure 2A In the UV-based imaging method 200, at block 206, a processor (e.g., processor 124 or remote processor 154) determines whether to filter the set of standard wavelength data. This determination can be based on the range of wavelength data detected (e.g., by detector 114), some of which may include a wide wavelength range. If no filter is applied, the UV-based imaging method 200 proceeds to block 210.

[0076] However, at box 208, a standard wavelength data filter is applied. The standard wavelength data filter may include a set of program instructions stored in the memory of the UV imaging system 100 (e.g., in memory 121 or remote memory 154). For example, in Figure 3A and Figure 3B In the list of computer programs, pseudo-Python instructions illustrate an example of a standard wavelength data filter (code snippet 6) for use in... Figure 4A Data points were removed from the original single-wavelength light range and the original UV absorbance value range to ensure optimal data quality for training the calibration model as described in this paper. Figure 3AIn the example, code segment 6 applies a filtering criterion set as a user-defined variable (code segment 4) to allow for flexible manipulation of the filtering range. For example, as shown in code segment 4, the filtering criterion for a standard protein sample (e.g., standard protein sample 140s) is configured to select data points with UV absorbance of 0.1 to 1.5 for the wavelength range of 280 to 315 nm.

[0077] At block 208, the processor (e.g., processor 124 or remote processor 154) implements a standard wavelength data filter for a series of single-wavelength beams to select or limit a first wavelength range. Alternatively, the processor may implement a standard wavelength data filter to select or limit a first absorbance-wavelength value pair for each first series based on UV absorbance values ​​within a first filter range of UV absorbance values. For example, the application of a standard wavelength data filter (e.g., at block 208) produces... Figure 4B Figure 450 shows Figure 4A Figure 400 plots a limited or filtered range of absorbance to wavelength pairs (both wavelength and absorbance values, i.e., absorbance to wavelength pairs).

[0078] Specifically, Figure 4B Figure 450 illustrates various series of first absorbance versus wavelength pairs representing a first wavelength range (e.g., wavelength axis 454) selected from the UV spectrum, according to various embodiments disclosed herein. Figure 4B As shown, absorbance to wavelength pairs for each standard protein sample (e.g., standard protein sample 140s) from 400a to 400m are illustrated; however, as described herein, some absorbance to wavelength pairs have been filtered by a standard wavelength data filter (e.g., at box 208). Specifically, Figure 4B One embodiment is shown in which a standard wavelength data filter is applied to... Figure 4A The absorbance to wavelength pair is plotted within the single-wavelength beam range of 200 to 400 nm.

[0079] As by Figure 4B As shown, a standard wavelength data filter (e.g., at box 208) has been applied. Figure 4A The single-wavelength beam ranges from 200 to 400 nm, allowing for selection. Figure 4B The first wavelength range is 280 to 315 nm. The first wavelength range of 280 to 315 nm is... Figure 4B It is plotted on the wavelength axis 454. Therefore, Figure 4B The first wavelength range, from 280 to 315 nm, has been extended from... Figure 4A The single-wavelength beam range of 200 to 400 nm is selected.

[0080] In addition, Figure 4B In the embodiments, a standard wavelength data filter (e.g., at block 208) is applied to select a first absorbance-wavelength value pair for each first series based on UV absorbance values ​​within a first filter range (e.g., UV absorbance axis 452). The first filter range of UV absorbance values ​​from 0.1 to 1.5 is within... Figure 4B It is plotted on the UV absorbance axis 452. Therefore, Figure 4B The first filter range, with UV absorbance values ​​ranging from 0.1 to 1.5, has been extended from... Figure 4A The UV absorbance value is selected from the range of 0 to 7. Therefore, Figure 4B The filtered UV spectra of calibration standards are shown in general, and specifically, the filtered UV spectra of protein calibration standards are shown within a limited range of wavelengths from 280 to 315 nm and absorbance values ​​from 0.1 to 1.5. When compared with... Figure 4A When making comparisons, the filter values ​​(wavelength values ​​from 280 to 315 nm and absorbance values ​​from 0.1 to 1.5) typically include those absorbance-wavelength pairs across each standard protein sample (e.g., 400a and standard protein sample 400m), which are not folded relative to each other or on top of each other, as shown in Figure 400.

[0081] exist Figure 4A and Figure 4B In some examples, the standard protein sample (e.g., standard protein sample 140s) is a protein standard solution in the range of 0.1 to 126 mg / mg, such as... Figure 4B As shown. This is at code segment 2. Figure 3A and Figure 3B The list of example computer programs further illustrates that standard protein sample values ​​are set to corresponding ranges. However, this document considers a wider range of such standard protein samples (e.g., standard protein sample 140s) / protein standard solutions, for example, where the set of standard protein samples can be protein standard solutions ranging from 0.1 to 220 mg per milliliter (ml).

[0082] about Figure 2A At box 210, the UV-based imaging method 200 includes generating a multi-wavelength calibration model using a processor (e.g., processor 124 or remote processor 154). The multi-wavelength calibration model is based on this set of standard wavelength data and its various absorbance value pairs for each standard protein sample. In some embodiments, where such data is filtered (e.g., at box 208), the multi-wavelength calibration model is based on each of a first series of first absorbance-wavelength value pairs from the set of standard wavelength data.

[0083] Figure 5AFigure 500 illustrates a multi-wavelength calibration model 501 consisting of a single-wavelength calibration model (i.e., a "calibration curve") according to various embodiments disclosed herein. Figure 5A In the embodiment, the multi-wavelength calibration model 501 consists of cross-wavelength calibration models. Figure 4B The UV spectrum is composed of single-wavelength calibration models. This includes models for each wavelength from 280 nm to 315 nm. For example, Figure 500 identifies single-wavelength calibration models 501s280 (which is a single-wavelength model for wavelength 280 nm), 501s294 (which is a single-wavelength model for wavelength 294 nm), and 501s300 (which is a single-wavelength model for wavelength 300 nm). Each single-wavelength calibration model is generated to predict protein concentration values ​​(a range of protein concentration values ​​across protein concentration axis 504) for a given UV absorbance value (a range of UV absorbance values ​​across UV absorbance axis 502) for a specific wavelength (e.g., 280 nm, 294 nm, and 300 nm). Figure 5A The UV absorbance axis 502 corresponds to Figure 4B The UV absorbance axis is defined. Therefore, for a given wavelength in the UV spectrum, protein concentration can be predicted by inputting the UV absorbance value into the corresponding single-wavelength calibration model or the calibration curve of the UV spectral wavelength. As described herein, these predictions can be used to determine the unknown protein concentration in an unknown protein sample (e.g., an unknown protein sample of 140 u).

[0084] exist Figure 5A In this embodiment, a linear regression algorithm is used to determine each single-wavelength calibration model (e.g., 501s280, 501s294, and 501s300). Each such model includes a y-intercept and a slope value, which are used to predict protein concentration. The y-intercept and slope values ​​are typically determined to be constant values, where the slope value is applied to a provided UV absorbance value, and the result is added to the y-intercept value to produce the corresponding protein concentration prediction or value. Figure 5A Figure 500 illustrates various predicted or calculated protein concentrations (across protein concentration axis 504) for each of the individual single-wavelength calibration models (e.g., 501s280, 501s294, and 501s300) when different UV absorbance values ​​are provided for each of the individual single-wavelength calibration models. Thus, the entire single-wavelength calibration model comprises a multi-wavelength calibration model that can accept absorbance values ​​for each wavelength and, as for... Figure 4A and Figure 4B Determine and describe the UV absorbance value.

[0085] Figure 5B Various embodiments disclosed herein are shown, indicating Figure 5AFigure 550 shows the slope 560s and y-intercept 560y of various single-wavelength calibration models (e.g., single-wavelength calibration models 501s280, 501s294 and 501s300). Figure 5B A graph showing paired slope values ​​560s and y-intercept values ​​560y at each wavelength between 280nm and 315nm across wavelength axis 554 is presented. Figure 5B Including with Figure 5A The same range of UV absorbance values ​​(e.g., UV absorbance axis 552) and the same range of wavelength values ​​(e.g., wavelength axis 554). The paired slope 560s and y-intercept 560y correspond to... Figure 5A Single-wavelength calibration models (e.g., single-wavelength calibration models 501s280, 501s294, and 501s300). For example, as by... Figure 5B As shown, the slope value 562a corresponds to the y-intercept value 562b at a wavelength of 280 nm, which corresponds to the single-wavelength calibration model 501s280. Similarly, as from... Figure 5B As shown, the slope value 564a corresponds to the y-intercept value 564b at a wavelength of 290 nm, which corresponds to the single-wavelength calibration model 501s290. The y-axis (UV absorbance axis 552) displays the UV absorbance metric for any given prediction of the single-wavelength calibration model, given either the slope value or the y-intercept value.

[0086] Therefore, at least relative to Figure 5B Implementation examples, Figure 5A The multi-wavelength calibration model 501 can be generated and includes a series of single-wavelength calibration models for use with this series of single-wavelength beams (e.g., such as...). Figure 4A The first wavelength range selected (e.g., as described in the wavelength range of 200 nm to 400 nm) Figure 4B Each of the described wavelengths (280 nm to 315 nm) includes a slope value 560s and a y-intercept value 560y corresponding to the UV absorbance of a specific single wavelength (e.g., wavelengths of 280 nm, 294 nm, and 300 nm, respectively). By inputting the wavelength value and the UV absorbance value, a processor (e.g., processor 124 and / or remote processor 154) implementing the multi-wavelength calibration model 501 can select the corresponding single-wavelength calibration model to correctly and accurately predict protein concentration values.

[0087] Figure 3A and Figure 3B The list of example computer programs illustrates the first wavelength range (e.g., such as...). Figure 4BThe generation of single-wavelength calibration models for each of the wavelengths described (280 nm to 315 nm). For example, at code segment 8, after the standard wavelength data is filtered, a linear regression algorithm (LinearRegression.fit) is performed on the data for each standard protein sample by a processor (e.g., processor 124 or remote processor 154), which results in the generation of coefficient values ​​(i.e., slope values) and intercept values ​​(i.e., y-intercept values) for the single-wavelength calibration model at a specific wavelength. Thus, each single-wavelength calibration model (e.g., 501s280, 501s294, and 501s300) can be generated in this way. The single-wavelength calibration models can be stored in memory (e.g., memory 121 or remote memory 159) and can be accessed together as a comprehensive multi-wavelength calibration model as described herein (e.g., via a wrapper API or as a data structure via a database). Figure 3A and Figure 3B In this embodiment, each single-wavelength calibration model includes a quality or fit score for a certain threshold. For example, in code segment 3, only single-wavelength calibration models with a minimum quality or fit value of 0.99 can be included in the multi-wavelength calibration model. In this way, the multi-wavelength calibration model 501 maintains and uses a set of high-quality prediction models with accurate predictive capabilities.

[0088] exist Figure 5A and Figure 5B In the above embodiments, a single-wavelength calibration is generated via linear regression and compiled into a generated multi-wavelength calibration model 501. However, other prediction technologies or techniques may also be used. For example, machine learning models can be trained using supervised or unsupervised machine learning programs or algorithms. Machine learning programs or algorithms may employ neural networks, which may be convolutional neural networks, deep learning neural networks, or combined learning modules or programs that learn from two or more features or feature datasets in a specific region of interest. Machine learning programs or algorithms may also include natural language processing, semantic analysis, automated reasoning, regression analysis, support vector machine (SVM) analysis, decision tree analysis, random forest analysis, K-nearest neighbor analysis, naive Bayes analysis, clustering, reinforcement learning, and / or other machine learning algorithms and / or techniques. Machine learning may include identifying and recognizing patterns in existing data (such as the UV absorbance values ​​of standard protein samples (e.g., standard protein sample 140s) in the UV spectrum) to facilitate prediction of subsequent data (e.g., predicting the unknown concentration of unknown protein samples (e.g., unknown protein sample 140u) as described herein).

[0089] One or more machine learning models, such as those described above, can be created and trained based on example (e.g., “training data”) inputs or data (which may be referred to as “features” and “labels”) to make effective and reliable predictions on new inputs, such as test-level or production-level data or inputs. In supervised machine learning, a machine learning program operating on a server, computing device, or other one or more processors can be set up with example inputs (e.g., “features”) and their associated or observed outputs (e.g., “labels”) so that the machine learning program or algorithm determines or discovers rules, relationships, or other machine learning “models” that map such inputs (e.g., “features”) to outputs (e.g., labels), for example, by determining and / or assigning weights or other metrics to the model across its various feature categories. Such rules, relationships, or other models can then be provided as subsequent inputs to a model executing on a server, computing device, or other one or more processors to predict the expected output based on the discovered rules, relationships, or models.

[0090] In unsupervised machine learning, a server, computing device, or other one or more processors may need to find its own structure in unlabeled example inputs, whereby, for example, the server, computing device, or other one or more processors perform multiple training iterations to train multiple generations of models until a satisfactory model is generated, for example, a model that provides sufficient predictive accuracy when given test-level or production-level data or inputs. The disclosures in this paper may utilize one or both of such supervised or unsupervised machine learning techniques.

[0091] Therefore, the single-wavelength calibration model or the multi-wavelength calibration model itself of the multi-wavelength calibration model can be trained using wavelength data of a standard protein sample (e.g., standard protein sample 140s), wherein wavelength values ​​and UV absorbance values ​​(e.g., various absorbance-wavelength value pairs as described herein) are input as feature data, and wherein protein concentration values ​​of the standard protein sample (e.g., standard protein sample 140s) can be used as labeling data. The trained single-wavelength calibration model or the multi-wavelength calibration model of the multi-wavelength calibration model can then be used to predict the protein concentration of a sample with an unknown concentration, as described herein.

[0092] about Figure 2A The UV-based imaging method 200 includes a processor (e.g., processor 124 or remote processor 154) that receives (at block 212) a set of unknown wavelength data (e.g., wavelength data 118) recorded by a detector (e.g., detector 114) for each of a set of unknown protein samples (e.g., unknown protein sample 140u).

[0093] In some embodiments, the set of unknown wavelength data (e.g., wavelength data 118) may be received as comma-separated value (CSV) data. In other embodiments, the set of standard wavelength data may be received in a database format as described herein, Extensible Markup Language (XML), JavaScript Object Notation (JSON), or other data or text formats. Figure 3A and Figure 3B In the embodiment of the computer program list, for example, at code segments 8 and 10, wavelength data of an unknown protein sample (e.g., unknown protein sample 140u) is read as CSV data into memory (e.g., memory 121 or remote memory 159) by a processor (e.g., processor 124 or remote processor 154), wherein the user inputs the data as a file containing wavelength data of the unknown protein sample (e.g., unknown protein sample 140u).

[0094] For each unknown protein sample, the set of unknown wavelength data (e.g., wavelength data 118) may include a second series of second absorbance-wavelength pairs across a second wavelength range selected from this series of single-wavelength beams. Each second absorbance-wavelength pair may include a UV absorbance value and a wavelength value. As described herein, it can be used in conjunction with... Figure 4A The same method is used to capture and / or measure the set of unknown wavelength data for standard protein samples (e.g., standard protein sample 140s).

[0095] like Figure 2A and Figure 2B As shown, the UV-based imaging method 200 continues (214), where, at block 216, a processor (e.g., processor 124 or remote processor 154) determines whether to filter the set of unknown wavelength data. This determination can be based on the range of wavelength data detected (e.g., by detector 114), some of which may include a wide wavelength range. If no filter is applied, the UV-based imaging method 200 proceeds to block 220.

[0096] However, at box 218, an unknown wavelength data filter is applied. The unknown wavelength data filter may include a set of program instructions stored in the memory of the UV imaging system 100 (e.g., in memory 121 or remote memory 154). For example, in Figure 3A and Figure 3B In the list of computer programs, pseudo-Python instructions illustrate an example of an unknown wavelength data filter (code snippet 7) for removing data points from the raw single-wavelength light range and raw UV absorbance value range, as measured by detector 114, to ensure optimal data quality for training calibration models as described herein. Figure 3AIn the example, code segment 7 applies filter criterion settings as user-defined variables (code segment 4) to allow for flexible manipulation of the filter range. For example, as shown in code segment 4, the filter criterion for an unknown protein sample (e.g., unknown protein sample 140u) is configured to select data points with UV absorbance values ​​of 0.2 to 1.2 for the wavelength range of 280 to 315 nm.

[0097] At block 218, the processor (e.g., processor 124 or remote processor 154) implements an unknown wavelength data filter for a series of single-wavelength beams to select or limit a second wavelength range. Alternatively, the processor may implement the unknown wavelength data filter to select or limit each second series of second absorbance-wavelength value pairs based on UV absorbance values ​​within a second filter range of UV absorbance values. For example, the application of the unknown wavelength data filter (e.g., at block 218) produces... Figure 6 Figure 600 shows a limited or filtered set of data (both wavelength and absorbance values, i.e., absorbance to wavelength pairs) of unknown protein sample data detected and / or received at box 212.

[0098] Specifically, Figure 6 Figure 600 illustrates various series of second absorbance versus wavelength pairs representing a second wavelength range (e.g., wavelength axis 604) selected from the UV spectrum, according to various embodiments disclosed herein. Figure 6 Similar to Figure 4B However, filtered data for unknown protein samples (e.g., unknown protein sample 140u) is shown. (As per...) Figure 6 As shown, the absorbance versus wavelength values ​​of 13 individual unknown protein samples (e.g., unknown protein sample 140u) are represented by a graph from unknown protein samples 600a to 600m (inclusive of each of 600a, 600c, 600f, and 600m). Figure 6 In some embodiments, as described herein, some absorbance-wavelength pairs have been filtered by an unknown wavelength data filter (e.g., at box 218) and are not shown in Figure 600. Specifically, Figure 6 An embodiment is shown in which an unknown wavelength data filter is applied to absorbance versus wavelength pairs plotted on a UV spectrum used to detect unknown wavelength data (e.g., wavelength data 118).

[0099] As by Figure 6 As shown, an unknown wavelength data filter (e.g., at box 218) is applied to select... Figure 6 The second wavelength range is 280 nm to 315 nm. The second wavelength range of 280 nm to 315 nm is... Figure 6 It is depicted on the wavelength axis 604. In various embodiments, Figure 6 The second wavelength range, from 280 to 315 nm, can be... Figure 4A The single-wavelength beam range is selected from 200nm to 400nm.

[0100] In addition, Figure 6 In one embodiment, an unknown wavelength data filter (e.g., at block 218) is applied to select a second absorbance-wavelength pair for each second series based on UV absorbance values ​​within a second filter range (e.g., UV absorbance axis 602) of UV absorbance values. The second filter range of UV absorbance values ​​from 0.2 to 1.2 is... Figure 6 It is plotted on the UV absorbance axis 602. Therefore, Figure 6 The second filter range, with UV absorbance values ​​ranging from 0.2 to 1.2, has been extended from... Figure 4A The UV absorbance value is selected from the range of 0 to 7. Therefore, Figure 6 The filtered UV spectrum of an unknown protein sample (e.g., unknown protein sample 140u) is shown. Specifically, the filtered UV spectrum of the unknown protein sample (e.g., unknown protein sample 140u) is shown in a limited range of wavelength values ​​from 280 to 315 nm and absorbance values ​​from 0.2 to 1.2.

[0101] exist Figure 6 In some instances, the unknown protein sample (e.g., unknown protein sample 140u) is a protein solution in the range of 0.1 to 126 mg / mg. At least for... Figure 6 In one embodiment, the protein solution of the unknown protein sample (e.g., unknown protein sample 140u) is used with the above-mentioned... Figure 4A and Figure 4B The same range of protein solutions.

[0102] about Figure 2B In block 220 of the UV-based imaging method 200, a processor (e.g., processor 124 or remote processor 154) uses a multi-wavelength calibration model to determine multiple protein concentration values ​​for each unknown protein sample in the set of unknown protein samples (e.g., unknown protein sample 140u). Each of these multiple protein concentration values ​​may correspond to a single-wavelength beam (e.g., ...) from the set of single-wavelength beams. Figure 4A The second wavelength range selected (e.g., the range of 200nm to 400nm) is the range of 200nm to 400nm. Figure 6 (ranging from 280nm to 315nm).

[0103] In some embodiments, at least one unknown protein sample in the group of unknown protein samples (e.g., unknown protein sample 140u) may include a protein therapeutic agent. In such embodiments, the protein therapeutic agent may be an antibody, an antigen-binding antibody fragment, an antibody protein product, or a bispecific T-cell conjugate. Any one of the following: molecule, bispecific antibody, trispecific antibody, Fc fusion protein, recombinant protein, and active fragment of recombinant protein.

[0104] exist Figure 3A and Figure 3B In the example list of computer programs, a processor (e.g., processor 124 or remote processor 154) uses a multi-wavelength calibration model to determine multiple protein concentration values ​​for each unknown protein sample in the set of unknown protein samples (e.g., unknown protein sample 140u). Specifically, in code segment 10, the "calc_unk" code function is called. The calc_unk code function (code segment 9) reads wavelength data (e.g., wavelength data 118) for each unknown protein sample in the set of unknown protein samples (e.g., unknown protein sample 140u) as CSV data. Then, the "calc_unk" code function (code segment 9) determines multiple protein concentration values ​​using a multi-wavelength calibration model (i.e., stored in memory as the variable "df"), and these values ​​are stored as the variable "conc". Figure 3A and Figure 3B In the embodiment, the "calc_unk" code function (code segment 9) is configured to sum and produce a second wavelength range (e.g., ...) by summing values ​​(e.g., various values ​​of the "conc" variable for each individual predicted concentration). Figure 6 The average predicted value of all protein concentrations in the 280nm to 315nm range was used to generate enhanced predictions.

[0105] Figure 7 Various embodiments disclosed herein are illustrated, depicting the use of Figure 5A Figure 700 shows the protein concentration values ​​determined by the multi-wavelength calibration model 501. Figure 7 In an embodiment, Figure 700 illustrates the application of the multi-wavelength calibration model 501 to a second wavelength range (e.g., for an unknown protein sample, such as an unknown protein sample 140u) for filtering. Figure 6 The range is 280 to 315 nm. Figure 700 depicts the protein concentration values ​​for each applicable wavelength.

[0106] As by Figure 7As shown, unknown protein samples (e.g., unknown protein sample 140u) are represented as unknown protein samples 700a to 700m, and are plotted as protein concentration values ​​along the protein concentration axis 702 (y-axis) and wavelength values ​​along the wavelength axis 704 (x-axis). The wavelength axis 704 (x-axis) includes... Figure 6 The wavelength values ​​are the same as those on wavelength axis 604 (i.e., the range of 280 nm to 315 nm). As described herein, each of the unknown protein samples 700a, 700c, 700f, and 700m corresponds to the same wavelength value (i.e., the range of 280 nm to 315 nm). Figure 6 Unknown protein samples 600a, 600c, 600f, and 600m.

[0107] Figure 7 The figure includes a zoom section 710, which depicts several unknown protein samples, including 700a and 700c, at a magnified scale. That is, zoom section 710 is equivalent to Figure 700, only at a different scale. Specifically, zoom section 710 displays protein concentration values ​​within a magnified range of 0 to 5 mg / ml, where several unknown protein samples (including 700a, 700c, and 700f) have protein concentration values ​​within this range. Each of the unknown protein samples 700a to 700f depicted by zoom section 710 corresponds to the unknown protein samples 700a to 700f depicted in Figure 700.

[0108] Each point of a given unknown protein sample (e.g., 700a to 700m) provides a prediction of the protein concentration value of the given sample at a given wavelength. For example, as shown by scaling section 710, the unknown protein sample 700c includes two points, 700c1 and 700c2. Scaling section 710 indicates that for point 700c1, the multi-wavelength calibration model 501 predicts that the unknown protein sample 700c has a protein concentration value prediction of approximately 0.5 at a wavelength of approximately 290 nm. Similarly, scaling section 710 indicates that for point 700c2, the multi-wavelength calibration model 501 predicts that the unknown protein sample 700c has a protein concentration value prediction of approximately 0.5 at a wavelength of approximately 298 nm.

[0109] As a further example, as shown by scaling portion 710, the unknown protein sample 700f includes two points 700f1 and 700f2. Scaling portion 710 indicates that for point 700f1, the multi-wavelength calibration model 501 predicts that the unknown protein sample 700f has a protein concentration value of approximately 4.0 at a wavelength of approximately 300 nm. Similarly, scaling portion 710 indicates that for point 700f2, the multi-wavelength calibration model 501 predicts that the unknown protein sample 700f has a protein concentration value of approximately 3.8 at a wavelength of approximately 303 nm.

[0110] As another example, as shown in Figure 700, the unknown protein sample 700m comprises two points, 700m1 and 700m2. Figure 400 shows that for point 700m1, the multi-wavelength calibration model 501 predicts a protein concentration value of approximately 122 for the unknown protein sample 700m at a wavelength of approximately 312 nm. Similarly, Figure 700 shows that for point 700m2, the multi-wavelength calibration model 501 predicts a protein concentration value of approximately 121 for the unknown protein sample 700m at a wavelength of approximately 313 nm.

[0111] Thus, as from Figure 7 As shown, a multi-wavelength calibration model (e.g., multi-wavelength calibration model 501) can take as input a series of absorbance-wavelength pairs (e.g., for...). Figure 6 The second series of absorbance-wavelength pairs described herein are used to predict each of the plurality of protein concentration values. In some embodiments, only the wavelength value or absorbance value of each absorbance-wavelength pair needs to be provided as input to the multi-wavelength calibration model 501 to generate a prediction. Thus, one or more wavelength values ​​and / or one or more absorbance values ​​for a specific wavelength (corresponding to a single-wavelength model) can be input to predict a protein concentration value for that specific absorbance value and / or that specific wavelength.

[0112] about Figure 2B At block 222 of the UV-based imaging method 200, the processor (e.g., processor 124 or remote processor 154) determines whether to generate enhancement prediction based on multiple protein concentration values ​​determined by the multi-wavelength calibration model 501. If enhancement prediction is not performed, the UV-based imaging method 200 proceeds to block 226.

[0113] At block 224 of the UV-based imaging method 200, the processor (e.g., processor 124 or remote processor 154) has determined to generate an enhancement prediction. This determination can be made, for example, by setting a threshold prediction quality value (e.g., in program instructions executed by the processor), thereby requiring the prediction to have sufficient quality or fit. This is, for example, by... Figure 3A Code snippet 3 proves this. This is also... Figure 3B This is demonstrated in code segment 9, where each protein concentration value is automatically averaged, thus within a certain wavelength range (e.g., for...). Figure 6 and Figure 7Enhanced predictions are created with enhanced accuracy within the described second wavelength range. In other embodiments, a flag may be set or provided to program instructions or code to indicate that enhanced predictions will be generated. Thus, in various embodiments, the generation of enhanced predictions may include averaging (or otherwise manipulating) each of the plurality of protein concentration values ​​corresponding to the second wavelength range. Enhanced predictions may include the average, median, derivative, or other manipulation of protein concentration values ​​determined by a multi-wavelength calibration model (e.g., multi-wavelength calibration model 501).

[0114] As by Figure 7 As shown, the enhanced prediction leverages the overlap between the wavelength range and predicted concentration values ​​for each sample (e.g., unknown protein samples 700a to 700m) and provides overlapping concentration measurements, resulting in enhanced accuracy compared to individual prediction points (e.g., points 700c1 or 700c2). In various embodiments, a multi-wavelength calibration model (e.g., multi-wavelength calibration model 501) can be updated using enhanced prediction or configured to calculate and / or output (e.g., ...) Figure 3B (As shown in code snippet 9) This enhanced prediction.

[0115] about Figure 2B In block 226 of the UV-based imaging method 200, a processor (e.g., processor 124 or remote processor 154) uses a multi-wavelength calibration model generated at block 220 or further updated or configured (e.g., manipulated at block 224) by enhanced predictions to monitor or measure the protein concentration of a protein-related product. In various embodiments, the protein-related product is a therapeutic product.

[0116] Figure 8 Various embodiments disclosed herein are illustrated, depicting applications. Figure 5A A flowchart 800 describes a multi-wavelength calibration model 501 used to monitor or measure the protein concentration of a protein-related product 860. As shown in Figure 800, a batch of unknown protein samples 840u are scanned by a UV imaging device (e.g., UV imaging device 101 as described herein) to generate wavelength data 818u for each unknown protein sample 840u. The unknown protein sample 840u can be the unknown protein sample 140u or it can be a new set of samples. The wavelength data 818u is consistent with the wavelength data 818u described herein. Figure 1 , 2A The wavelength data 818u is the same or of the same type as described in wavelength data 118, such as 2B. Wavelength data 818u can be input into a multi-wavelength calibration model 501, which can be previously generated from standard protein sample data, as referenced herein. Figure 2B , 2B As described in 3A, 3B, 4, and 5A.

[0117] exist Figure 8 In some embodiments, the multi-wavelength calibration model 501 is loaded into the memory (not shown) of the process control device 850. The process control device 850 includes a processor (not shown) for implementing the multi-wavelength calibration model 501 to monitor or measure the protein concentration of the protein-related product 860. In some embodiments, the process control device 850 receives the multi-wavelength calibration model 501 from the UV imaging device 101 and / or the computing device 151 via a computer network 130, such as via data packet transmission via Transmission Control Protocol and Internet Protocol (TCP / IP) or other similar protocols.

[0118] Process control device 850 can apply multi-wavelength calibration model 501 by outputting prediction 852 (an enhanced prediction or a prediction based on one or more calculated protein concentration values ​​for each standard protein sample as described herein). Prediction 852 can then be used to monitor or measure the protein concentration of protein-related product 860. For example, in some embodiments, prediction 852, or the signal or data transmission based on prediction 852, can be transmitted or otherwise provided to a sub-device (not shown) of a process control or manufacturing system, such as a field device communicatively coupled to process control device 850 via a process control network, for monitoring or measuring the protein concentration of protein-related product 860. Thus, prediction 852 can be output from process control device 850 and used to monitor, test, measure, or develop protein concentrations in end-use products. For example, in one embodiment, prediction 852 can be used to monitor or measure the protein concentration 860 of protein-related product 865 during the manufacturing process of protein-related product 865. Protein-related product 865 can be a product developed via an assembly or automation system in a process control location using process control device 850. For example, based on the measured protein concentration, the protein concentration of a protein-related product can be increased or decreased to fall within a specific range of protein concentration. For example, a batch of protein-related products can be accepted or rejected depending on whether the measured protein concentration falls within or outside a specific range. Due to the robustness of the multi-wavelength calibration model 501 implemented on the process control device 850, the protein concentration of the protein-related product 865 can be measured or monitored over a wide range of UV spectra, including the UV spectral range as described herein.

[0119] In a further example, a prediction 852 from a multi-wavelength calibration model 501 provided by process control device 850 can be used to monitor or measure the protein concentration 860 of protein-related product 865 against known specifications of the protein-related product, or to compare it with known specifications of the protein-related product. Known specifications can be defined, for example, by standard protein samples 140s. In such embodiments, process control device 850 can apply prediction 852 to protein-related product 865 to compare it with known specifications and measure compliance with regulatory requirements, efficacy or quality thresholds, or other regulatory or quality issues.

[0120] Therefore, as provided herein, the UV-based imaging system and method of this disclosure improves upon the limitations of currently known instruments and methods. The UV-based imaging system and method described herein overcome the linear limitation of the Beer-Lambert law by generating and applying a unique multi-wavelength calibration model 501 calibrated with a set of protein sample standards. Furthermore, unlike other instruments, such as those that mechanically scan an extended linear range of protein concentrations with mechanical path lengths, the UV-based imaging system and method of this disclosure generates and uses a multi-wavelength calibration model (e.g., multi-wavelength calibration model 501) that avoids mechanical error problems.

[0121] Therapeutic proteins come in a variety of forms and formats; this approach is independent of the type of therapeutic protein. In the case of antibodies as therapeutic proteins, for example, antibody architectures have been used to generate an increasing number of alternative forms, spanning a molecular weight range of at least about 12–150 kDa and having a range of titers (n) from monomers (n=1), to dimers (n=2), to trimers (n=3), to tetramers (n=4), and possibly higher; such alternative forms are referred to herein as “antibody protein products.” Antibody protein products include those based on the intact antibody structure and those mimicking antibody fragments that retain the intact antigen-binding capacity, such as scFv, Fab, and VHH / VH (discussed below). The smallest antigen-binding antibody fragment that retains its intact antigen-binding site is the Fv fragment, which consists entirely of a variable (V) region. The molecule is stabilized by linking the V region to the scFv fragment (a variable single-chain fragment) using a soluble, flexible amino acid peptide linker, or by adding a constant (C) domain to the V region to generate the Fab fragment [antigen-binding fragment]. scFv and Fab fragments can be readily generated in host cells (e.g., prokaryotic host cells). Other antibody protein products include disulfide-stabilized scFv (ds-scFv), single-chain Fab (scFab), and dimer and polymeric antibody forms such as bifunctional, trifunctional, and tetrafunctional antibodies, or mini-abs comprising various forms of scFv linked to oligomeric domains. The smallest fragments are the VHH / VH of camel heavy chain Abs and single-domain Abs (sdAbs). The most commonly used building blocks for constructing novel antibody forms are single-chain variable (V) domain antibody fragments (scFvs) containing V domains (VH and VL domains) from the heavy and light chains linked by peptide linkers having approximately 15 amino acid residues. Peptibody, or peptide-Fc fusion, is another antibody protein product. The structure of a peptibody consists of a bioactive peptide grafted onto an Fc domain. Peptibody has been well described in the art. See, for example, Shimamoto et al., mAbs 4(5):586-591 (2012).

[0122] Other antibody protein products include single-chain antibodies (SCAs), bifunctional antibodies, triantibodies, tetraantibodies, and bispecific or trispecific antibodies. Bispecific antibodies can be divided into five main categories: BsIgG, attached IgG, BsAb fragments, bispecific fusion proteins, and BsAb conjugates. See, for example, Spiess et al., Molecular Immunology 67(2) Part A:97 / -106 (2015).

[0123] In an exemplary aspect, the therapeutic protein comprises any of these antibody protein products. In an exemplary aspect, the therapeutic protein comprises any of the following: scFv, Fab VHH / VH, Fv fragment, ds-scFv, scFab, dimer antibody, multimeric antibody (e.g., biantibody, triantibody, tetraantibody), miniAb, peptide, VHH / VH of camel heavy chain antibody, sdAb, biantibody; triantibody; tetraantibody; bispecific or trispecific antibody, BsIgG, additional IgG, BsAb fragment, bispecific fusion protein, and BsAb conjugate.

[0124] In an exemplary aspect, the therapeutic protein is a bispecific T-cell conjugate. molecular. A molecule is a bispecific antibody construct or a bispecific fusion protein containing two antibody-binding domains (or targeting regions) linked together. The molecule consists of two single-stranded variable segments (scFv) linked together by a flexible linker (scFc-scFc, Huehls et al., 2015). One arm of the molecule is engineered to bind to a protein found on the surface of cytotoxic T cells, while the other arm is designed to bind to a specific protein primarily found on tumor cells. When both targets are bound, The molecule forms a bridge between cytotoxic T cells and tumor cells, enabling T cells to recognize tumor cells and fight them by injecting the toxic molecule. The tumor-binding arm of the molecule can be altered to create different molecules targeting different types of cancer. Antibody constructs. (About...) The molecular term "binding domain" refers to a domain that (specifically) binds to, interacts with, or recognizes a given target epitope or target site of a target molecule (antigen). The structure and function of the first binding domain (recognizing tumor cell antigens), and preferably also the structure and / or function of a second binding domain (cytotoxic T-cell antigens), are based on the structure and / or function of the antibody (e.g., a full-length or intact immunoglobulin molecule). For example, The molecule comprises a first binding domain characterized by the presence of three light chain CDRs (i.e., CDR1, CDR2, and CDR3 in the VL region) and three heavy chain CDRs (i.e., CDR1, CDR2, and CDR3 in the VH region). The second binding domain preferably also includes the minimum structural requirements for antibody binding to the target. More preferably, the second binding domain comprises at least three light chain CDRs (i.e., CDR1, CDR2, and CDR3 in the VL region) and / or three heavy chain CDRs (i.e., CDR1, CDR2, and CDR3 in the VH region). It is envisioned that the first and / or second binding domains are generated or available through phage display or library screening methods, rather than by transplanting CDR sequences from a pre-existing (monoclonal) antibody into a scaffold. The binding domain may typically comprise both the antibody light chain variable region (VL) and the antibody heavy chain variable region (VH); however, it does not necessarily contain both. For example, Fd fragments have two VH regions and typically retain some antigen-binding functionality of the intact antigen-binding domain. Examples of (modified) antibody fragments that bind antigens include (1) Fab fragments, a monovalent fragment having VL, VH, CL and CH1 domains; (2) F(ab')2 fragments, a bivalent fragment having two Fab fragments connected in the hinge region by disulfide bridges; (3) Fd fragments having two VH and CH1 domains; (4) Fv fragments having VL and VH domains in a single arm of the antibody; (5) dAb fragments having VH domains (Ward et al., (1989) Nature 341:544-546); (6) separated complementarity-determining regions (CDRs); and (7) single-chain Fv (scFv), the latter being preferred (e.g., derived from scFV libraries).

[0125] In exemplary aspects, the therapeutic protein is a bispecific antibody, a trispecific antibody, a molecule, or an Fc fusion protein. In some aspects, the therapeutic protein includes a single-chain variable fragment (scFV) that binds to a first target, optionally wherein the therapeutic protein includes a second scFV that binds to a second target (optionally different from the first target).

[0126] In some respects, therapeutic proteins are recombinant therapeutic proteins or their active fragments. In other respects, therapeutic proteins are endogenous to cells or organisms and are isolated from cells or organisms.

[0127] All aspects of this disclosure

[0128] The following aspects of this disclosure are merely exemplary and are not intended to limit the scope of this disclosure.

[0129] 1. An ultraviolet (UV)-based imaging system configured to determine the protein concentration of an unknown protein sample based on automatic multi-wavelength calibration, the UV-based imaging system comprising: a light source configured to project a multi-wavelength beam; a monochromator configured to receive the multi-wavelength beam and project a series of single-wavelength beams of a UV spectrum based on the multi-wavelength beam; a sample holder operable to receive a protein sample, the protein sample being (1) a standard protein sample selected from a set of standard protein samples, or (2) an unknown protein sample selected from a set of unknown protein samples, the unknown protein sample having an unknown protein concentration, wherein the sample holder is positioned to allow the series of single-wavelength beams to pass through at least a portion of the protein sample; a detector operable to detect the series of single-wavelength beams passing through at least a portion of the protein sample; a memory storing program instructions; and a processor communicatively coupled to the memory, wherein the processor is configured to execute the program instructions to cause the processor to: receive, as determined by the detector, the protein concentration of the set of standard protein samples. The set of standard wavelength data recorded for each standard protein sample includes, for each standard protein sample, a first series of first absorbance-wavelength pairs across a first wavelength range selected from the series of single-wavelength beams, each first absorbance-wavelength pair including a UV absorbance value and a wavelength value. A multi-wavelength calibration model is generated based on each of the first series of first absorbance-wavelength pairs of the set of standard wavelength data. A set of unknown wavelength data recorded by the detector for each unknown protein sample in the set of unknown protein samples is received, wherein for each unknown protein sample, the set of unknown wavelength data includes a second series of second absorbance-wavelength pairs across a second wavelength range selected from the series of single-wavelength beams, each second absorbance-wavelength pair including a UV absorbance value and a wavelength value. The multi-wavelength calibration model is used to determine multiple protein concentration values ​​for each unknown protein sample in the set of unknown protein samples, each of the multiple protein concentration values ​​corresponding to the second wavelength range selected from the series of single-wavelength beams.

[0130] 2. The UV-based imaging system as described in aspect 1, wherein the light source, the monochromator, the sample holder, and the detector comprise part of a spectrophotometer apparatus.

[0131] 3. The UV-based imaging system as described in any of the preceding aspects, wherein the processor is communicatively coupled to the detector.

[0132] 4. The UV-based imaging system as described in any of the preceding aspects, wherein the memory or the processor is located remotely from the detector.

[0133] 5. The UV-based imaging system as described in any of the preceding aspects, wherein the set of standard protein samples comprises a plurality of standard protein samples, and wherein the set of unknown protein samples comprises a single protein sample having an unknown protein concentration.

[0134] 6. The UV-based imaging system as described in any of the preceding aspects, wherein the set of standard protein samples is a protein standard solution in the range of 0.1 to 220 mg per milliliter (ml).

[0135] 7. The UV-based imaging system as described in any of the preceding aspects, wherein the series of single-wavelength beams comprises single-wavelength beams in the UV spectrum from 200 nanometers (nm) to 400 nanometers (nm).

[0136] 8. The UV-based imaging system as described in aspect 7, wherein each single-wavelength beam in the UV spectrum is separated by a one-nanometer interval.

[0137] 9. The UV-based imaging system as described in any of the preceding aspects, wherein generating the multi-wavelength calibration model includes generating a single-wavelength calibration model for each of the first wavelength range selected from the series of single-wavelength beams, each single-wavelength calibration model including a slope value and a y-intercept value corresponding to the UV absorbance of a particular single wavelength.

[0138] 10. The UV-based imaging system as described in any of the preceding aspects, wherein the multi-wavelength calibration model is trained as a machine learning model.

[0139] 11. The UV-based imaging system as described in any of the preceding aspects, further comprising a standard wavelength data filter, the processor being further configured to execute the program instructions to cause the processor to apply the standard wavelength data filter to the series of single-wavelength beams, thereby selecting the first wavelength range.

[0140] 12. The UV-based imaging system of aspect 12, wherein the processor is further configured to execute these program instructions to cause the processor to implement the standard wavelength data filter, thereby limiting each first series of first absorbance-wavelength value pairs based on UV absorbance values ​​within a first filter range of UV absorbance values.

[0141] 13. The UV-based imaging system as described in any of the preceding aspects, further comprising an unknown wavelength data filter, the processor being further configured to execute the program instructions to cause the processor to implement the unknown wavelength data filter for the series of single-wavelength beams, thereby selecting the second wavelength range.

[0142] 14. The UV-based imaging system of aspect 13, wherein the processor is configured to execute these program instructions to cause the processor to implement the unknown wavelength data filter, thereby limiting each second series of second absorbance-wavelength value pairs based on UV absorbance values ​​within a second filter range of UV absorbance values.

[0143] 15. The UV-based imaging system as described in any of the preceding aspects, wherein the processor is further configured to generate enhanced predictions based on the plurality of protein concentration values.

[0144] 16. The UV-based imaging system of aspect 15, wherein the generation of the enhancement prediction includes averaging each of the plurality of protein concentration values ​​corresponding to the second wavelength range.

[0145] 17. The UV-based imaging system as described in any of the preceding aspects, wherein the multi-wavelength calibration model is used to monitor or measure the protein concentration of protein-related products.

[0146] 18. The UV-based imaging system as described in aspect 17, wherein the protein concentration is monitored or measured during the manufacturing process of the protein-related product.

[0147] 19. A UV-based imaging system as described in any one or more of aspects 17 or 18, wherein the protein concentration of the protein-related product is compared with known specifications of the protein-related product.

[0148] 20. The UV-based imaging system as described in any one or more of aspects 17 to 19, wherein the protein-related product is a therapeutic product.

[0149] 21. A UV-based imaging system as described in any one or more of aspects 17 to 20, wherein the multi-wavelength calibration model is input to these second series of second absorbance-wavelength value pairs to predict each of the plurality of protein concentration values.

[0150] 22. A UV-based imaging method for determining the protein concentration of an unknown protein sample based on automated multi-wavelength calibration, the UV-based imaging method comprising: receiving at a processor a set of standard wavelength data as recorded by a detector for each of a set of standard protein samples, wherein for each standard protein sample, the set of standard wavelength data includes a first series of first absorbance-wavelength pairs spanning a first wavelength range selected from a series of single-wavelength beams of the UV spectrum, each first absorbance-wavelength pair including a UV absorbance value and a wavelength value; and using the processor, generating a multi-wavelength calibration model based on each of these first series of first absorbance-wavelength pairs of the set of standard wavelength data. The processor receives a set of unknown wavelength data as recorded by the detector for each unknown protein sample in a set of unknown protein samples, wherein for each unknown protein sample, the set of unknown wavelength data includes a second series of second absorbance-wavelength value pairs spanning a second wavelength range selected from the series of single-wavelength beams, each second absorbance-wavelength value pair including a UV absorbance value and a wavelength value; and determines a plurality of protein concentration values ​​for each unknown protein sample in the set of unknown protein samples using the multi-wavelength calibration model as implemented on the processor, each of the plurality of protein concentration values ​​corresponding to the second wavelength range selected from the series of single-wavelength beams.

[0151] 23. The UV-based imaging method as described in aspect 22, wherein the detector includes part of a spectrophotometer apparatus.

[0152] 24. The UV-based imaging method as described in any one or more of aspects 22 to 23, wherein the processor is communicatively coupled to the detector.

[0153] 25. The UV-based imaging method as described in any one or more of aspects 22 to 24, wherein the processor is located remotely from the detector.

[0154] 26. The UV-based imaging method as described in any one or more of aspects 22 to 25, wherein the set of standard protein samples comprises a plurality of standard protein samples, and wherein the set of unknown protein samples comprises a single protein sample having an unknown protein concentration.

[0155] 27. The UV-based imaging method as described in any one or more of aspects 22 to 26, wherein the set of standard protein samples is a protein standard solution in the range of 0.1 to 220 mg per milliliter (ml).

[0156] 28. The UV-based imaging method as described in any one or more of aspects 22 to 27, wherein the series of single-wavelength beams includes single-wavelength beams in the UV spectrum from 200 nanometers (nm) to 400 nanometers (nm).

[0157] 29. The UV-based imaging method as described in aspect 28, wherein each single-wavelength beam in the UV spectrum is separated by a one-nanometer interval.

[0158] 30. The UV-based imaging method as described in any one of aspects 22 to 28, wherein generating the multi-wavelength calibration model includes generating a single-wavelength calibration model for each of the first wavelength range selected from the series of single-wavelength beams, each single-wavelength calibration model including a slope value and a y-intercept value corresponding to the UV absorbance of a particular single wavelength.

[0159] 31. The UV-based imaging method as described in any one or more of aspects 22 to 30, wherein the multi-wavelength calibration model is trained as a machine learning model.

[0160] 32. The UV-based imaging method as described in any one or more of aspects 22 to 31, further comprising a standard wavelength data filter, wherein the processor is further configured to execute the program instructions to cause the processor to implement the standard wavelength data filter for the series of single-wavelength beams, thereby selecting the first wavelength range.

[0161] 33. The UV-based imaging method as described in aspect 32, wherein the processor is further configured to execute these program instructions to cause the processor to implement the standard wavelength data filter, thereby limiting each first series of first absorbance-wavelength value pairs based on UV absorbance values ​​within a first filter range of UV absorbance values.

[0162] 34. The UV-based imaging method as described in any one or more of aspects 22 to 33, further comprising an unknown wavelength data filter, wherein the processor is further configured to execute the program instructions to cause the processor to implement the unknown wavelength data filter for the series of single-wavelength beams, thereby selecting the second wavelength range.

[0163] 35. The UV-based imaging method as described in aspect 34, wherein the processor is configured to execute these program instructions to cause the processor to implement the unknown wavelength data filter, thereby limiting each second series of second absorbance-wavelength value pairs based on UV absorbance values ​​within a second filter range of UV absorbance values.

[0164] 36. The UV-based imaging method as described in any one or more of aspects 22 to 35, wherein the processor is further configured to generate enhancement predictions based on the plurality of protein concentration values.

[0165] 37. The UV-based imaging method as described in aspect 36, wherein the generation of the enhancement prediction includes averaging each of the plurality of protein concentration values ​​corresponding to the second wavelength range.

[0166] 38. The UV-based imaging method as described in any one or more of aspects 22 to 37, wherein the multi-wavelength calibration model is used to monitor or measure the protein concentration of a protein-related product.

[0167] 39. The UV-based imaging method as described in aspect 38, wherein the protein concentration is monitored or measured during the manufacturing process of the protein-related product.

[0168] 40. The UV-based imaging method as described in aspect 38 or 39, wherein the protein concentration of the protein-related product is compared with known specifications of the protein-related product.

[0169] 41. The UV-based imaging system as described in any one or more of aspects 38 to 40, wherein the protein-related product is a therapeutic product.

[0170] 42. The UV-based imaging method as described in any one or more of aspects 22 to 41, wherein the multi-wavelength calibration model is input to these second series of second absorbance-wavelength value pairs to predict each of the plurality of protein concentration values.

[0171] 43. A tangible, non-transitory computer-readable medium storing instructions for determining the protein concentration of an unknown protein sample based on automatic multi-wavelength calibration, the instructions, when executed by one or more processors of a computing device, causing the computing device to: receive at the processor a set of standard wavelength data, such as that recorded by a detector for each of a set of standard protein samples, wherein for each standard protein sample, the set of standard wavelength data includes a first series of first absorbance-wavelength value pairs across a first wavelength range selected from a series of single-wavelength beams of a UV spectrum, each first absorbance-wavelength value pair including a UV absorbance value and a wavelength value; and, using the processor, determine the protein concentration of an unknown protein sample based on these first series of first absorbance-wavelength value pairs of the set of standard wavelength data. Each of the absorbance-wavelength value pairs generates a multi-wavelength calibration model; the processor receives a set of unknown wavelength data as recorded by the detector for each unknown protein sample in a set of unknown protein samples, wherein for each unknown protein sample, the set of unknown wavelength data includes a second series of second absorbance-wavelength value pairs spanning a second wavelength range selected from the series of single-wavelength beams, each second absorbance-wavelength value pair including a UV absorbance value and a wavelength value; and using the multi-wavelength calibration model as implemented on the processor, a plurality of protein concentration values ​​are determined for each unknown protein sample in the set of unknown protein samples, each of the plurality of protein concentration values ​​corresponding to the second wavelength range selected from the series of single-wavelength beams.

[0172] 44. An imaging apparatus for determining the protein concentration of an unknown protein sample based on automated multi-wavelength calibration, the imaging apparatus comprising: means for receiving a set of standard wavelength data for each of a set of standard protein samples, wherein for each standard protein sample, the set of standard wavelength data includes a first series of first absorbance-wavelength pairs across a first wavelength range selected from a series of single-wavelength beams of a UV spectrum, each first absorbance-wavelength pair including a UV absorbance value and a wavelength value; and means for generating a multi-wavelength calibration model based on each of the first series of first absorbance-wavelength pairs of the set of standard wavelength data. The apparatus includes: means for receiving a set of unknown wavelength data for each unknown protein sample in a set of unknown protein samples, wherein for each unknown protein sample, the set of unknown wavelength data includes a second series of second absorbance-wavelength pairs spanning a second wavelength range selected from the series of single-wavelength beams, each second absorbance-wavelength pair including a UV absorbance value and a wavelength value; and means for determining a plurality of protein concentration values ​​for each unknown protein sample in the set of unknown protein samples using the multi-wavelength calibration model, each of the plurality of protein concentration values ​​corresponding to the second wavelength range selected from the series of single-wavelength beams.

[0173] 45. The UV-based imaging method as described in any one or more of aspects 22 to 41, wherein at least one of the unknown protein samples in the group of unknown protein samples contains a protein therapeutic agent.

[0174] 46. ​​The UV-based imaging method as described in aspect 45, wherein the protein therapeutic agent is selected from the group consisting of: antibodies, antigen-binding antibody fragments, antibody protein products, and bispecific T-cell conjugates. Molecules, bispecific antibodies, trispecific antibodies, Fc fusion proteins, recombinant proteins, and active fragments of recombinant proteins.

[0175] Additional considerations

[0176] While the disclosure herein sets forth detailed descriptions of many different embodiments, it should be understood that the legal scope of this specification is defined by the words of the claims set forth at the end of this patent and their equivalents. This detailed description is to be construed as exemplary only and does not describe every possible embodiment, as describing every possible embodiment would be impractical. Many alternative embodiments may be implemented using current technology or technology developed after the filing date of this patent, which still fall within the scope of these claims.

[0177] The following additional considerations apply to the foregoing discussion. Throughout this specification, multiple instances can implement components, operations, or structures described as single instances. Although the various operations of one or more methods are shown and described as separate operations, one or more of the various operations can be performed simultaneously, and the operations do not need to be performed in the order shown. Structures and functions presented as separate components in the example configuration can be implemented as composite structures or components. Similarly, structures and functions presented as single components can be implemented as separate components. These and other variations, modifications, additions, and improvements all fall within the scope of this document.

[0178] Additionally, certain embodiments herein are described as including logic or a plurality of routines, subroutines, applications, or instructions. These may constitute software (e.g., code embodied on a machine-readable medium or in transmitted signals) or hardware. In hardware, routines, etc., are tangible units capable of performing certain operations and can be configured or arranged in a certain manner. In example embodiments, one or more computer systems (e.g., standalone client or server computer systems) or one or more hardware modules (e.g., processors or a group of processors) of a computer system may be configured by software (e.g., an application or an application portion) to operate as hardware modules to perform certain operations as described herein.

[0179] In various embodiments, hardware modules can be implemented mechanically or electronically. For example, a hardware module may include a dedicated circuit system or logic that is permanently configured (e.g., as a dedicated processor, such as a field-programmable gate array (FPGA), or as an application-specific integrated circuit (ASIC)) to perform specific operations. A hardware module may also include programmable logic or circuit systems that are temporarily configured by software to perform specific operations (e.g., as contained within a general-purpose processor or other programmable processor). It will be understood that cost and time considerations can drive the decision to implement the hardware module mechanically in a dedicated and permanently configured circuit system or in a temporarily configured circuit system (e.g., configured by software).

[0180] Accordingly, the term "hardware module" should be understood to include tangible entities, meaning entities that are physically constructed, permanently configured (e.g., hardwired), or temporarily configured (e.g., programmed) to operate in a particular manner or perform the specific operations described herein. Consider embodiments where hardware modules are temporarily configured (e.g., programmed), each hardware module does not need to be configured or instantiated at any given time. For example, in cases where the hardware modules include a general-purpose processor configured using software, the general-purpose processor can be configured as distinct hardware modules at different times. The software can accordingly configure the processor, for example, to constitute a specific hardware module at one time and different hardware modules at different times.

[0181] Hardware modules can provide and receive information from other hardware modules. Therefore, the described hardware modules can be considered communicatively coupled. When multiple such hardware modules exist simultaneously, communication can be achieved through signal transmission connecting the hardware modules (e.g., via appropriate circuitry and buses). In embodiments where multiple hardware modules are configured or instantiated at different times, such communication between hardware modules can be achieved, for example, by storing and retrieving information in a memory structure accessible to the multiple hardware modules. For example, one hardware module can perform an operation and store the output of that operation in a storage device communicatively coupled to it. Another hardware module can then access the storage device at a later time to retrieve and process the stored output. Hardware modules can also initiate communication with input or output devices and can operate on resources (e.g., collections of information).

[0182] The various operations of the example methods described herein can be performed, at least in part, by one or more processors that are temporarily configured (e.g., by software) or permanently configured to perform the relevant operations. Whether temporarily or permanently configured, such processors can constitute processor-implemented modules that operate to perform one or more operations or functions. In some example embodiments, the modules mentioned herein may include processor-implemented modules.

[0183] Similarly, the methods or routines described herein can be implemented at least in part by a processor. For example, at least some operations of the method can be performed by one or more processors or hardware modules implemented by processors. The execution of certain operations can be distributed across one or more processors, residing not only within a single machine but also deployed across multiple machines. In some example embodiments, one or more processors may reside in a single location, while in other embodiments, processors may be distributed across multiple locations.

[0184] The execution of certain operations in an operation can be distributed across one or more processors, residing not only within a single machine but also deployed across multiple machines. In some example embodiments, one or more processors or processor-implemented modules may reside in a single geographic location (e.g., within a home environment, office environment, or server cluster). In other embodiments, one or more processors or processor-implemented modules may be distributed across multiple geographic locations.

[0185] This detailed description is to be interpreted as exemplary only and does not describe every possible embodiment, as describing every possible embodiment would be impractical if not impossible. Many alternative embodiments can be implemented by those skilled in the art using current technology or technology developed after the filing date of this application.

[0186] Those skilled in the art will understand that various modifications, alterations, and combinations can be made to the embodiments described above without departing from the scope of the invention, and such modifications, alterations, and combinations can be considered to be within the scope of the inventive concept.

[0187] Unless conventional device-plus-function language is explicitly enumerated, such as the language of "means for..." or "steps for..." as explicitly enumerated in the (multiple) claims, the patent claims at the end of this patent application are not intended to be interpreted in accordance with 35 USC §112(f). The systems and methods described herein relate to improving computer functionality and improving the operation of conventional computers.

Claims

1. A UV-based imaging system configured to determine the protein concentration of an unknown protein sample based on automated multi-wavelength calibration, the UV-based imaging system comprising: A light source (102) is configured to project a multi-wavelength beam (104); A monochromator (106) is configured to receive the multi-wavelength beam and project a series of single-wavelength beams of UV spectrum based on the multi-wavelength beam; A sample holder (110) operable to receive a protein sample, which is (1) a standard protein sample selected from a set of standard protein samples (140s) or (2) an unknown protein sample selected from a set of unknown protein samples (140u) having an unknown protein concentration, wherein the sample holder is positioned to allow the series of single-wavelength beams to pass through at least a portion of the protein sample. Detector (114), operable to detect the series of single-wavelength beams passing through at least a portion of the protein sample; Memory that stores program instructions; as well as The processor is communicatively coupled to the memory, wherein The processor is configured to execute these program instructions to cause the processor to: Receive a set of standard wavelength data recorded by the detector for each standard protein sample in the set of standard protein samples, wherein for each standard protein sample, the set of standard wavelength data includes a first series of first absorbance-wavelength value pairs across a first wavelength range selected from the series of single-wavelength beams, each first absorbance-wavelength value pair including a UV absorbance value and a wavelength value. A multi-wavelength calibration model is generated based on each of the first series of first absorbance-wavelength value pairs from this set of standard wavelength data. Receive a set of unknown wavelength data recorded by the detector for each unknown protein sample in the set of unknown protein samples, wherein for each unknown protein sample, the set of unknown wavelength data includes a second series of second absorbance-wavelength value pairs spanning a second wavelength range selected from the series of single-wavelength beams, each second absorbance-wavelength value pair including a UV absorbance value and a wavelength value, and Using this multi-wavelength calibration model, multiple protein concentration values ​​are determined for each unknown protein sample in the group of unknown protein samples, each of which corresponds to the second wavelength range selected from the series of single-wavelength beams.

2. The UV-based imaging system of claim 1, wherein the light source, the monochromator, the sample holder, and the detector comprise a portion of a spectrophotometer apparatus.

3. The UV-based imaging system of claim 1 or claim 2, wherein the processor is communicatively coupled to the detector.

4. The UV-based imaging system of claim 1 or claim 2, wherein the memory or the processor is located remotely from the detector.

5. The UV-based imaging system of claim 1 or claim 2, wherein the set of standard protein samples comprises a plurality of standard protein samples, and wherein the set of unknown protein samples comprises a single protein sample having an unknown protein concentration.

6. The UV-based imaging system of claim 1 or claim 2, wherein the set of standard protein samples is a protein standard solution in the range of 0.1 to 220 mg per milliliter (ml).

7. The UV-based imaging system of claim 1 or claim 2, wherein the series of single-wavelength beams comprises single-wavelength beams in the UV spectrum from 200 nanometers (nm) to 400 nanometers (nm).

8. The UV-based imaging system of claim 7, wherein each single-wavelength beam in the UV spectrum is separated by a one-nanometer interval.

9. The UV-based imaging system of claim 1 or claim 2, wherein generating the multi-wavelength calibration model comprises generating a single-wavelength calibration model for each of the first wavelength range selected from the series of single-wavelength beams, each single-wavelength calibration model comprising a slope value and a y-intercept value corresponding to the UV absorbance of a particular single wavelength.

10. The UV-based imaging system of claim 1 or claim 2, wherein the multi-wavelength calibration model is trained as a machine learning model.

11. The UV-based imaging system of claim 1 or claim 2, further comprising a standard wavelength data filter, the processor being further configured to execute the program instructions to cause the processor to implement the standard wavelength data filter for the series of single-wavelength beams, thereby selecting the first wavelength range.

12. The UV-based imaging system of claim 11, wherein the processor is further configured to execute the program instructions to cause the processor to implement the standard wavelength data filter, thereby limiting each first series of first absorbance-wavelength value pairs based on UV absorbance values ​​within a first filter range of UV absorbance values.

13. The UV-based imaging system of claim 1 or claim 2, further comprising an unknown wavelength data filter, the processor being further configured to execute the program instructions to cause the processor to implement the unknown wavelength data filter for the series of single-wavelength beams, thereby selecting the second wavelength range.

14. The UV-based imaging system of claim 13, wherein the processor is configured to execute the program instructions to cause the processor to implement the unknown wavelength data filter, thereby limiting each second series of second absorbance-wavelength value pairs based on UV absorbance values ​​within a second filter range of UV absorbance values.

15. The UV-based imaging system of claim 1 or claim 2, wherein the processor is further configured to generate enhanced predictions based on the plurality of protein concentration values.

16. The UV-based imaging system of claim 15, wherein generating the enhancement prediction includes averaging each of the plurality of protein concentration values ​​corresponding to the second wavelength range.

17. The UV-based imaging system of claim 1 or claim 2, wherein the multi-wavelength calibration model is used to monitor or measure the protein concentration of protein-related products.

18. The UV-based imaging system of claim 17, wherein the protein concentration is monitored or measured during the manufacturing process of the protein-related product.

19. The UV-based imaging system of claim 17, wherein the protein concentration of the protein-related product is compared with known specifications of the protein-related product.

20. The UV-based imaging system of claim 17, wherein the protein-related product is a therapeutic product.

21. The UV-based imaging system of claim 1 or claim 2, wherein the multi-wavelength calibration model takes into account these second series of second absorbance-wavelength value pairs to predict each plurality of protein concentration values.

22. A UV-based imaging method for determining the protein concentration of an unknown protein sample based on automated multi-wavelength calibration, the UV-based imaging method comprising: At the processor, a set of standard wavelength data is received, as recorded by detector (114) for each of a set of standard protein samples (140s), wherein for each standard protein sample, the set of standard wavelength data includes a first series of first absorbance-wavelength value pairs across a first wavelength range selected from a series of single-wavelength beams from the UV spectrum, each first absorbance-wavelength value pair including a UV absorbance value and a wavelength value. Using this processor, a multi-wavelength calibration model is generated based on each of these first series of first absorbance-wavelength value pairs of the set of standard wavelength data. The processor receives a set of unknown wavelength data recorded by the detector for each of a set of unknown protein samples (140u), wherein for each unknown protein sample, the set of unknown wavelength data includes a second series of second absorbance-wavelength value pairs across a second wavelength range selected from the series of single-wavelength beams, each second absorbance-wavelength value pair including a UV absorbance value and a wavelength value. as well as Using the multi-wavelength calibration model implemented on the processor, multiple protein concentration values ​​are determined for each unknown protein sample in the group of unknown protein samples, each of the multiple protein concentration values ​​corresponding to the second wavelength range selected from the series of single-wavelength beams.

23. The UV-based imaging method of claim 22, wherein the detector comprises part of a spectrophotometer device.

24. The UV-based imaging method of claim 22 or claim 23, wherein the processor is communicatively coupled to the detector.

25. The UV-based imaging method of claim 22 or claim 23, wherein the processor is located remotely from the detector.

26. The UV-based imaging method of claim 22 or claim 23, wherein the set of standard protein samples comprises a plurality of standard protein samples, and wherein the set of unknown protein samples comprises a single protein sample having an unknown protein concentration.

27. The UV-based imaging method of claim 22 or claim 23, wherein the set of standard protein samples is a protein standard solution in the range of 0.1 to 220 mg per milliliter (ml).

28. The UV-based imaging method of claim 22 or claim 23, wherein the series of single-wavelength beams comprises single-wavelength beams in the UV spectrum from 200 nanometers (nm) to 400 nanometers (nm).

29. The UV-based imaging method of claim 28, wherein each single-wavelength beam in the UV spectrum is separated by a one-nanometer interval.

30. The UV-based imaging method of claim 22 or claim 23, wherein generating the multi-wavelength calibration model comprises generating a single-wavelength calibration model for each of the first wavelength range selected from the series of single-wavelength beams, each single-wavelength calibration model comprising a slope value and a y-intercept value corresponding to the UV absorbance of a particular single wavelength.

31. The UV-based imaging method of claim 22 or claim 23, wherein the multi-wavelength calibration model is trained as a machine learning model.

32. The UV-based imaging method as described in claim 22 or claim 23, further comprising: A standard wavelength data filter is provided, and the processor is further configured to execute these program instructions to enable the processor to implement the standard wavelength data filter for the series of single-wavelength beams, thereby selecting the first wavelength range.

33. The UV-based imaging method of claim 32, wherein the processor is further configured to execute the program instructions to cause the processor to implement the standard wavelength data filter, thereby limiting each first series of first absorbance-wavelength value pairs based on UV absorbance values ​​within a first filter range of UV absorbance values.

34. The UV-based imaging method as described in claim 22 or claim 23, further comprising: An unknown wavelength data filter is provided, and the processor is further configured to execute these program instructions to enable the processor to implement the unknown wavelength data filter for the series of single-wavelength beams, thereby selecting the second wavelength range.

35. The UV-based imaging method of claim 34, wherein the processor is configured to execute the program instructions to cause the processor to implement the unknown wavelength data filter, thereby limiting each second series of second absorbance-wavelength value pairs based on UV absorbance values ​​within a second filter range of UV absorbance values.

36. The UV-based imaging method of claim 22 or claim 23, wherein the processor is further configured to generate enhanced predictions based on the plurality of protein concentration values.

37. The UV-based imaging method of claim 36, wherein generating the enhancement prediction includes averaging each of the plurality of protein concentration values ​​corresponding to the second wavelength range.

38. The UV-based imaging method of claim 22 or claim 23, wherein the multi-wavelength calibration model is used to monitor or measure the protein concentration of protein-related products.

39. The UV-based imaging method of claim 38, wherein the protein concentration is monitored or measured during the manufacturing process of the protein-related product.

40. The UV-based imaging method of claim 38, wherein the protein concentration of the protein-related product is compared with known specifications of the protein-related product.

41. The UV-based imaging method of claim 38, wherein the protein-related product is a therapeutic product.

42. The UV-based imaging method of claim 22 or claim 23, wherein the multi-wavelength calibration model takes into account these second series of second absorbance-wavelength value pairs to predict each plurality of protein concentration values.

43. The UV-based imaging method of claim 22, wherein at least one of the unknown protein samples in the group of unknown protein samples contains a protein therapeutic agent.

44. The UV-based imaging method of claim 43, wherein the protein therapeutic agent is selected from the group consisting of: antibodies, antigen-binding antibody fragments, antibody protein products, bispecific T-cell conjugate molecules, bispecific antibodies, trispecific antibodies, Fc fusion proteins, recombinant proteins, and active fragments of recombinant proteins.

45. A tangible, non-transitory computer-readable medium storing instructions for determining the protein concentration of an unknown protein sample based on automatic multi-wavelength calibration, the instructions, when executed by one or more processors of a computing device, causing the computing device to perform the following operations: At the processor, a set of standard wavelength data is received, as recorded by detector (114) for each of a set of standard protein samples (140s), wherein for each standard protein sample, the set of standard wavelength data includes a first series of first absorbance-wavelength value pairs across a first wavelength range selected from a series of single-wavelength beams from the UV spectrum, each first absorbance-wavelength value pair including a UV absorbance value and a wavelength value. Using this processor, a multi-wavelength calibration model is generated based on each of these first series of first absorbance-wavelength value pairs of the set of standard wavelength data. The processor receives a set of unknown wavelength data recorded by the detector for each of a set of unknown protein samples (140u), wherein for each unknown protein sample, the set of unknown wavelength data includes a second series of second absorbance-wavelength value pairs across a second wavelength range selected from the series of single-wavelength beams, each second absorbance-wavelength value pair including a UV absorbance value and a wavelength value. as well as Using the multi-wavelength calibration model implemented on the processor, multiple protein concentration values ​​are determined for each unknown protein sample in the group of unknown protein samples, each of the multiple protein concentration values ​​corresponding to the second wavelength range selected from the series of single-wavelength beams.

46. ​​An imaging device for determining the protein concentration of an unknown protein sample based on automated multi-wavelength calibration, the imaging device comprising: A means for receiving a set of standard wavelength data for each of a set of standard protein samples (140s), wherein for each standard protein sample, the set of standard wavelength data includes a first series of first absorbance-wavelength value pairs across a first wavelength range selected from a series of single-wavelength beams of UV spectrum, each first absorbance-wavelength value pair including a UV absorbance value and a wavelength value. A device for generating a multi-wavelength calibration model for each of the first series of first absorbance-wavelength value pairs based on the set of standard wavelength data; A means for receiving a set of unknown wavelength data for each of a set of unknown protein samples (140u), wherein for each unknown protein sample, the set of unknown wavelength data includes a second series of second absorbance-wavelength value pairs across a second wavelength range selected from the series of single-wavelength beams, each second absorbance-wavelength value pair including a UV absorbance value and a wavelength value. as well as A device for determining multiple protein concentration values ​​for each unknown protein sample in a set of unknown protein samples using the multi-wavelength calibration model, each of the multiple protein concentration values ​​corresponding to the second wavelength range selected from the series of single-wavelength beams.

Images