Systems and methods for targeted analysis of DIA-pasef acquisition methods

Abstract

The present invention determines how to extract data of peptide precursors and fragments thereof in an efficient manner when processing a diagonal series of so-called dia-PASEF acquisition methods, such as synchro-PASEF or midia-PASEF. In one implementation of the invention, the entire ion cloud is considered to be acquired through a single diagonal DIA window.

Description

System and method for targeted analysis of diagonal series acquired by the dia-PASEF acquisition method Technical Field

[0001] This section covers compound analysis using mass spectrometry, and more specifically, instruments and methods for peptide analysis. Background Technology

[0002] Data-independent acquisition (DIA) targeting analysis is a powerful mass spectrometry approach for comprehensive, reproducible, and precise proteomics quantification. It is a variant of targeted proteomics where the targeting aspect is introduced only at the data analysis level. Unlike multiple reaction monitoring (MRM), this method requires no preliminary method design prior to sample injection. Because LC-MS acquisition covers the full analyte content of the sample across the entire mass and retention time (RT) range, posterior mining of data can be performed on any peptide / precursor of interest. Data is acquired in a data-independent manner across the entire mass range (e.g., 200 to 2000 Thomson) and throughout the entire chromatography, regardless of sample content. This is typically achieved by progressively stepping the selection window of the mass analyzer across the entire mass range. In effect, this data acquisition method generates a complete fragment ion map for all analytes present in the sample and again correlates the fragment ion spectra with the precursor ion selection window where the fragment ion spectra are acquired. This is achieved by widening the precursor isolation window on the mass analyzer and thus pre-accounting for multiple precursors that co-elute and jointly participate in the fragmentation patterns recorded during analysis. Such a precursor window is called a precursor selection window. This ultimately yields complex fragment ion spectra from multiple precursor fragments, requiring more challenging data analysis.

[0003] DIA requires a spectral library containing peptide precursor ions for targeted extraction from DIA data. This targeted extraction can be performed across multiple dimensions, such as retention time, ion mobility, etc. Existing methods cannot rapidly or efficiently analyze multidimensional data to generate quantifications of all data (for as many peptides as possible). One way to achieve targeted extraction is to use the full range of available data when searching for a given peptide precursor ion. However, this reduces the overall quality of the analysis due to increased interference in the data. Furthermore, it leads to a significant increase in analysis time due to the need to process more data. Another way to achieve targeted extraction is to use only the retention time dimension. However, this does not consider the ion mobility dimension. Therefore, systems and methods are needed that can achieve targeted extraction in addition to the retention time dimension, also considering the ion mobility dimension.

[0004] US 7,838,826 B1 (MA Park, 2008) and its corresponding patent family disclose a small ion mobility analyzer / spectrometer, which is known by the acronym "TIMS" analyzer / spectrometer (TIMS = Trapped Ion Mobility Spectrometer). The terms ion mobility analyzer and ion mobility spectrometer are used interchangeably herein. A TIMS analyzer includes a gas flow that drives ions against a counteracting electric field barrier, such that the ions are initially trapped along the axis of the TIMS analyzer. The ions are confined in the radial direction by an RF electric field. After the ions have been transferred from the ion source to the electric field barrier, the height of the electric field barrier or the gas velocity is adjusted so that the ions are released from the electric field barrier in order of their mobility.

[0005] Data acquisition independent of ion mobility has recently been enhanced with the development of timsTOF Pro. TM It became popular after its release.

[0006] EP-A-3054473 discloses the operation of such a captured ion mobility spectrometer, which is based on the DC electric field barrier against which ions are propelled by a gas flow, preferably in combination with a mass analyzer as an ion detector. The present invention places an additional RF ion trap upstream of the captured ion mobility spectrometer, wherein the RF ion trap operates as an accumulation unit in parallel with the captured ion mobility spectrometer, allowing a first set of ions to be analyzed in the captured ion mobility spectrometer while a second set of ions from the ion source is simultaneously collected in the accumulation unit. Due to its parallel accumulation-serial fragmentation (PASEF) strategy, it achieves a high duty cycle and shows promise for biological applications. However, early acquisition methods proposed for this type of instrument were inefficient in ion cloud sampling (Figure 1).

[0007] Recently, researchers have proposed new DIA acquisition methods that sample ion clouds more effectively; namely, synchro-PASEF (Skowronek P. et al., “Synchro-PASEF allows precursor-specific fragmention extraction and interference removal in data-independent acquisition”; Mol.Cel. Prot. 2022; 22(2)) and midia-PASEF (Distler U. et al., “midia-PASEF maximizes information content in data-independent acquisition proteomics”; bioRxiv.2023). The latter is distinguished from the former by using an overlapping mass (DIA) window and is described in US-A-2022034840, disclosing an apparatus and method for data-independent combined ion mobility and mass spectrometry analysis, including introducing precursor ions into an ion mobility spectrometer (IMS), continuously releasing precursor ions from the IMS according to the ion mobility of the precursor ions, introducing the released precursor ions into a mass filter, fragmenting the precursor ions transported through the mass filter to generate fragment ions, and performing mass spectrometry measurements on the fragment ions. The IMS and mass filter are controlled synchronously to perform multiple IM scans, wherein adjacent mass windows associated with continuous mass spectrometry measurements of fragment ions overlap during the IM scans, such that during the IM scans, precursor ions transported through the mass filter are located in at least one continuous scan region in the m / z-IM plane, the continuous scan region extending in a generally diagonal direction in the m / z-IM plane.

[0008] Both synchro-PASEF and midia-PASEF methods work by seamlessly and continuously following the natural shape of the ion cloud in both the ion mobility or collision cross section (CCS) and peptide precursor mass dimensions (Figure 2a). In this process, the DIA window takes the shape of a diagonal slice rather than a vertical box covering the top of the ion cloud. The immediate benefit of these types of acquisition methods is improved ion cloud sampling, as it is more efficient than a vertical box.

[0009] However, analyzing such data remains an open challenge and is often quite complex. Each diagonal window slice consists of hundreds of individual scans with continuously varying m / z and ion mobility ranges. To date, research has focused on deconvolution of such DIA data to create data-dependent acquisition (DDA) type data with relatively non-chimeric MS2 spectra due to narrow isolation widths. This is achieved by utilizing the concept of so-called precursor slices, where knowledge of the expected slices of the precursor and their relationship to fragments is used to assign fragments belonging to a specific peptide precursor. This results in cleaner DDA-type MS2 spectra.

[0010] However, these methods are computationally slow and assume that DIA data can be cleaned to resemble DDA data without significant information loss. Additionally, if the method relies on MS1 ​​for the deconvolution step, it may be affected for peptides lacking high-quality MS1 signals. In such methods, by design, most peptide precursor ions are expected to be sliced ​​by two or more diagonal DIA windows. This complicates data analysis but also provides novel fractions utilizing this acquisition scheme. Summary of the Invention

[0011] While current efforts are focused on deconvolving DIA data to make it resemble narrow-window DDA data, a potential approach is to analyze such data in a way that adapts to its complexity.

[0012] This invention identifies how to efficiently extract data on peptide precursors and their fragments when processing diagonal series acquisition methods such as synchro-PASEF or midia-PASEF, which are based on so-called dia-PASEF. In one implementation of this invention, the entire ion cloud is considered to be acquired through a single diagonal DIA window (Fig. 2b). In one method of this implementation, data is extracted for precursor ions measured by multiple diagonal DIA windows. A novel method is then used that enables the processing of this type of DIA data in a standard targeted manner. Additionally, a new category of fractions utilizing the slicing phenomenon can be used. In one implementation of this method, the invention enables seamless querying of mass spectrometry data in a manner that allows for efficient extraction of both complete (across all diagonal DIA windows, Fig. 3a, Fig. 3b) and partial (specific diagonal DIA window, Fig. 3c, Fig. 3d) ion traces of fragment ions and / or precursor ions in both retention time (RT) (Fig. 4d, Fig. 4e) and ion mobility (IM) dimensions (Fig. 4f, Fig. 4g). One advantage of this invention is its significantly faster speed than existing solutions, and because it queries data in all possible ways across both the RT and IM dimensions, it should achieve a higher degree of identification. Furthermore, it does not rely on the presence of a high-quality MS1 signal for deconvolution. Those skilled in the art will understand that, in addition to proteomics, this invention can be applied to other mass spectrometry-based omics data, such as metabolomics. Furthermore, those skilled in the art will understand that, besides synchronous-PASEF or midia-PASEF, this invention can be used with other similar methods. For example, the shape of the diagonal slice does not need to be described by a linear function, but can be described by a nonlinear function such as a polynomial function. Additionally, this method can also be used to analyze basic diagonal acquisition schemes, such as slice-PASEF (Szyrwiel L. et al., “Slice-PASEF: fragmenting all ions for maximum sensitivity in proteomics”; bioRxiv. 2022). Specific references to synchronous-PASEF and midia-PASEF in this specification are for illustrative purposes only and are not intended to be limiting.

[0013] The present invention relates to the method as defined in claim 1 and further detailed in the corresponding dependent claims.

[0014] definition

[0015] LC-MS / MS: Tandem liquid chromatography-mass spectrometry, a technique in instrumental analysis in which one or more mass analyzers are coupled together after a liquid chromatography system using additional reaction steps to increase their ability to analyze chemical samples.

[0016] MS1, MS2: In LC-MS / MS experiments, molecules of a given sample are ionized, and their mass-to-charge ratios (usually given as m / z or m / Q) are measured / selected by a mass analyzer (designated MS1). Ions with a specific m / z ratio from MS1 are selected and then fragmented into smaller fragment ions, for example, by collision-induced dissociation, ion-molecule reactions, or photodissociation. These fragments are then introduced into the mass analyzer (MS2), which measures the fragments by their m / z ratio. This fragmentation step allows for the identification and separation of ionized molecules with very similar m / z ratios but producing different fragmentation patterns in MS2. The undisturbed peptide ions that dissociate into smaller fragment ions (often as a result of collision-induced dissociation in MS / MS experiments) are generally referred to as precursors.

[0017] Data-dependent acquisition (DDA): LC-MS / MS, or "shotgun" MS method, is based on the generation of fragment ions from precursor ions automatically selected from the precursor ion distribution in the first (MS1) dimension. The window for the second (MS2) dimension is automatically selected by the machine based on the MS1 output (a single precursor peak). This means that in this mode, the MS2 dimension is not sampled continuously, but selectively based only on the MS1 signal. In a typical shotgun acquisition method, the MS selects the first 10 precursor ions of each MS1 scan for fragmentation to allow measurement in MS2 using a relatively narrow isolation width of 1 to 2 Thomsons. In subsequent scans, the MS typically ignores the precursor ions already selected for fragmentation to allow for the fragmentation of new precursor ions.

[0018] Data-independent acquisition (DIA): An LC-MS / MS method in which all ionized compounds of a given sample falling within a specified mass range in the first MS1 dimension are systematically and unbiasedly fragmented to produce the corresponding spectrum in the MS2 dimension. In contrast to DDA, the MS2 space is continuously sampled in this case. This not only results in a larger data volume but also has the effect that the spectrum measured in the MS2 space includes not only fragments from a single precursor in the MS1 dimension but potentially fragments from several such precursors. A common feature of DIA methods is that instead of selecting and sorting individual precursor peaks, fragmentation is performed over a wider m / z window, resulting in complex spectra containing fragment ions from several precursors. This avoids the missing peptide ID data points common in shotgun methods and potentially allows for the sorting of the entire proteome in a single run, providing a significant advantage over SRM, which can only monitor a small number of peptides per run. Furthermore, DIA exhibits excellent sensitivity and a large dynamic range. To identify peptides present in a sample, fragment ion spectra can be searched against theoretical spectra or mined using transformations similar to SRM. The detected fragments were then arranged in a peak group similar to that of SRM. In DIA acquisition, the window size in the MS2 dimension is typically greater than 30 Thomson. This means that a typical MS2 scan in DIA is more complex than in DDA because significantly more precursor ions are co-fragmented.

[0019] Protein database: Preferably, it is a database selectively targeting only the organism from which the sample is derived, which includes peptide and protein data from that organism, meaning sequence information but no spectral information.

[0020] Spectral library: A database containing information about peptide and protein systems and their fragments, specifically correlating spectral information from LC-MS / MS experiments (including (indexed) retention times, ion mobilities, m / z ratios, and expected relative intensities of fragment ions) with these peptides, proteins, and fragments. A spectral library can be a predictive library, an empirical library, or an intermediate output of library-free search analysis.

[0021] Experience (spectral) libraries (also known as computer-simulated spectral libraries): These are spectral libraries obtained by analyzing data using LC-MS / MS experiments typically employing DDA, as well as protein databases and spectral-centric analyses.

[0022] Predictive (spectral) libraries (also known as computer-simulated spectral libraries): These are spectral libraries obtained using the results of computer simulations, particularly through predictions made by pre-existing machine learning models (similar to, but not limited to, deep learning neural networks). These models are typically trained on large-scale empirical data, which enables them to predict various aspects of the spectral library (similar to, but not limited to, fragmentation, retention time, or ion mobility).

[0023] Deconvolution: Analyzing complex MS2 spectra to determine the processing of the underlying precursors that constitute these spectra.

[0024] Calibration: Calibration is used to detect deviations between theoretical quantities and their empirical counterparts, and is a process to reduce their influence. For example, an untuned MS may cause a relatively large shift in the measured m / z in the MS2 of a precursor. Calibration steps are typically based on some regression analysis to detect and correct for this shift. Calibration is usually performed for m / z, ion mobility, and retention time (or iRT).

[0025] Library-free search analysis: In this field, DIA's library-free search analysis refers to the process of obtaining MS measurements without the need for the specific purpose of creating a spectral library.

[0026] Spectrum-centric analysis: Data analysis of data obtained in LC-MS / MS experiments (which may be DDA or DIA data), where the search is spectrum-centric. This means that spectra in the MS2 dimension are scanned to obtain all theoretical peptides and their fragments that may match, derived from protein databases that typically lack prior spectral information or have limited prior spectral information. Typically, the parent precursor ions of the MS2 spectrum are matched with the theoretical m / z 508 of all precursors in the search space 506 (see also Figure 5 and further corresponding descriptions below) with a specific m / z tolerance, thus giving a set of candidate peptides 509. The theoretical fragment ions 511 with the highest spectral descriptiveness of the candidate peptide are then identified as true peptide spectrum matches (PSM). No further prior information about the fragments is required.

[0027] Peptide-centered analysis / peptide-centered search: Data analysis of data obtained in LC-MS / MS experiments (which may be DDA or DIA data), where the search is precursor-centered. Comparing the spectra in MS1 ​​and MS2 dimension 604, the search queries for predicted possible peptides and their fragments derived from the predicted spectral library or empirical spectral library 601 (see also Figure 6 and further corresponding descriptions below). In this analysis, spectral information of the peptide is required, particularly retention time, ion mobility, and the possibility of observing fragment ions with relative fragment intensities. This information is used to narrow the peptide search space by querying only spectra falling within a specific m / z, iRT, or IM tolerance 605, and is used to score matches. Having this additional information greatly improves the sensitivity of the analysis by generating a stronger score 608.

[0028] Mobility plot: The trace of an ion along the ion mobility dimension at a specific retention time. It is generated by tracking the ion in a (typically) single 5D scan. The x-axis is a measure of ion mobility (e.g., 1 / K0, reverse ion mobility, collision cross-section), and the y-axis is intensity. Mobility plots from multiple retention time points can be summarized into a new mobility plot.

[0029] Extracted Ion Chromatography (XIC): The traces of ions along the retention time dimension. It is generated based on 5D scans corresponding to multiple retention time points. A single data point at a specific retention time in the XIC can be based on vertex intensity, area under the curve, or the sum of all data points in the mobility plot. The x-axis is the retention time dimension, and the y-axis is intensity. It is beneficial to observe both dimensions (and thus both in the mobility plot and in the extracted ion chromatogram) because, at the XIC level, the ion mobility level is compressed.

[0030] Ion Mobility (IM) Index: In the context of this invention, an ion mobility index refers to an index within a master table containing all unique ion mobility dimension values ​​for a given LC-MS measurement. Therefore, it is an index mapped to a reference or master table containing all unique ion mobility values ​​for that scan (see further example 425 below).

[0031] Isolation Index: In the context of this invention, an isolation index refers to an index within the main table (further see Example 418 below) that contains all unique m / z isolation ranges applied in a given LC-MS measurement for isolating the MS1 signal for subsequent MS2 fragmentation.

[0032] DIA Window: The scan area, i.e., the region covered in the m / z-IM plane during a single IM scan, is generally preferably arranged diagonally. The term "generally diagonal" will be interpreted in a broad sense, and in particular, it does not require the window to extend along a straight line. Rather, the term "generally diagonal" primarily reflects the fact that both the m / z range and the IM range vary simultaneously in a continuous scan area.

[0033] Targeted peptides: Peptides derived from the query search space. For example, these could be all peptides known to exist in the human proteome.

[0034] Decoy peptides: Artificially constructed peptides added to the search space. These are primarily used for machine learning and error detection estimation to distinguish true target peptides (peptides that are actually present in the sample) from false target peptides (peptides that are not present in the sample and therefore behave like artificially constructed decoy peptides) in the search space.

[0035] Microscan: Performs an MS1 ​​or MS2 scan with a specific precursor isolation width at a specific ion mobility and retention time. The x-axis is m / z, and the y-axis is intensity. Therefore, it is a single TIMS push event that produces a list of m / z intensity pairs for a single point in the ion mobility dimension.

[0036] TIMS Scan: In a single TIMS scan, ions from a selected mass range are fragmented to record the ion mobility-resolved MS2 spectra of all precursors. A TIMS scan typically consists of hundreds of microscans, each of which is an MS2 spectrum of a selected mass range of precursors fragmented at a specific ion mobility.

[0037] According to a first aspect of the invention, the present invention relates to a method for identifying data-independent acquisition (DIA) targeted peptide precursors from mass spectrometry intensity data of a sample, the mass spectrometry intensity data being acquired based on mass-to-charge ratio (m / z), retention time (RT), and ion mobility (IM).

[0038] This was accomplished using data acquired in the following manner: introducing precursor ions from the sample into an ion mobility separator, continuously releasing precursor ions from the ion mobility separator according to their ion mobility, introducing the released precursor ions into a mass filter that selectively transmits precursor ions with m / z values ​​falling within a controllable m / z window, fragmenting the precursor ions transmitted through the mass filter to generate fragment ions, and performing mass spectrometry measurements on the fragment ions, wherein each fragment ion is associated with a mass window and an ion mobility (IM) range, and the detected fragments are associated with their corresponding precursor ions.

[0039] The ion mobility separator and the mass filter are controlled synchronously to perform multiple ion mobility scans, during which precursor ions that increase or decrease ion mobility (IM) are successively released from the ion mobility separator, and during this period, the mass window of the mass filter is continuously or gradually moved toward lower or higher m / z values, respectively.

[0040] For a given RT value, the obtained data includes: at least two quality windows (DIA windows) of the quality filter, each quality window being able to be assigned a DIA window index, and each quality window including multiple microscans performed at a specific ion mobility with a specific precursor m / z isolation width, each microscan being able to be assigned an isolation index.

[0041] To enable querying fragment ions for a given precursor ion, at each retention time point, all ions in the MS2 scan are associated with their ion mobility and isolation index and stored in a (at least) 5D data matrix, which includes or consists of dimension m / z, intensity, ion mobility index, isolation index, and DIA window. This 5D data scan can be used to return intensity based on m / z tolerance, ion mobility tolerance, and precursor m / z.

[0042] To correlate detected fragments with their corresponding precursor ions, data from at least two mass windows are combined into a (at least) 5D dataset (for a given RT value) based on mass-to-charge ratio (m / z), intensity, ion mobility (index) (IM), DIA window (index), and isolation index.

[0043] According to a first preferred embodiment of the method, in order to associate the detected fragment with its corresponding precursor ion, for each target (and decoy) peptide precursor in the spectral library, ion traces are extracted from a 5D dataset, and a score is calculated, preferably for multiple scores, and preferably for at least one or all of the fully extracted ion chromatograms (XIC) and partially extracted ion chromatograms and mobility maps, to separate the target from the decoy, preferably by using the calculated scores, including at least one based on correlation, peak shape, and intensity, particularly by using machine learning, preferably by analyzing the false discovery rate based on the target / decoy.

[0044] The mass windows associated with continuous mass spectrometry measurements of the target ion preferably do not overlap.

[0045] In a single IM scan, the regions covering the m / z-IM plane are generally arranged roughly diagonally.

[0046] Similarly, the term "approximately diagonal" will be interpreted in a broad sense, and in particular, does not necessarily require the frame to extend along a straight line. Rather, the term "approximately diagonal" primarily reflects the fact that both the m / z range and the IM range vary simultaneously within a continuous scan region. For example, the shape of a diagonal slice does not need to be linear but can be polynomial.

[0047] Preferably, for each target and decoy peptide precursor in the spectral library, ion traces are extracted from the 5D dataset preferably in the form of fully extracted ion chromatograms by the following operations: in the RT dimension, iteration is performed from the first scan to the last scan on all 5D IM scans measured for a given ion within the range of interest of a given RT, wherein, during the iteration process, for each 5D IM scan, the first module determines the starting index of the peak to be iterated by performing a binary search of the lower bound of the m / z window of the fragment ions of the peaks preferably sorted and classified by m / z in the 5DIM scan, until it encounters the last peak in the array or a peak whose m / z is higher than the upper bound of the m / z window, wherein, in each iteration step, the module checks whether the IM index of the current peak falls within the IM window and whether the m / z of the parent peptide precursor is within the isolation range of the microscan measured for that peak, and if yes, the peak intensity is added to the run and the peak index is then incremented, and if false, the peak index is directly incremented and then returned to the first module and repeated until the iteration is complete.

[0048] The process preferably includes the following steps: ensuring that peaks measured only in a specific DIA window are used to generate extracted ion chromatograms (XIC) in the RT dimension.

[0049] According to yet another preferred embodiment, preferably for each target and bait peptide precursor in the spectral library, ion traces are extracted from the 5D dataset in the form of a complete mobility map by: finding an IM scan corresponding to a specific RT for which the mobility map is to be created, preferably this is done for the vertex RT point of the extracted ion chromatogram, wherein, in the obtained 5D IM scan, a preferably binary search is performed to find a peak index that is closest to but still within the lower bound of the m / z window, and then the exit condition of the current peak index is checked, wherein if the current peak index is greater than 5D... If the index of the last peak in the IM scan, or the m / z of the peak at the current peak index, is not within the m / z window, the exit condition is met. If the exit condition is not met, the IM index of the current peak is checked to see if it falls within the IM window, and if the m / z of the precursor peptide is within the isolation range of the microscan measured for that peak. If true, the intensity value of the current peak is added to the mobility map array of the current peak's IM index, and then the current peak index is incremented. If false, the current peak index is incremented, and the exit condition is checked again for the given new peak index.

[0050] Preferably, for each target and decoy peptide precursor in the spectral library, ion traces can be extracted from the 5D dataset in the form of a partial mobility map by: finding an IM scan corresponding to a specific RT for which the mobility map is to be created, preferably done for the vertex RT point of the extracted ion chromatogram, wherein, in the resulting 5D IM scan, a preferably binary search is performed to find a peak index that is closest to but still within the lower bound of the m / z window, and then the exit condition of the current peak index is checked, wherein if the current peak index is greater than 5D... If the index of the last peak in the IM scan, or the m / z of the peak at the current peak index, is not within the m / z window, the exit condition is met. If the exit condition is not met, it is checked whether the IM index of the current peak falls within the IM window and whether the m / z of the precursor peptide is within the isolation range of the microscan measured for that peak. If true, the intensity value of the current peak is added to the mobility map array of the IM index of the current peak, and then the current peak index is incremented. If false, the current peak index is incremented, and the exit condition is checked again for the given new peak index. This process includes the following steps: ensuring that only peaks measured in a specific DIA window are used to generate the mobility map.

[0051] For the analysis of 5D datasets, data can be loaded in virtual batches with arbitrary m / z widths.

[0052] In LC-MS / MS experiments, the data are preferably a set of data obtained from the sample (preferably a digested proteome sample) that are independent of acquisition data.

[0053] The data is preferably in the form of sample mass spectrum intensity data acquired using LC tandem mass spectrometry, preferably LC-DIA, based on the mass-to-charge ratio (m / z), retention time (RT), and ion mobility (IM).

[0054] More preferably, in the LC-MS / MS experiment, the data are a set of data acquired independently of the sample, wherein the sample is a complex mixture of at least one protein of interest and other proteins and / or other biomolecules, in the form of a complex natural biological matrix that has been decomposed prior to LC-MS / MS analysis.

[0055] The at least one protein of interest can be a protein based solely on protein amino acids, or a protein based on protein amino acids and carrying post-translational modifications.

[0056] The ion mobility separator is preferably a TIMS analyzer, and more preferably a TIMS analyzer with parallel accumulation and separation, particularly operating using a method comprising the following steps: (a) accumulating ions in an RF ion trap; (b) transferring at least a subset of the accumulated ions to a captured ion mobility separator, wherein the transferred ions are radially confined by an RF field and pushed by a gas flow against the rising edge of an axial DC electric field barrier, such that the transferred ions are spatially separated along the rising edge according to their ion mobility; (c) releasing the transferred ions sequentially according to their ion mobility by lowering the height of the DC electric field barrier, while ions from the ion source are further accumulated in the RF ion trap; and (d) restoring the height of the DC electric field barrier, triggering a continuous transfer of the accumulated ions from the RF ion trap to the captured ion mobility separator.

[0057] A TIMS analyzer includes a gas flow that drives ions against a counteracting electric field barrier, causing the ions to be initially trapped along the axis of the TIMS analyzer. The ions are confined in the radial direction by an RF electric field. After transferring the ions from the ion source to the electric field barrier, the height of the electric field barrier or the gas velocity is adjusted so that the ionic material is released from the electric field barrier in order of its mobility.

[0058] Typically, the ion mobility separation unit of a TIMS analyzer is only about 5 cm long. A radial RF quadrupole field is generated within a small tube with an inner diameter of approximately 8 mm to hold ions near the axis. The gas flow within the tube drives the ions entrained in the gas flow against a tilted, counteracting DC electric field barrier, where ions are captured and separated according to their mobility at their position on the electric field ramp, where the friction of the moving gas equals the reaction force of the DC electric field on the ramp. After ion loading into the TIMS, the height of the DC electric field barrier decreases; the scan releases ionic material in order of ion mobility. Unlike many other experiments building small ion mobility spectrometers, this small device has achieved high R0 values ​​while reducing scan speed. mob An ion mobility resolution of 400 is very high.

[0059] At the start of an LC observation retention time window, preferably within the range of 1 to 15 seconds, particularly preferably within the range of 3 to 10 seconds, a full scan is preferably performed, wherein, in this full scan, the full ion mobility width and the full m / z width of interest are scanned, and wherein, based on the full scan of the remaining portion of the LC observation window, the second ion mobility separator and the mass filter are synchronously controlled to perform multiple IM scans, during which precursor ions that increase or decrease IM are successively released from the IMS, and during this period, the mass window of the mass filter is continuously or gradually moved toward lower or higher m / z values, respectively, to avoid identifying peptides of no interest in the full scan, and wherein the step of associating detected fragments with their corresponding precursor ions is based on determining or utilizing the corresponding mass window and IM range associated with each occurrence of the fragment in the mass spectrometry measurement.

[0060] According to another aspect of the invention, the invention relates to the application of methods according to any of the foregoing aspects for determining at least one of the compositions of a sample including quantitative information about the components or a medically relevant conformation of the components; for determining a protein-based drug or its effect; for the effect of a drug or other ligand on a protein; or for quality control of a protein-based pharmaceutical preparation.

[0061] Finally, the present invention relates to a computer program product that analyzes data using the methods described in the above description, or a computer-readable medium having such a computer program product stored thereon.

[0062] Further embodiments of the invention are described in the dependent claims. Attached Figure Description

[0063] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings, which are for the purpose of illustrating the presently preferred embodiments of the invention and not for the purpose of limiting the invention. In the drawings,

[0064] Figure 1 shows a schematic diagram of an ion cloud (ellipse) using the standard dia PASEF measurement scheme, where the y-axis represents ion mobility and the x-axis represents m / z, with the vertical box depicting the DIA window.

[0065] Figure 2 shows a schematic diagram of an ion cloud (ellipse) using the synchro-PASEF scheme in a), where the ion mobility is on the y-axis and the m / z is on the x-axis, and the DIA window is represented by a diagonal box. Figure 2 also shows a schematic diagram of how the present invention conceptually processes the method depicted in Figure 1.

[0066] Figure 3 shows example ion traces of fragments belonging to the peptide across all DIA windows in the RT dimension in a), example ion traces of fragments belonging to the peptide in the IM dimension at retention time when the peptide has vertex intensity by merging diagonal scans in b), example ion traces of fragments belonging to the peptide in each measured DIA window in the RT dimension in c), and example ion traces of fragments belonging to the peptide in each measured DIA window in the RT dimension when the peptide has vertex intensity at retention time by merging diagonal scans in d), as well as the individual DIA windows in which the fragments are measured separately.

[0067] Figure 4 shows the 5D IM scan 400 and its relationship to the raw data from the mass spectrometer in a), describes in b) the input of the 5D IM scan that may need to be queried in an efficient manner in one implementation, describes in c) a high-level flowchart depicting one implementation, describes in d) an implementation of a flowchart for creating a complete XIC, describes in e) an implementation of a flowchart for creating a partial XIC, describes in f) an implementation of a flowchart for creating a module of a complete mobility map, and describes in g) an implementation of a flowchart for creating a partial mobility map.

[0068] Figure 5 shows an example of a complete XIC constructed from 5D IM scans in a), and an implementation of a method for constructing ion traces from 5D IM scans in b).

[0069] Figure 6 shows examples of partial mobility maps of fragment ions constructed based on 5D IM scans in a) to c).

[0070] Figure 7 illustrates one implementation of a method to improve memory usage in a data processing pipeline by loading data in virtual batches. Detailed Implementation

[0071] In one implementation of the invention, it is operated by first converting mass spectrometry data into a series of 5D ion mobility (IM) scans having dimensions m / z

[401] , intensity

[402] , IM index

[403] , isolation index

[404] , and DIA window

[405] . To generate a 5D scan from a measurement of a value of RT (or more precisely, an RT window or a single time point), all m / z intensity pairs from all individual MS2 DIA windows 204, 205, 206 are combined into a list. Each data point in this list is annotated with a parent DIA window index 405 (identifying which of the original DIA windows 204, 205, 206 the data point originates from), the actual isolation experienced by the m / z intensity pair (encoded as an isolation index of a reference master table (as depicted in Figure 4a)), and the corresponding ion mobility (also encoded as an IM index 402 of a master table referencing all ion mobility values). The resulting list is then categorized by m / z (401) and stored in a matrix of dimension n multiplied by 5. Figure 4a shows an example 5D IM scan at a specific RT point 406.

[0072] Then, during targeted analysis using spectral libraries (predicted or empirical), data was queried for each peptide precursor (Figures 4d to 4g). Data was queried in two ways.

[0073] First, the data is queried by treating it as if it were collected by a single diagonal DIA window (Figure 4d, Figure 4f). This is achieved by not using any DIA window filters in the query.

[0074] Secondly, the data is queried using DIA window filters to collect ion traces for each diagonal DIA window (Fig. 4e, Fig. 4g). These queries are used to create ion traces in both retention time (RT) and ion mobility (IM) dimensions, yielding four different types of ion traces, each potentially containing partially different information. Furthermore, the four types of ion traces can be combined. The resulting ion traces can be used for scoring in different ways (e.g., correlation scoring, shape scoring, etc.), as is typically done in targeted analysis (Fig. 4c). Additionally, we describe a novel method for processing this type of data through virtual blocks to efficiently utilize memory (Fig. 7).

[0075] Figure 1 shows a schematic diagram of the ion cloud (ellipse), with ion mobility on the y-axis 101 and m / z on the x-axis 102. The vertical frame depicts the DIA window measured in a loop consisting of eight full TIMS ion mobility scans. In the ion mobility dimension, the DIA window with a continuous frame depicts a full TIMS scan 104a, 104b. As in 104a and 104b, each full TIMS scan is a collection of microscans and can be considered an array of true ion mobility indices 103 when the m / z dimension is ignored. The m / z isolation range for all microscans within the DIA window is identical and unique for that DIA window. A diagram of peptide 105 measured in DIA window 104a is shown. The vertical DIA window frame is geometrically “awkward” when following the ion cloud and results in inefficient sampling of the ion cloud.

[0076] Figure 2a shows a schematic diagram of the ion cloud (ellipse), with ion mobility on the y-axis 201 and m / z on the x-axis 202. Unlike Figure 1, the DIA windows are represented by diagonal boxes 204, 205, and 206. These diagonal DIA windows may overlap or not overlap in the m / z dimension. Each diagonal DIA window represents a full TIMS scan and is a collection of microscans. However, in this scheme, each microscan within the DIA window can have different isolation ranges in both the ion mobility 201 and the m / z dimension 202. This allows for efficient sampling of the ion cloud by following its natural shape, which is not possible with vertical boxes having the same m / z isolation range as depicted in Figure 1. In this type of scheme, it is generally desirable to measure peptides through multiple DIA windows. A diagram of peptide 200 measured in all three DIA windows 204, 205, and 206 is shown. In one variant, a small number of consecutive microscans in the diagonal DIA window can have the same m / z range but different IM ranges.

[0077] In the classic dia-PASEF method (Figure 1), in each DIA window, the acquisition software measures a fixed m / z isolation window 106 within a specific range of ion mobility 101. This results in several rectangular regions for acquiring MS2 ions, as shown in boxes 104a / 104b. In contrast, a simpler form of diagonal dia-PASEF moves the isolation window by a fixed amount with each increment in the ion mobility dimension 201. This results in the acquired MS2 window having a distinctive diamond shape (Figure 2a).

[0078] Figure 2b illustrates a schematic diagram of how the present invention conceptually processes the method depicted in Figure 2a. Multiple diagonal DIA windows 204, 205, 206 are merged as if there were only one DIA window 207 per MS1 cycle, but this is done in a manner without information loss. This enables seamless extraction of fragment ion traces of peptide precursor ions measured in both the RT and ion mobility dimensions through multiple diagonal DIA windows. An illustration of peptide 200 is shown, which can now be considered as being measured in a single DIA window 207. Additionally, because we have preserved information, we can also extract fragment ion traces of peptide 200 in each DIA window 204, 205, 206 in which peptide 200 is measured.

[0079] Figure 3a shows an example ion trace belonging to fragments of peptide 200 across all DIA windows in the RT dimension. It was constructed by extracting data from several merged diagonal scans 207 measured between time points 13.25 min and 13.50 min. We refer to it as the complete XIC 301 because it is based on all measurements related to this peptide in both the retention time and ion mobility dimensions.

[0080] Figure 3b shows an example ion trace belonging to fragments of peptide 200 in the IM dimension at the retention time when the peptide has a peak intensity (13.375 min) using a merged diagonal scan 207. Similar mobility maps can be created in the RT dimension from other measurement time points. We refer to this as the complete mobility map 302 because it is based on all measurements associated with the peptide at a given retention time point.

[0081] Figure 3c shows example ion traces belonging to fragments of peptide 200 in each DIA window measuring peptide 200 along the RT dimension. Partial XIC 303 is constructed by extracting data from several merged diagonal scans by selecting only the data measured by a specific diagonal DIA window. 301 can be considered the sum of all these individual partial XICs. However, additional information can be obtained by examining the individual partial XICs, as they may potentially have different interferences. In another implementation, they can also be connected across the DIA window and additionally across the RT dimension, also as two additional variations of the mobility plot used for scoring.

[0082] Figure 3d shows an example ion trace belonging to fragments of peptide 200 in the IM dimension at retention time using a merged diagonal scan 207, with individual DIA windows measured within it. Similar mobility maps can be created in the RT dimension from other measurement time points. We refer to this as partial mobility map 304 because it is based on measurements associated with the peptide at a given retention time point and a specific DIA window.

[0083] Figure 4a illustrates the 5D IM scan 400 and its relationship to the raw data from the mass spectrometer. In one implementation, at each given RT position, the mass spectrometer can separate compounds by their ion mobility. Data representing the full IM scan is recorded across multiple microscans, where each microscan, corresponding to a specific ion mobility 403 (the actual measured ion mobility value 419 is used in index form, as given in the corresponding master table 425), is a set of measured peaks with associated m / z dimensions 401 and intensities 402. Additionally, for each microscan, a specific m / z range is used to isolate precursor ions 407, which can be encoded as an isolation index 404 that references the actual range isolation used, returned based on the global isolation master table 418. Finally, we can encode the DIA window 405 in which a specific microscan is measured. These dimensions can be implemented as a single multidimensional array with five dimensions, as shown below. In one implementation, a single full IM scan can consist of hundreds of microscans, each with hundreds of data points. In another embodiment, microscans can also be merged (an array for data processing that preserves the individual attribution information). In yet another embodiment, each microscan in the full scan can be represented by its ion mobility scan index 403 (as a placeholder for the actual ion mobility value), precursor isolation index 404 (as a placeholder for the actual parent precursor isolation m / z range 407), and DIA window 405. In one implementation, different microscans are converted into a single 5D IM scan 400 by merging all individual microscans in the full IM scan into a single array with five dimensions (m / z 401, intensity 402, IM index 403, precursor isolation index 404, and DIA window 405). This 5D IM scan can be categorized by m / z 401 for quick m / z-based access.

[0084] Figure 4b illustrates that in one implementation, it may be necessary to query the 5D IM scan in an efficient manner to generate inputs 301, 302, 303, and 304. Window (range) 410 (m / z window), 411 (RT window), and 412 (IM window) are typically determined empirically based on prior analysis to determine the optimal tolerance window.

[0085] Figure 4c illustrates a high-level flowchart depicting one implementation of the roles of modules 420 and 421 within the context of other modules typically used for target analysis of DIA data 422, 423, and 424. The modules described in this implementation are responsible for converting the raw data into 5D IM scans 420, which are then queried to create different types of ion traces 301, 302, 303, and 304 for each target and decoy peptide precursor using module 421. To generate a single 5D scan from a single time point, all m / z intensity pairs from all individual MS2 frames 204, 205, and 206 are combined into a list. Each data point in this list is annotated with the parent DIA window index 405, the actual isolation experienced by the m / z intensity pair (encoded as an isolation index in the reference master table), and the corresponding ion mobility (also encoded as an IM index 402 in the reference master table of all ion mobility values). The resulting list is then categorized by m / z (401) and stored in a matrix of dimension n multiplied by 5. These ion traces are subsequently scored using various scores based on correlation, peak shape, etc., and used for peptide precursor identification in a manner known in the art. Without encoded information about the parent isolation for each data point, the query used to extract XIC data would also include data points that could not possibly originate from the query ion in question. Encoded information including the parent isolation for each data point produces cleaner (less noisy) data for subsequent data analysis.

[0086] Figure 4d illustrates one implementation of the flowchart for module 421a (see Figure 5a for an example of 421a) used to create a complete XIC 301. In the RT dimension, the new method iterates over all 5D IM scans measured within the RT range of interest for a given ion 411, from the first scan to the last. During the iteration process 430, for each 5D IM scan, module 431 determines the starting index of the peak to be iterated over by performing a binary search of the lower bound of the m / z window of the fragment ions 410 pre-sorted by m / z in the 5D IM scan 401. 432 Iteration continues until it encounters the last peak in the array or a peak whose m / z is above the upper bound of the m / z window 410. In each iteration step, module 433 checks whether the IM index 403 of the current peak falls within the IM window 412, and whether the m / z of the parent peptide precursor 408 is within the isolation range 407 of the microscan measuring that peak. If true, the peak intensity is added to run and 434, and then the peak index 435 is incremented. If false, the peak index 435 is incremented directly. It then returns to module 431 and continues repeating until the iteration is complete. At this point, module 436 sets the current run and to the intensity value of the current RT in the XIC array and resets run and to zero. Modules 437 and 430 repeat this entire process until the last 5D IM scan is processed.

[0087] Figure 4e illustrates one implementation of the flowchart of module 421b used to create partial XIC 303. One difference from 421a is that 421b introduces a submodule 440, which ensures that only peaks measured in a specific DIA window are used to generate XIC in the RT dimension.

[0088] Figure 4f illustrates one implementation of the flowchart for module 421c used to create the complete mobility map 302. Module 441 searches for a 5D IM scan corresponding to the specific RT for which the mobility map is to be created. Typically, this is done for the vertex RT point of the XIC, but it can also be done for all or a subset of the measurement points in the RT dimension. In the 5D IM scan, a binary search is performed to find a peak index that is closest to the lower bound of the m / z window 410 but still within its range. Then, in module 443, an exit condition is checked for the current peak index. If the current peak index is greater than the last peak index in the 5D IM scan, or if the m / z of the peak at the current peak index is not within the m / z window 410, the exit condition is met. If the exit condition is not met, module 444 checks whether the IM index 403 of the current peak falls within the IM window 412, and whether the m / z of the precursor peptide 408 is within the isolation range of the microscan of the peak measured 407. If true, the intensity value of the current peak is added to the mobility graph array of the current peak's IM index 445, and then the current peak index is incremented 446. If false, the current peak index is incremented 446. Given a new peak index, module 443 checks the exit condition again and repeats the entire process 444 to 446. The method stops when the exit condition is met.

[0089] Figure 4g illustrates one implementation of the flowchart for module 421d used to create partial mobility plot 304 (see Figure 6a for an example of module 421d). One difference from 421 is that module 447 ensures that only peaks measured within a specific DIA window are used to generate the mobility plot.

[0090] Figure 5a illustrates an example 421a of constructing a complete XIC of fragment ions based on a 5D IM scan 520 using an example. Similar to other implementations and examples mentioned in this disclosure, this example is intended to be non-limiting and for illustrative purposes only. In one implementation, a 5D IM scan 520 is created for each full IM scan at a given RT. Ion traces 518, 519 are constructed for fragments belonging to a parent precursor ion having a specific expected m / z 500, a retention time window 503, and an ion mobility window 504. Example fragments have a specific expected m / z 501 and an m / z window 502. Each 5D IM scan within a specific RT window 503 is processed iteratively. Here, we illustrate an example 5D IM scan 520 with an RT index of 30. The method of the present invention will first use a binary search in a pre-classified m / z dimension 512 to find the starting position corresponding to the lower bound of the m / z window. It then iterates through each position in the array, starting from the first index and continuing to the end position 513, where m / z 506 is still within the m / z window 502. For each position it iterates through, it sums all the intensity values, where IM index 508 is within the expected IM window 504 and isolation index 509 corresponds to the parent isolation range 511 containing the target precursor m / z 500. Even in this example, for simplicity, only one isolation index corresponds to the target precursor (index 2). In practice, there could be several isolation indices where the parent precursor is measured. In this example, the cells at positions [i, 3] and [i, 5] match all conditions, and their intensities 507 are summed 514 to give a single data point 515 of the intensity 517 of the fragment ion at the current RT position 516. Tracking such data points for all fragment ions belonging to the peptide precursor across the entire range of RT window 501 will create a complete XIC of 301.

[0091] Figure 5b illustrates an implementation of the method for constructing ion traces of fragment ions based on a 5D IM scan, using an example. Continuing from the example in Figure 5a, at RT index 31, we iterate to the next 5D IM scan 527. Following the same processing as described in Figure 5a, the three data points at [i, 4], [i, 5], and [i, 6] are summed 523 to obtain the intensity of the fragment ion 524 at RT index 31. The updated XIC plot with traces derived based on RT indices 30 and 31 is shown as 526.

[0092] Figure 6a illustrates an example of 421d to construct a partial mobility map 304 of fragment ions based on a 5D IM scan. Initially, a two-dimensional mobility map trace array 606 is initialized with a length based on the range of an IM window 605. IM indices 607 in the array are initialized to the range of an IM window (12 to 19), where intensity values ​​are set to zero. Subsequently, a binary search in a pre-classified m / z dimension 512 is used to find the starting position corresponding to the lower bound of the m / z window. It then iterates through each position in the array starting from the starting index until it satisfies an exit condition 443. At the starting position ([i, 1] in this example), it checks whether acceptance conditions 444 and 447 are true. In this example, this translates to: 1) isolation index 614 equals 2 because the target precursor 600 is covered by precursor isolation 617 from 520 to 580, 2) IM index 613 values ​​are in the range of 12 to 19, and 3) DIA window 615 equals 1 604. At the starting position, acceptance condition 619 is not met, which means that the strength value 608 for [i, j] is set to zero, where j is calculated by equation 609 (in this example, j=5).

[0093] Figure 6b shows an example of 421d to construct a partial mobility map 304 of fragment ions based on a 5D IM scan. Continuing from Figure 6a, the current position 621 is incremented by 1. When the acceptance condition passes 622, the intensity value 608 for [i, 3] 623 is set to the intensity value 612 at the current position.

[0094] Figure 6c shows an example of 421d to construct a partial mobility map 304 of fragment ions based on a 5D IM scan. Continuing from Figure 6b, we skip several iteration points and consider the final position 631 before the exit condition is met. When the acceptance condition passes 632, the intensity value 608 for [i, 0] 633 is set to the intensity value 612 at the current position. The resulting mobility map 635 is a partial ion trace of a single fragment ion for a DIA window equal to 1.

[0095] Figure 7 illustrates one implementation of a method to improve memory footprint in a data processing pipeline by loading data in virtual batches. As depicted in Figure 3, there is a high memory cost associated with processing data. We propose a novel batch loading method that facilitates implementation using a 5D IM scan data structure as shown in Figure 4a. Instead of loading all microscans belonging to the 5D IM scan (described here as rectangular blocks) into memory, we can load them using virtual blocks of arbitrary m / z width. Two examples of such virtual blocks are shown in Figures 803 and 804. Each block has microscans associated with it that fall entirely or partially within the block's boundaries. To process block 803, all microscans of types 805 and 807 must be loaded. Similarly, to process block 804, all microscans of types 806 and 807 need to be loaded. Microscans belonging to block 807 must be loaded at least twice because they are associated with two blocks. The advantage of this approach is that not all microscans need to be loaded into memory simultaneously.

[0096] the term

[0097] 1. TIMS Scan: In a single TIMS scan, ions from a selected mass range are fragmented to record the ion mobility-resolved MS2 spectra of all precursors. A TIMS scan typically consists of hundreds of microscans, each of which is an MS2 spectrum of a precursor fragmented at a specific ion mobility for a selected mass range.

[0098] 2. Microscan: Perform MS1 or MS2 scans with a specific precursor isolation width at specific ion mobilities and retention times. The x-axis represents m / z, and the y-axis represents intensity.

[0099] 3. Mobility Map: The traces of ions along the ion mobility dimension. It is typically generated by tracking ions in a single 5D scan. The x-axis is a measure of ion mobility (e.g., reverse ion mobility, collision cross-section), and the y-axis is intensity.

[0100] Those skilled in the art will understand that, in addition to proteomics, this invention can be applied to other mass spectrometry-based omics data, including metabolomics. Furthermore, those skilled in the art will understand that, besides synchronous-PASEF or midia-PASEF, this invention can be used with other similar methods. Specific references to synchronous-PASEF and midia-PASEF in this specification are for illustrative purposes only and are not intended to be limiting. Those skilled in the art will also understand that this system and method can be extended or applied to any additional dimension.

[0101] While certain aspects of the invention have been specifically shown and described with reference to exemplary embodiments thereof, those skilled in the art will understand that various changes in form and detail may be made without departing from the spirit and scope of the invention as defined by the appended claims. It will also be understood that components of this disclosure may include hardware components or combinations of hardware and software components. Hardware components, methods, and workflows may include any suitable tangible components constructed or arranged to operate as described herein. Some of the hardware components may include processing circuitry (e.g., a processor or a group of processors) to perform the operations described herein. Software components may include code recorded on a tangible computer-readable medium. Processing circuitry may be configured by the software components to perform the described operations. Therefore, this embodiment is intended to be illustrative rather than restrictive in all respects.

[0102] List of Figure Labels: DDA Data Dependent Acquisition, DIA Data Independent Acquisition, IM Ion Mobility, iRT Indexed Retention Time, m / z Mass-to-charge Ratio, MRM Multiple Reaction Monitoring, MS1 First Spectral Dimension in LC-MS / MS Experiment, MS2 Second Spectral Dimension in LC-MS / MS Experiment, PASEF Parallel Accumulation Serial Fragmentation, RT Retention Time, SRM Selected Reaction Monitoring, TIMS Captured Ion Mobility Spectrum, XIC Extracted Ion Chromatogram, 101 y-axis, ion mobility, 102 x-axis, mass-to-charge ratio, 103 Ion Mobility Index, 104 Microscan, 105 Peptide shown, 106 m / z isolation window of a DIA window, 200 Peptide shown, 201 y-axis, 202 x-axis.Mass-to-charge ratio 203 Ion mobility index 204-206 Diagonal DIA window 207 Diagonal DIA window for analysis / scan 301 Full XIC 302 Full mobility plot 303 Partial XIC 304 Partial mobility plot 400 5 DIM scan example 401 m / z dimension 402 Intensity 403 Specific ion mobility 404 Isolation index 405 DIA window 406 RT point 407 Precursor isolation (m / z) 408 m / z of parent peptide precursor 409 m / z of fragment ions 410 m / z window (range) 411 RT window (range) 412 IM window (range) 413 DIA window filter 418 Isolation master table 419 Ion mobility values ​​420 Merging into 5D data 421 Extracting ion traces 422 Calculating scores 423 Using scores alone 424 Peptide identification 425 Ion mobility master table 430 Iterative processing 431 Finding the starting peak index 432 Crossover point 43 3 Crosspoint 434 Increase Intensity 435 Incremental Index 436 Set Intensity Value 437 Crosspoint 440 Submodule 441 Search 442 Search Index 443 Crosspoint 444 Crosspoint 445 Set Intensity 446 Incremental Index 447 Crosspoint 500 Target Precursor 501 Extract Fragment 502 m / z Window 503 RT Window 504 IM Window 505 DIA Window Filter 506 m / z 507 Intensity 508 IM Index 509 Isolation Index 510 DIA Window 511 Precursor Isolation 512 Start Position 513 End Position 514 Sum 515 Intensity Data Entry 516 RT Index 518 XIC Array 519 XIC 520 5D IM Scan 523 Total 524 Data Entry Intensity 525 XIC Array 526 XIC 527 5D IM Scan 600 Targeted Precursor 601 Fragment Extraction 602 m / z Window 603 IM Window 604 DIA Window 605 Mobility Map Trace Length 606 Mobility Map Trace 607 IM Index 608 Intensity 609 Index 6105D IM Scan 611 m / z 612 Intensity 613 IM Index 614 Isolation Index 615 DIA Window 616 Index 617 Precursor Isolation 618 Starting Position 619 Not Satisfied 620 Mobility Map 621 Current Position 622 Passing 623 Index 624 Mobility Map Trace 625 Mobility Map 631 Current Last Position 632 Passing 633 Index 634 Mobility Map Trace 635 Mobility Map 801 Ion Mobility 802 m / z 803-804 Virtual Block 805-807 Micro Scan

Claims

1. A method for identifying targeted peptide precursors from sample mass spectrometry intensity data using data obtained through the following operations: introducing precursor ions from the sample into an ion mobility separator; sequentially releasing precursor ions from the ion mobility separator according to the ion mobility of the precursor ions; introducing the released precursor ions into a mass filter; selectively transmitting precursor ions with m / z values ​​falling within a controllable m / z window; fragmenting the precursor ions transmitted through the mass filter to generate fragment ions; and performing mass spectrometry measurements on the fragment ions, wherein... Each fragment ion is associated with a mass window and an ion mobility (IM) range, and detected fragments are associated with their corresponding precursor ions. The ion mobility separator and the mass filter are controlled synchronously to perform multiple ion mobility scans, during which precursor ions with increasing or decreasing ion mobility (IM) are sequentially released from the ion mobility separator. During this process, the mass window of the mass filter is moved continuously or gradually toward lower or higher m / z values, respectively. For a given RT value, the obtained data includes at least two mass windows of the mass filter (204-). 206), each mass window can be assigned a DIA window index (405), and each mass window (204-206) includes multiple microscans (104) performed at a specific precursor m / z isolation width at a specific ion mobility, each microscan being assigned an isolation index (404), wherein, in order to associate the detected fragments with their corresponding precursor ions, the data of the at least two mass windows (204-206) are combined into at least a 5D dataset (207) according to the mass-to-charge ratio (m / z), intensity, ion mobility index (IM), DIA window index (405) and isolation index (404).

2. The method according to claim 1, wherein, In order to associate the detected fragments with their corresponding precursor ions, for each target and decoy peptide precursor in the spectral library, ion traces (421) are extracted from the 5D dataset (207), and scores (422) are calculated, preferably for multiple scores, and preferably for at least one or all of the fully extracted ion chromatograms (XIC) and partially extracted ion chromatograms (XIC) and mobility maps, to separate the target from the decoy (423), preferably by using the calculated scores, including at least one of correlation, peak shape, and intensity, particularly by using machine learning, preferably by analyzing the false discovery rate based on the target / decoy.

3. The method according to any one of the preceding claims, wherein, The mass windows associated with continuous mass spectrometry measurements of the target ion do not overlap.

4. The method according to any one of the preceding claims, wherein, In a single IM scan, the regions covered in the m / z-IM plane are arranged approximately diagonally.

5. The method according to any one of the preceding claims, wherein, Preferably, for each target and decoy peptide precursor in the spectral library, ion traces (421) are extracted from the 5D dataset (207) in the form of fully extracted ion chromatography (XIC) by the following operation: in the RT dimension, iteration is performed on all 5D IM scans measured for a given ion within the range of interest of a given RT, from the first scan to the last scan, wherein, during the iterative process, for each 5D IM scan, the first module (431) performs the operation preferably intended for the 5D dataset. The starting index of the peak to be iterated is determined by a binary search of the lower bound of the m / z window of the fragment ions of the peaks pre-sorted by m / z in the IM scan (401) until the last peak in the array is encountered or a peak whose m / z is higher than the upper bound of the m / z window (410) is encountered. In each iteration step, module (433) checks whether the IM index (403) of the current peak falls within the IM window (412) and whether the m / z of the precursor peptide (408) is within the isolation range of the microscan in which the peak is measured. If yes, the intensity of the peak is added to the run and (434) and the peak index (435) is then incremented. If no, the peak index (435) is incremented directly and then returned to the first module (432) and repeated until the iteration is complete.

6. The method according to claim 5, wherein, The process includes the following steps: ensuring that peaks measured only in a specific DIA window are used to generate extracted ion chromatograms (XIC) in the RT dimension.

7. The method according to any one of the preceding claims, wherein, Preferably, for each target and bait peptide precursor in the spectral library, ion traces (421) are extracted from the 5D dataset (207) in the form of a complete mobility map (302) by: finding an IM scan corresponding to a specific RT for which the mobility map is to be created, preferably this is done for the vertex RT point of the extracted ion chromatogram, wherein, in the obtained 5D IM scan, a preferably binary search is performed to find a peak index that is closest to but still within the lower bound of the m / z window (410), and then the exit condition of the current peak index is checked, wherein if the current peak index is greater than the 5D If the index of the last peak in the IM scan, or the m / z of the peak at the current peak index, is not within the m / z window (410), then the exit condition is true. If the exit condition is not met, then it is checked whether the IM index (403) of the current peak falls within the IM window (412) and whether the m / z of the precursor peptide (408) is within the isolation range (407) of the microscan in which the peak is measured. If true, then the intensity value of the current peak is added to the mobility map array (445) of the IM index of the current peak, and then the current peak index is incremented (446). If false, then the current peak index is incremented, and the exit condition is checked again for the given new peak index.

8. The method according to any one of the preceding claims, wherein, Preferably, for each target and bait peptide precursor in the spectral library, ion traces (421) are extracted from the 5D dataset (207) in the form of a partial mobility map (302) by: finding an IM scan corresponding to a specific RT for which the mobility map is to be created, preferably this is done for the vertex RT point of the extracted ion chromatogram, wherein, in the obtained 5D IM scan, a preferably binary search is performed to find a peak index that is closest to but still within the lower bound of the m / z window (410), and then the exit condition of the current peak index is checked, wherein if the current peak index is greater than the 5D If the index of the last peak in the IM scan, or the m / z of the peak at the current peak index, is not within the m / z window (410), then the exit condition is true. If the exit condition is not met, then it is checked whether the IM index (403) of the current peak falls within the IM window (412), and whether the m / z of the precursor peptide (408) is within the isolation range (407) of the microscan in which the peak was measured. If true, then the intensity value of the current peak is added to the mobility map array (445) of the IM index of the current peak, and then the current peak index is incremented (446). If false, then the current peak index is incremented, and the exit condition is checked again for the given new peak index. The process includes the following steps: ensuring that only peaks measured in a specific DIA window are used to generate the mobility map.

9. The method according to any one of the preceding claims, wherein, For the analysis of the 5D dataset (207), data is loaded in virtual batches with arbitrary m / z widths.

10. The method according to any one of the preceding claims, wherein, In LC-MS / MS experiments, the data are a set of data obtained from a sample that is independent of the sample acquisition process, and the sample is preferably a digested proteome sample.

11. The method according to any one of the preceding claims, wherein, The data is in the form of sample mass spectrometry intensity data acquired using LC tandem mass spectrometry, preferably LC-DIA, with mass-to-charge ratio (m / z), intensity, and ion mobility index (IM). Preferably, the data is a set of data acquired independently from the sample in an LC-MS / MS experiment. The sample is a complex mixture of at least one protein of interest and other proteins and / or other biomolecules, in the form of a complex natural biological matrix that has been degraded prior to LC-MS / MS analysis. The sample is also a complex mixture of at least one protein of interest and other proteins and / or other biomolecules, in the form of a complex natural biological matrix that has been degraded prior to LC-MS / MS analysis. The sample is also a complex mixture of at least one protein of interest, or a protein of interest that is based solely on protein amino acids or carries post-translational modifications.

12. The method according to any one of the preceding claims, wherein, The ion mobility separator is a TIMS analyzer, preferably a TIMS analyzer with parallel accumulation and separation, and is particularly operated using a method comprising the following steps: (a) accumulating ions in an RF ion trap; (b) transferring at least a subset of the accumulated ions to a captured ion mobility separator, wherein the transferred ions are radially confined by an RF field and pushed by a gas flow against the rising edge of an axial DC electric field barrier, such that the transferred ions are spatially separated along the rising edge according to their ion mobility; (c) continuously releasing the transferred ions according to their ion mobility by lowering the height of the DC electric field barrier, while ions from the ion source are further accumulated in the RF ion trap; and (d) restoring the height of the DC electric field barrier, triggering a continuous transfer of the accumulated ions from the RF ion trap to the captured ion mobility separator.

13. The method according to any one of the preceding claims, wherein, At the start of an LC observation retention time window, preferably within a range of 1 to 15 seconds, particularly preferably within a range of 3 to 10 seconds, a full scan is performed, wherein, in this full scan, the full ion mobility width and the full m / z width of interest are scanned, and wherein, based on this full scan of the remaining portion of the LC observation window, a second ion mobility separator and the mass filter are synchronously controlled to perform multiple IM scans, during which precursor ions that increase or decrease IM are successively released from the ion mobility separator, and during this period, the mass window of the mass filter is continuously or gradually moved toward lower or higher m / z values, respectively, to avoid identifying peptides of no interest in the full scan, and wherein the step of associating detected fragments with their corresponding precursor ions is based on determining or utilizing the corresponding mass window and IM range associated with each occurrence of the fragment in the mass spectrometry measurement.

14. Use of a method according to any one of the preceding claims for determining at least one of the compositions of a sample including quantitative information about the components or a medically relevant conformation of said components; for determining a protein-based drug or its effect; for the effect of a drug or other ligand on a protein; or for quality control of a protein-based pharmaceutical preparation.

15. A computer program product for analyzing data using the method according to any one of claims 1 to 13, or a computer-readable medium having such a computer program product stored thereon.

Images