Techniques for predicting collision cross section values ​​in ion mobility spectrometry

Abstract

The present invention describes techniques and apparatus for generating predicted collision cross section (CCS) values ​​based on a hybrid CCS prediction process. In one embodiment, for example, the apparatus may include at least one memory and a logic component coupled to the at least one memory. The logic component may be configured to: implement a predictive CCS process, for example, by receiving analytical information from an analysis of a sample using an ion mobility spectrometer, the sample comprising at least one component; generate an approximate molecular model of the component via an approximate molecular modeling process, and generate a predicted CCS value via a computational model based on the approximate molecular model. The present invention describes other embodiments.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims the benefit of and priority to U.S. Provisional Patent Application No. 62 / 834,084, filed on April 15, 2019, which is hereby incorporated by reference in its entirety. Technical Field

[0003] Embodiments herein relate generally to processing analytical data, and more particularly, to predicting collision cross section values ​​based on analytical information derived from ion mobility analysis of a sample. Background Art

[0004] Ion mobility spectrometry or separation (IMS) is a technique widely used to identify and quantify compounds in samples. Generally speaking, the operation of IMS is as follows: by subjecting ions to a weak electric field in the presence of an inert buffer gas (e.g., nitrogen or helium), gas-phase ions are separated based on their shape, size, and charge. When the ions flow through the separation cell or migration cell of the IMS device, the collision frequencies of ions of different sizes and shapes and the inert buffer gas will be different. Under the influence of the electric field, the collision between the ions and the inert buffer gas will produce a drift time difference of the ions (i.e., the drift time of ions with a higher collision frequency will be longer than that of ions with a lower collision frequency). Characteristics associated with IMS include drift time, ion mobility, collision cross section (CCS) or (Ω) and / or reduced CCS (Ω'). The ions separated by IMS can be introduced into a mass analyzer (e.g., a mass spectrometer (MS)) to determine additional physicochemical properties, such as mass-to-charge ratio (m / z).

[0005] Using CCS to identify and confirm compound identity in a sample experiment provides orthogonal compound signatures for retention time and tandem MS information under given experimental conditions. However, conventional methods are not suitable for widespread use due to the lack of CCS measurements in compound libraries and the lack of computational tools that can be used to quickly generate theoretical CCS values. Therefore, analysts who use CCS to assess sample composition would benefit from a more effective and efficient method for obtaining CCS values ​​than conventional systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0006] Figure 1 An embodiment of a first operating environment is shown.

[0007] Figure 2 An embodiment of a second operating environment is shown.

[0008] Figure 3 An implementation of a first logic flow is shown.

[0009] Figure 4The experimental collision cross section (CCS) for various processes is shown in relation to the predicted CCS.

[0010] Figure 5 The error percentages for various CCS prediction procedures are shown.

[0011] Figure 6 CCS prediction results of a CCS prediction process according to some embodiments are shown.

[0012] Figure 7 An embodiment of a computing architecture is shown.

[0013] Figure 8 Several graphs showing how the model drift time from a single case study is related to the empirical CCS (and other parameters) are depicted. DETAILED DESCRIPTION

[0014] Various embodiments as a whole may relate to systems, methods and / or devices for predicting collision cross sections (CCS) or (Ω). In some embodiments, a CCS prediction (hybrid modeling or hybrid CCS prediction) process that combines molecular modeling techniques with machine learning techniques (including but not limited to artificial intelligence processes, neural networks, etc.) can be used to determine predicted CCS or reduced CCS values ​​(pCCS). In various embodiments, the CCS prediction process can include generating one or more pCCS values ​​for various forms of a compound, such as different charge states, conformers, isomers, isobars, etc.

[0015] In some embodiments, analytical information associated with a sample or a portion of a sample may be obtained. In various embodiments, a sample may contain multiple components or a single component, such as a single molecule of interest. The analytical information may include various properties, features, descriptors, elements, etc. of the components. For example, a property may include a description (e.g., a 2D or connectivity description) of a molecule alone or in combination with known and / or theoretical physicochemical properties of the molecule.

[0016] In some embodiments, the characteristic may be theoretical, determined by analysis, or a combination thereof. For example, in an exemplary embodiment, the analytical information may include data obtained by mass analysis (i.e., IMS, MS, IM-MS, CCS analysis (i.e., derived from drift time), etc.) of a sample that can be used to form an approximate molecular model of a compound of interest. In various embodiments, the approximate molecular model may be or may include, for example, a pseudo 3D molecular descriptor that can be provided to a machine learning process to generate a pCCS value. In some embodiments, the approximate molecular model may include a 2D description, a connectivity description, etc. of the molecule. The approximate molecular model may be generated by performing an approximate, rough, abbreviated, and / or similar molecular modeling process to determine a set of conformations of the molecule, the time of which (e.g., at the level of seconds to minutes) is significantly less than the time required to form a complete model using conventional techniques (at the level of hours to days). For example, the approximate molecular modeling process may only perform a limited number of modeling cycles, may only determine a limited number of characteristics, may determine relative values, descriptors, factors, or other distinguishing elements, etc. In some embodiments, the approximate molecular model may include possible energy states, ionization states, or vibration states of the molecule. The approximate molecular model can be provided to a CCS computational model (e.g., a machine learning process, a neural network, an artificial intelligence process, and / or the like) that is used to generate pCCS values ​​based on (2D) molecular descriptors and properties (including the approximate molecular model generated according to some embodiments) to supplement the measured data to establish a predictive model.

[0017] IMS and / or IM-MS can be applied in a variety of fields, including but not limited to drug development (i.e., drug metabolism and pharmacokinetics (DMPK)), food and environment (F&E) research, toxicology, metabolomics and other "omics" research, etc. For example, the use of CCS (e.g., rotation average CCS) to identify and confirm compound identity can provide an orthogonal physicochemical complement to retention time and MS information. At least one limiting factor in conventional systems is the lack of CCS measurements in compound libraries and the lack of computational tools that can be used to quickly generate theoretical CCS values ​​from structures, which are calculated using molecular modeling methods or predicted using tools based on machine learning (i.e., neural networks, artificial intelligence processes, etc.). Therefore, some embodiments provide a hybrid model that combines molecular modeling principles into a machine learning-based approach, which can take advantage of the advantages of both approaches.

[0018] In some embodiments, the approximate molecular model can be used to "train" a CCS calculation or determination algorithm, process, model, etc. In various embodiments, training the CCS calculation algorithm can include experimental information. In an exemplary embodiment, training the CCS calculation algorithm does not require actual experimental information, or can use a combination of experimental information and non-experimental (e.g., library or theoretical information) information. For example, CCS values ​​can be derived from a library of known or previously determined CCS values, from complex ab initio molecular modeling and CCS calculation workflows, etc.

[0019] In various embodiments, the CCS prediction process may include using an approximate molecular model (alone or in combination with other known physicochemical information) to configure a CCS calculation algorithm for a set of first components with known CCS values, and using the CCS calculation algorithm to calculate CCS values ​​for one or more second components that may not have known CCS values. Generally speaking, it may not be necessary to experimentally measure the CCS values ​​of the second components. The calculated CCS values ​​of the second components can be used to optimize or plan possible future experiments.

[0020] In this description, many specific details, such as components and system configurations, may be set forth in order to provide a more thorough understanding of the described embodiments. However, those skilled in the art will appreciate that the described embodiments may be practiced without such specific details. In addition, some well-known structures, elements, and other features are not shown in detail to avoid unnecessarily obscuring the described embodiments.

[0021] In the following description, references to "one embodiment," "an embodiment," "an exemplary embodiment," "various embodiments," etc., indicate that an embodiment of the described technology may include a particular feature, structure, or characteristic, but more than one embodiment may include the particular feature, structure, or characteristic and not every embodiment must include the particular feature, structure, or characteristic. Furthermore, some embodiments may have some, all, or none of the features described for other embodiments.

[0022] As used in this specification and the claims, unless otherwise stated, the use of ordinal adjectives "first", "second", "third", etc. to describe elements merely refers to a specific instance of the referenced element or different instances of similar elements, and is not meant to imply that the elements so described must be in a specific order in time, space, order, or in any other manner.

[0023] Figure 1 An example of an operating environment 100 is shown that may represent some embodiments. Figure 1As shown, mass analysis device 105 (e.g., IM-MS device) may include ion source 130, wherein first ion transfer region 135 is disposed downstream of ion source 130. IMS device 160 may be disposed downstream of ion source 130 and first transfer region 135. IMS device 160 may include IMS trap region 140 and IMS separation region 145. An optional second transfer region 150 may be disposed downstream of IMS device 160.

[0024] In some embodiments, one or more downstream devices or stages 155 may be optionally arranged downstream of the second transfer region 150. The downstream device 155 may include various devices, such as, but not limited to, a mass analyzer, a mass filter, and / or one or more other analytical devices. For example, the downstream device 155 may include a time-of-flight ("ToF") mass analyzer, one or more quadrupole mass filters, one or more ion traps, etc. In some embodiments, an IMS device 160 may be coupled between the ion source 130 and the one or more downstream devices 155. The embodiments are not limited in this context.

[0025] In various embodiments, mass analysis device 105 and / or portions thereof may be operably coupled to computing device 110 (see, e.g., Figure 2 and Figure 7 In some embodiments, computing device 110 may be operable to control various functions of the mass analysis device and / or data processing functions associated with analysis information obtained from mass analysis device 105. For example, computing device 110 may be operable to determine various physicochemical properties of sample components analyzed via the mass analysis device.

[0026] Figure 2 An example of an operating environment 200 is shown that may represent some embodiments. Figure 2 As shown, operating environment 200 may include analysis system 205, which may be used to manage analysis information 232 associated with, for example, an ion mobility spectrometer 270. In some embodiments, ion mobility spectrometer 270 may be or may include an IMS device, an IM-MS device, etc. (see, e.g., Figure 1 In various embodiments, the ion mobility spectrometer 270 may include other specific components, including but not limited to a chromatography system, a liquid chromatography (LC) system, a gas chromatography (GC) system, a mass analyzer system, a mass detector system, a high performance liquid chromatography (HPLC) system, an ultra high performance liquid chromatography (ULHPLC) system, a mass analyzer system, a mass detector system, a mass analyzer ... System, ultra high performance liquid chromatography (UHPLC) system, ultraviolet (UV) detector, visible light detector, solid phase extraction system, sample preparation system, sample introduction system, pump system, capillary electrophoresis instrument, combinations thereof, components thereof, variations thereof, etc.

[0027] In some embodiments, the ion mobility spectrometer 270 is operable to perform an analysis and generate analysis information 232. In various embodiments, the analysis information 232 may include information, data, files, charts, graphs, images, spectra, peak lists, mass values, retention time values, concentration values, compound identification information, etc. generated by the analysis instrument as a result of performing an analysis method. For example, the ion mobility spectrometer 270 may generate mass-charge (m / z) information, drift time (t d ) information, ion migration information, CCS information, etc. in the form of analytical information 232.

[0028] In various embodiments, analytical information 232 may include non-experimental information, including but not limited to theoretical information, library information of known data, etc. The analytical information may include various properties, characteristics, descriptors, elements, etc. of the components. For example, the properties may include descriptions (e.g., 2D or connectivity descriptions) of the molecules alone or in combination with known and / or theoretical physicochemical properties of the molecules. In some embodiments, the properties may be theoretical, determined via analytical analysis, or a combination thereof.

[0029] In various embodiments, the analysis system 205 may include a computing device 210 that is communicatively coupled to an ion mobility spectrometer 270 or otherwise configured to receive and store analysis information 232 associated with the analysis device 215. In some embodiments, the computing device 210 may receive at least a portion of the analysis information 232 from the ion mobility spectrometer 270. In various embodiments, the computing device 210 may receive at least a portion of the analysis information 232 from data sources 254a-n via a network 250. For example, the ion mobility spectrometer 270 may be operable to provide the analysis information 232 directly to the computing device 210 and / or a location on the network 250 accessible to the computing device 210 (e.g., a cloud computing environment).

[0030] In some embodiments, the computing device 210 may be used to control, monitor, manage, or otherwise handle various operational functions of the ion mobility spectrometer 270. In some embodiments, the computing device 210 may be used to provide analytical information 232 to a location on the network 250 via a secure or authenticated connection. In some embodiments, the computing device 210 may be or may include a stand-alone computing device, such as a personal computer (PC), a server, a tablet computing device, a cloud computing device, a mobile computing device (e.g., a smart phone, a tablet computing device, etc.), a data appliance, etc. In various embodiments, the computing device 210 may be or may include a controller or control system integrated into the ion mobility spectrometer 270 to control operational aspects thereof.

[0031] although Figure 2Only one computing device 210 is depicted in the figure, but the embodiments are not limited thereto. In various embodiments, the functions, operations, configurations, data storage functions, applications, logic, etc. described with respect to the computing device 210 may be performed by one or more other computing devices and / or stored in one or more other computing devices. For illustrative purposes only, a single computing device 210 is depicted to simplify the drawings.

[0032] like Figure 2 As shown, computing device 210 may include processor circuit 220, memory unit 230, and transceiver 260. Processor circuit 220 is communicatively coupled to memory unit 230 and / or transceiver 260.

[0033] Processor circuit 220 may include and / or access various logic components for performing processes according to some embodiments. For example, processor circuit 220 may include and / or access compound identification logic component 222 and / or CCS prediction logic component 224. Processing circuit 220 and / or compound identification logic component 222 and / or CCS prediction logic component 224 and / or their parts may be implemented in hardware, software or a combination thereof. As used in this application, the terms "logic component", "component", "layer", "system", "circuit", "decoder", "encoder" and / or "module" are intended to refer to computer-related entities, which may be hardware, a combination of hardware and software, software or software in execution, and their examples are provided by exemplary computing architecture 400. For example, logic, circuits, or layers may be and / or may include, but are not limited to, a process running on a processor, a processor, a hard drive, multiple storage drives (optical and / or magnetic storage media), an object, an executable program, an execution thread, a program, a computer, a hardware circuit, an integrated circuit, an application specific integrated circuit (ASIC), a programmable logic device (PLD), a digital signal processor (DSP), a field programmable gate array (FPGA), a system on a chip (SoC), a memory cell, a logic gate, a register, a semiconductor device, a chip, a microchip, a chipset, a software component, a program, an application, firmware, a software module, computer code, any combination of the foregoing, and the like.

[0034] although Figure 2 220, the compound identification logic 222 is depicted as being within the processor circuit 220, but the embodiments are not limited thereto. In addition, although the CCS prediction logic 224 is depicted as being a logic component of the processor circuit 220, the embodiments are not limited thereto, as the data processing logic 224 can be an independent logic component. For example, the compound identification logic 222 and / or any of its components can be located within an accelerator, a processor core, an interface, a single processor die, fully implemented as a software application (e.g., the compound identification application 240), etc.

[0035] The memory unit 230 may include various types of computer-readable storage media and / or systems in the form of one or more higher-speed memory units, such as read-only memory (ROM), random access memory (RAM), dynamic RAM (DRAM), double data rate DRAM (DDRAM), synchronous DRAM (SDRAM), static RAM (SRAM), programmable ROM (PROM), erasable programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, polymer memory (such as ferroelectric polymer memory, bidirectional memory, phase change or ferroelectric memory), silicon-oxide-nitride-oxide-silicon (SONOS) memory, magnetic card or optical card, device array (such as independent disk redundant array (RAID) drive), solid-state memory device (e.g., USB memory, solid-state drive (SSD), and any other type of storage medium suitable for storing information. In addition, the memory unit 230 may include various types of computer-readable storage media in the form of one or more lower-speed memory units, including internal (or external) hard disk drives (HDD), magnetic floppy disk drives (FDD), optical disk drives for reading or writing removable optical disks (e.g., CD-ROM or DVD), solid-state drives (SSD), etc.

[0036] The memory unit 230 may store a compound identification application 240, which may be operated alone or in combination with the compound identification logic component 222 to perform various analytical services according to some embodiments. For example, the compound identification application 240 may be operated to identify compounds in a sample. In another example, according to some embodiments, the compound identification application 240 may be operated to perform a hybrid CCS prediction process, for example, via the CCS prediction logic component 224 to generate a pCCS value. In an exemplary embodiment, the compound identification application 240 may generate a pCCS value as part of a process for identifying compounds in a sample. In various embodiments, identifying a compound may include identifying a "known compound" (i.e., a known compound that is expected to be part of a sample), a "known unknown compound" (i.e., a known compound that is not expected to be part of a sample) and / or an "unknown unknown compound" (i.e., an unknown compound within a sample).

[0037] In various embodiments, compound identification application 240 may use computational model information 234 to identify compounds and / or generate pCCS information (see, e.g., Figure 3). In some embodiments, the computational model information 234 may include molecular modeling information, processes, etc. for modeling molecules based on the analytical information. A non-limiting example of a modeling process may include a structure calculation step, followed by a CCS calculation step for one or more of the structures identified thereby. The structure calculation step may include generating an initial 3D structure based on 2D coordinates, identifying a set of one or more possible conformers (e.g., about five conformers) from these initial 3D structures, and obtaining an optimized structure for each conformer. The step of calculating CCS values ​​based on the structure may include one or more of the trajectory method, the exact hard sphere scattering method, the projection approximation, the electron density isosurface scattering (SEDI), the projection superposition approximation (PSA), and many other CCS calculation methods. The step of generating an optimized structure may include energy minimization. Examples of software that can be used to calculate structures and perform energy minimization may include Avogadro, Gaussian, GAMESS, etc. Examples of software that can be used to calculate CCS values ​​for a given structure include MobCal, IMOS, CCScalc, etc. Embodiments are not limited to these exemplary modeling processes, as any molecular modeling process capable of operating in accordance with some embodiments is contemplated herein.

[0038] In various embodiments, the molecular modeling process may be configured to model at least a portion of the analyzed molecule or other compound. In some embodiments, the molecular model may include various factors, elements, descriptors (e.g., pseudo 3D geometry or relative 3D geometry) associated with the molecule, etc. Non-limiting examples of descriptors may include charge, size, shape, etc. In various embodiments, the descriptor may include all or substantially all descriptors associated with a specific molecular modeling process. In an exemplary embodiment, the descriptor may include a subset of descriptors associated with a specific modeling process. In some embodiments, a subset of descriptors may be selected based on the contribution of the subset of descriptors to determining CCS (or pCCS).

[0039] In an exemplary embodiment, the computational model information 234 may include various machine learning processes, algorithms, and the like. In some embodiments, the machine learning process may be trained using training data (such as actual CCS values ​​of known molecules determined experimentally). Non-limiting examples of machine learning processes may include MetCCS (see, e.g., Zhou et al., "MetCCSPredictor: a Web Server for Predicting Collision Cross-Section Values ​​of Metabolite in Metabolomics", Bioinformatics, 2017, 33, 2235-2237), DeepCCS (see, e.g., Plante et al., "Predicting Ion Mobility Collision Cross-Sections Using a Deep Neural Network: DeepCCS", Analytical Chemistry (April 2019). In some embodiments, the machine learning process may be or may include a cross-validation (CV) process. The embodiments are not limited to these exemplary machine learning processes, as any machine learning process that can operate according to some embodiments is contemplated herein.

[0040] In some embodiments, the compound identification application 240 may generate analytical information 232 in the form of experimental information. In various embodiments, the experimental information may include compound information of known compounds, such as drift time, CCS, m / z, etc. The experimental information of known compounds may be used as part of a modeling process and / or a machine learning process according to some embodiments to determine specific features of unknown compounds. For example, experimental CCS data may be obtained from calibrated traveling wave based IMS measurements obtained using IMS-Q-oaToF and Q-IMS-oaToF configurations. Generally speaking, the experimental data may represent an average of at least three technical replicates. In summary, here, according to some embodiments, 4100 CCS values ​​from nearly 3000 different compounds representing various adduct forms and ten chemical classes of compounds can be used for training, validation and / or testing of the model.

[0041] This article includes one or more logic flows representing exemplary methods for performing novel aspects of the disclosed architecture. Although one or more methods shown herein are shown and described as a series of actions for the purpose of simplifying the description, it will be understood and appreciated by those skilled in the art that these methods are not limited by the order of actions. Therefore, some actions may occur in different orders and / or occur simultaneously with other actions shown and described herein. For example, it will be understood and appreciated by those skilled in the art that the method may alternatively be represented as a series of related states or events, such as in a state diagram. In addition, not all behaviors shown in the method may be necessary for the novel implementation. The box designated by the dotted line may be an optional box of the logic flow.

[0042] The logic flow may be implemented in software, firmware, hardware, or any combination thereof. In software and firmware implementations, the logic flow may be implemented by computer executable instructions stored on a non-transitory computer readable medium or machine readable medium. The implementation is not limited in this context.

[0043] Figure 3 An embodiment of a logic flow 300 is shown. The logic flow 300 may represent some or all of the operations performed by one or more embodiments described herein, such as by computing device 110 and / or device 210. For example, the logic flow 300 may represent some or all of the operations of generating a pCCS value according to some embodiments.

[0044] At block 302 , the logic flow 300 can receive analytical data. For example, the analytical information 232 obtained by analyzing a sample via the ion mobility spectrometer 270 can be accessed by the compound identification application 240 .

[0045] The logical flow 300 may perform approximate molecular modeling at block 304. For example, at least a portion of the analytical information (e.g., drift time, m / z, ion mobility, etc.) may be provided to a molecular modeling process to perform an approximate molecular modeling process. A full molecular modeling process using known techniques requires a large amount of computational resources and time (e.g., in the range of hours to days) to generate a molecular model. Thus, some embodiments may perform an approximate molecular modeling process where only a subset of descriptors, loops, etc. are determined or executed. For example, a conventional molecular modeling process may require X number of loops or optimization energy steps; however, some embodiments may only perform N loops, where N < X (or even N << X). For example, X may be greater than 100 and N may be less than 100. In some embodiments, N may be 1, 2, 3, 4, 5, 10, 15, 20, 50, 100, 200, 250, 500, 1000, 5000, and / or any value or range between any two of these values (including the end values). The approximate molecular modeling process may generate an approximate molecular model, representing a rough model of the molecule. In various embodiments, the approximate molecular model may include possible forms of the molecule, including charge states, conformational isomer states, and / or similar forms. Compared to a full model, the approximate molecular model can be generated in a shorter time, e.g., in the range of seconds or minutes. For example, the approximate molecular model may be generated in about 100 milliseconds (ms), about 1 second (s), about 5 s, about 10 s, about 30 s, about 1 minute (m), about 5 m, about 10 m, about 30 m, and / or any value or range between any two of these values (including the end values).

[0046] In some embodiments, a structure (or approximate structure or relative structure) may be determined based on the analytical information. The number of possible conformational isomers that may be generated may depend at least in part on the structure of the compound. Additionally, in some embodiments, the structure may determine the time taken to optimize the energy / conformation of each conformational isomer (e.g., 10 s for a rigid small molecule, up to 5 to 6 minutes for a more flexible / larger molecule, etc.).

[0047] In some embodiments, the approximate molecular modeling process may generate multiple molecular models, e.g., each molecular model having a different charge state or configuration. For example, the CCS of an ion may be related to the shape, size, and / or charge state of the ion. Ions of the same molecule may have different charge states. Thus, multiple approximate molecular models of the molecule may be generated, e.g., multiple models of different models with different charge states. In some embodiments, the approximate molecular model may be or may provide (pseudo) molecular descriptors associated with the sample component. In various embodiments, a machine learning process according to some embodiments may use these pseudo molecular descriptors to generate pCCS values.

[0048] The logic flow 300 may provide an approximate molecular model to the CCS computational model at box 306. For example, the computational model information 234 may include various machine learning processes for determining a molecular model based on pCCS. The compound identification application 240 may provide an approximate molecular model to the machine learning process of the computational model information 234. At box 308, the logic flow may generate pCCS. For example, the machine learning process may generate pCCS information 236 including at least one pCCS value. In various embodiments, multiple pCCS values ​​may be generated for a molecule, an approximate molecular model, etc. For example, multiple pCCS values ​​may be generated for different conformations, charge states, etc. of a molecule. In some embodiments, a rough model may be used to more effectively predict CCS and have improved accuracy compared to conventional processes. In various embodiments, a CCS prediction process (e.g., via a machine learning or artificial intelligence process) may be operated to generate pCCS values ​​based on pseudo-molecular descriptors of an approximate molecular model generated according to some embodiments.

[0049] At block 310, the logic flow 310 may determine compound information. For example, the compound identification application 240 may use the pCCS information 236 to determine or estimate the identity of the molecule of interest based on the pCCS value.

[0050] Example: Hybrid CCS Prediction Process

[0051] A hybrid CCS prediction process is performed according to some embodiments. Experimental CCS data are obtained from traveling wave based IMS measurements obtained using IMS-Q-oaToF and Q-IMS-oaToF geometries. Generally speaking, the experimental data may represent the average of at least three technical replicates. In summary, 4100 CCS values ​​from nearly 3000 different compounds representing various adduct forms of compounds and ten chemical categories are used for model training, validation and testing. For molecular modeling, Avogadro (oBabel) and CCScalc (included in DriftScope) are applied to provide structural / geometric information in various constant forms, respectively, as well as projection approximate calculation results, and a gradient boosting algorithm to train a prediction model with features, which include relevant molecular descriptors and structural features obtained using molecular modeling methods (e.g., via an approximate molecular modeling process).

[0052] A cross-validation (CV) strategy, external data, and multiple use cases were used to build and evaluate a hybrid CCS prediction model using experimental CCS data for singly charged compounds. CV results showed significant improvements over a baseline model based only on molecular weight, with the relative mean absolute error of the baseline model improving from 4.6% to 1.8% in a non-hybrid (machine learning only) model, and were observed for all molecules and chemical classes, even more complex chemical superclasses such as benzene compounds and organic heterocyclic compounds that are more difficult to predict.

[0053] The process results were also compared with different conventional machine learning methods and showed an overall improved performance. In some embodiments, it is not necessary to train several models; for example, a single model according to some embodiments can produce high-quality results as long as a sufficiently diverse training data set is used.

[0054] In the final evaluation, according to some embodiments, the predictive model was applied to the problem of CCS prediction of positional isomers and isobaric molecules. In most of these cases, the model was able to correctly predict the rank order of analyte CCS values, demonstrating that the predicted CCS values ​​can enhance downstream analysis. To further demonstrate the impact on downstream analysis, an evaluation metric that reflects the model's ability to discriminate isobaric analytes was used. The metric assigns the measured CCS library data to isobaric bins, and for each compound in the bin, specifies the closest observed CCS value and predicted CCS value and compares these values ​​with the CCS value of the actual compound. In most cases, the closest predicted CCS and observed CCS pairs were found to be the actual compounds of interest, indicating that combining predicted CCS values ​​can improve the discrimination of isobaric compounds.

[0055] Figure 4 and Figure 5 Depicted is a comparison of predicted CCS results for certain conventional processes compared to a hybrid CCS prediction process according to some embodiments. Figure 4The relationship between the experimental CCS and the predicted CCS of various processes is depicted. Specifically, graph 410 depicts the standard CCS value, graph 415 depicts the first conventional CCS prediction process, graph 420 depicts the second conventional CCS prediction process, and graph 425 depicts the hybrid CCS prediction process according to some embodiments. The first conventional CCS prediction process may include a process as described by Bijlsma, Lubertus et al. "Prediction of Collision Cross-Section Values ​​for Small Molecules:Application to Pesticide Residue Analysis." Analytical chemistry 89 12 (2017): 6583-6589 ("Bijlsma"). The second conventional CCS process may include MetCCS.

[0056] Figure 5 A graph 505 is depicted showing the percentage of compounds that fall outside of a 2% error tolerance for CCS predictions of a hybrid CCS prediction process of Bijlsma 520 and MetCCS according to some embodiments 525. A graph 510 depicts information of the percentage error of experimental CCS to Bijlsma 540, the percentage error of experimental CCS to MetCCS 545, and the percentage error of experimental CCS to the hybrid CCS prediction process according to some embodiments 550.

[0057] like Figure 4 and Figure 5 As shown, the hybrid CCS prediction process produces more accurate predictions compared to conventional processes (including Bijlsma and MetCCS).

[0058] Figure 6 Depicted are CCS prediction results for tinic acid metabolites using a hybrid CCS prediction process according to some embodiments. Figure 6 As shown, the percentage difference between pCCS and experimental CCS (eCCS) is less than about 2%.

[0059] Figures 4 to 6 The generated results are determined using a machine learning process without an approximate modeling process according to some embodiments.

[0060] Figure 8 Another example is depicted where the "Van der Waals" radius is used to derive 3D properties. These 3D properties can then be used by an AI-based algorithm for predicting CCS values. In one example, the model drift time can be calculated based on the following logic:

[0061] Model drift = -40.0 + 2.0 * cross-sectional area / z

[0062] / / Collision cross-sectional area (Ω) is

[0063] / / The function of the total van der Waals volume is as follows:

[0064] / / W=pi*(3 / (4*pi))^(2 / 3)*V(2 / 3)=1.209*V^(2 / 3)

[0065] Ω=1.209*Math.Pow(Volume,(2.0 / 3.0));

[0066] / / "Volume" represents the sum of the "residual volume" of each amino acid

[0067] / / It is the van der Waals volume, in units of

[0068] / / Cubic Angstrom

[0069] Figure 8 Plotted is how the model drift time (peptide.modeldrift) relates to the empirical CCS and other parameters.

[0070] Figure 7 An embodiment of an exemplary computing architecture 700 that may be suitable for implementing various embodiments as previously described is shown. In various embodiments, the computing architecture 700 may include or be implemented as part of an electronic device. In some embodiments, the computing architecture 700 may represent, for example, a distributed processing system that implements or utilizes one or more components described herein. In some embodiments, the computing architecture 700 may represent, for example, a computing node in a distributed processing system described herein that implements or utilizes one or more techniques described herein. The embodiments are not limited to this context.

[0071] As used in this application, the terms "system" and "component" and "module" may be intended to refer to a computer-related entity, which may be hardware, a combination of hardware and software, software, or software in execution, examples of which may be provided by the exemplary computing architecture 700. For example, a component may be, but may not be limited to, a process running on a processor, a processor, a hard drive, multiple storage drives (optical and / or magnetic storage media), an object, an executable file, an execution thread, a program, and / or a computer. For example, both an application running on a server and a server may be a component. One or more components may reside within a process and / or an execution thread, and a component may be located on a computer and / or distributed between two or more computers. In addition, components may be communicatively coupled to each other through various types of communication media to coordinate operations. Coordination may involve unidirectional or bidirectional information exchange. For example, a component may convey information in the form of a signal conveyed through a communication medium. The information may be implemented as a signal assigned to various signal lines. In such an assignment, each message may be a signal. However, another embodiment may alternatively employ a data message. Such data messages may be sent through various connections. Exemplary connections include parallel interfaces, serial interfaces, and bus interfaces.

[0072] The computing architecture 700 includes various general computing elements, such as one or more processors, multi-core processors, co-processors, memory units, chipsets, controllers, peripherals, interfaces, oscillators, timing devices, video cards, audio cards, multimedia input / output (I / O) components, power supplies, etc. However, embodiments may not be limited to the implementation of the computing architecture 700.

[0073] like Figure 7 As shown, computing architecture 700 includes a processing unit 704, a system memory 706, and a system bus 708. Processing unit 704 can be any of various commercially available processors, including but not limited to: and processor; Application, embedded and security processors; and and Processors; IBM and Cell processor; Core(2) and processor; and similar processors. Dual microprocessors, multi-core processors, and other multi-processor architectures may also be used as the processing unit 704.

[0074] The system bus 708 provides an interface for system components, including but not limited to an interface connecting the system memory 706 to the processing unit 704. The system bus 708 can be any of several types of bus structures, which can be further interconnected to a memory bus (with or without a memory controller), a peripheral bus, and a local bus using any of a variety of commercially available bus architectures. Interface adapters can be connected to the system bus 708 via a slot architecture. Exemplary slot architectures can include but are not limited to accelerated graphics port (AGP), card bus, (extended) industry standard architecture ((E)ISA), micro channel architecture (MCA), NuBus, peripheral component interconnect (extended) (PCI (X)), PCI Express, personal computer memory card international association (PCMCIA), etc.

[0075] The system memory 706 may include various types of computer-readable storage media in the form of one or more higher-speed memory cells, such as read-only memory (ROM), random access memory (RAM), dynamic RAM (DRAM), double data rate DRAM (DDRAM), synchronous DRAM (SDRAM), static RAM (SRAM), programmable ROM (PROM), erasable programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory (i.e., one or more flash memory arrays), polymer memory (such as ferroelectric polymer memory, bidirectional memory, phase change or ferroelectric memory), silicon-oxide-nitride-oxide-silicon (SONOS) memory, magnetic or optical cards, device arrays (such as redundant array of independent disks (RAID) drives), solid-state memory devices (e.g., USB memory, solid-state drives (SSDs), and any other type of storage medium suitable for storing information. Figure 7 In the illustrated embodiment, system memory 706 may include non-volatile memory 710 and / or volatile memory 712. In some embodiments, system memory 706 may include main memory. A basic input / output system (BIOS) may be stored in non-volatile memory 710.

[0076] The computer 702 may include various types of computer-readable storage media in the form of one or more slower memory units, including an internal (or external) hard disk drive (HDD) 714, a magnetic floppy disk drive (FDD) 716 for reading or writing to a removable magnetic disk 718, and an optical drive 720 (e.g., a CD-ROM or DVD) for reading or writing to a removable optical disk 722. The HDD 714, FDD 716, and optical drive 720 may be connected to the system bus 708 via a HDD interface 724, a FDD interface 726, and an optical drive interface 728, respectively. The HDD interface 724 for external drive implementations may include at least one or both of a universal serial bus (USB) and an Institute of Electrical and Electronics Engineers (IEEE) 994 interface technology. In various embodiments, these types of memory may not be included in the main memory or system memory.

[0077] The drives and associated computer-readable media provide volatile and / or nonvolatile storage of data, data structures, computer-executable instructions, etc. For example, a number of program modules may be stored in the drives and memory units 710, 712, including an operating system 730, one or more application programs 732, other program modules 734, and program data 736. In one embodiment, the one or more application programs 732, other program modules 734, and program data 736 may include, for example, various applications and / or components of the message controller 104.

[0078] A user may enter commands and information into the computer 702 through one or more wired / wireless input devices, such as a keyboard 738 and a pointing device such as a mouse 740. Other input devices may include a microphone, an infrared (IR) remote control, a radio frequency (RF) remote control, a game controller, a stylus, a card reader, a dongle, a fingerprint reader, a glove, a graphics tablet, a joystick, a keyboard, a retina reader, a touch screen (e.g., capacitive, resistive, etc.), a trackball, a trackpad, a sensor, a stylus, etc. These and other input devices may typically be connected to the processing unit 704 through an input device interface 742 that may be coupled to the system bus 708, but may be connected through other interfaces such as a parallel port, an IEEE 994 serial port, a game port, a USB port, an IR port, etc.

[0079] A monitor 744 or other type of display device may also be connected to the system bus 708 via an interface, such as a video adapter 746. The monitor 744 may be internal or external to the computer 702. In addition to the monitor 744, computers typically include other peripheral output devices, such as speakers, printers, and the like.

[0080] Computer 702 can be operated in a networked environment via wired and / or wireless communication to one or more remote computers such as remote computer 748 using logical connections. In various embodiments, one or more migrations can be performed via a networked environment. Remote computer 748 can be a workstation, server computer, router, personal computer, portable computer, microprocessor-based entertainment facilities, peer-to-peer device or other public network node, and generally includes many or all elements in the elements described relative to computer 702, but for the sake of brevity, only memory / storage device 750 can be shown. Depicted logical connection includes wired / wireless connection to local area network (LAN) 752 and / or larger network such as wide area network (WAN) 754. Such LAN and WAN network environments may be common in offices and companies, and facilitate enterprise-wide computer networks, such as intranets, which can all be connected to global communication networks, such as the Internet.

[0081] When used in a LAN networking environment, the computer 702 can be connected to the LAN 752 through a wired and / or wireless communication network interface or adapter 756. The adapter 756 can facilitate wired and / or wireless communication to the LAN 752, which can also include a wireless access point provided thereon for communicating with the wireless functionality of the adapter 756.

[0082] When used in a WAN networking environment, the computer 702 may include a modem 758, or may be connected to a communications server on the WAN 754, or have other means for establishing communications over the WAN 754, such as over the Internet. The modem 758 may be internal or external and a wired and / or wireless device that is connected to the system bus 708 via the input device interface 742. In a networked environment, program modules depicted relative to the computer 702, or portions thereof, may be stored in the remote memory / storage device 750. It will be appreciated that the network connections shown may be exemplary and other means of establishing a communications link between the computers may be used.

[0083] The computer 702 is operable to communicate with wired and wireless devices or entities using the IEEE 802 family of standards, such as wireless devices operatively arranged in wireless communications (e.g., IEEE 802.16 over-the-air modulation techniques). This includes at least Wi-Fi (or Wireless Fidelity), WiMax, and Bluetooth. TMWireless technology, etc. Therefore, the communication can be a predefined structure like a conventional network, or just an ad hoc communication between at least two devices. Wi-Fi networks use radio technologies called IEEE 802.11x (a, b, g, n, etc.) to provide secure, reliable, and fast wireless connectivity. Wi-Fi networks can be used to connect computers to each other, to the Internet, and to wired networks (using IEEE 802.3 related media and functions).

[0084] Many specific details have been set forth herein to provide a thorough understanding of the embodiments. However, those skilled in the art will appreciate that the embodiments may be practiced without these specific details. In other cases, well-known operations, components, and circuits are not described in detail to avoid obscuring the embodiments. It is understood that the specific structural and functional details disclosed herein may be representative and do not necessarily limit the scope of the embodiments.

[0085] The expressions "coupled" and "connected," as well as their derivatives, may be used to describe some embodiments. These terms are not intended to be synonymous with each other. For example, the terms "connected" and / or "coupled" may be used to describe some embodiments to indicate that two or more elements are in direct physical or electrical contact with each other. However, the term "coupled" may also mean that two or more elements are not in direct contact with each other, but still cooperate or interact with each other.

[0086] Unless expressly stated otherwise, it is understood that terms such as "processing", "computing", "calculating", "determining", etc. refer to the operation and / or process of a computer or computing system or similar electronic computing device, which processes and / or converts data represented as physical quantities (e.g., electronic) within the registers and / or memories of the computing system into physical quantities similarly represented as memories, registers or other such information storage devices, transmission or display devices of the computing system. The embodiments are not limited to this context.

[0087] It should be noted that the methods described herein do not have to be performed in the order described or in any particular order. In addition, the various activities described with respect to the methods identified herein can be performed in series or in parallel.

[0088] Although specific embodiments have been depicted and described herein, it should be understood that any arrangement that is expected to achieve the same purpose can replace the specific embodiments shown. The present disclosure is intended to cover any and all adaptations or changes of various embodiments. It should be understood that the above description is carried out in an illustrative manner, rather than restrictive. After reading the above description, the combination of the above-mentioned embodiments and other embodiments not specifically described herein will be apparent to those skilled in the art. Therefore, the scope of the various embodiments includes any other application in which the above-mentioned compositions, structures and methods are used.

[0089] Although the subject matter has been described in language specific to structural features and / or methodological acts, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts are disclosed as example forms of implementing the claims.

Claims

1. A device, comprising: at least one memory; and a logic component, the logic component coupled to the at least one memory, the logic component for implementing a predicted collision cross section (CCS) process, the logic component being configured to: receiving analytical information associated with a sample comprising at least one component, generating one or more approximate molecular models for the components via an approximate molecular modeling process, the approximate molecular models identifying the three-dimensional structures of the predicted components, applying a computational model configured to apply machine learning to the predicted three-dimensional structure identified by the approximate molecular model, and The predicted CCS values ​​are generated via a computational model.

2. The apparatus of claim 1, wherein the at least one ion mobility spectrometer comprises one of an ion mobility spectrometer (IMS) or an ion mobility mass spectrometer (IM-MS). 3 . The apparatus according to claim 1 , wherein the analysis information comprises at least one of drift time information, ion mobility information, or mass-to-charge ratio (m / z) information. The apparatus according to claim 1 , wherein the approximate molecular model comprises at least one pseudo-molecular descriptor.

5. The apparatus of claim 1, wherein the approximate molecular modeling process comprises executing a subset of a cycle of a complete molecular modeling process. The apparatus of claim 5 , wherein the subset of cycles is determined based on the structure of the component.

7. The apparatus of claim 1, wherein the computational model comprises a machine learning process.

8. The apparatus of claim 1, the approximate molecular model comprising a plurality of possible charge states of the modeled molecule.

9. The apparatus of claim 1, wherein the predicted CCS value is within 2% or less of the corresponding experimental CCS value.

10. The apparatus of claim 1, wherein the approximate molecular model is generated in a duration of less than 5 minutes.

11. A method for performing a predicted collision cross section (CCS) process, the method comprising: receiving analytical information associated with a sample comprising at least one component; generating an approximate molecular model of the component via an approximate molecular modeling process, the approximate molecular model identifying the predicted three-dimensional structure of the component, applying a computational model configured to apply machine learning to the predicted three-dimensional structure identified by the approximate molecular model, and The predicted CCS values ​​are generated via a computational model.

12. The method of claim 11, wherein the at least one ion mobility spectrometer comprises one of an ion mobility spectrometer (IMS) or an ion mobility mass spectrometer (IM-MS).

13. The method of claim 11, the analytical information comprising at least one of drift time information, ion mobility information, or mass-to-charge ratio (m / z) information. The method of claim 11 , wherein the approximate molecular model comprises at least one pseudo 3D molecular descriptor.

15. The method of claim 11, the approximate molecular modeling process comprising performing a subset of cycles of a full molecular modeling process.

16. The method of claim 15, wherein the subset of cycles is determined based on the structure of the component.

17. The method of claim 11, wherein the computational model comprises a machine learning process.

18. The method of claim 11, the approximate molecular model comprising a plurality of possible charge states of the modeled molecule.

19. The method of claim 11, wherein the predicted CCS value is within 2% or less of the corresponding experimental CCS value.

20. The method of claim 11, wherein the approximate molecular model is generated in a duration of less than 5 minutes.

Images