Computer-implemented method of analyzing a tween mixture

Abstract

The application provides a computer-implemented analysis method of a Tween mixture for determining the type of Tween components contained in the Tween mixture, comprising the following steps: obtaining multiple sets of mass spectrum data of a Tween mixture sample under different dissociation conditions; extracting multiple sets of three-dimensional feature components from the multiple sets of mass spectrum data according to a mass-to-charge ratio range set by a user, and performing cluster analysis on the multiple sets of three-dimensional feature components corresponding to each Tween component; displaying a three-dimensional graph constructed based on the multiple sets of three-dimensional feature components, and displaying a point corresponding to a set of three-dimensional feature components clustered as corresponding to the same Tween component in the three-dimensional graph, and displaying a point cluster corresponding to each Tween component in the three-dimensional graph; and analyzing the type of Tween components contained in the Tween mixture sample based on a mass-to-charge ratio range corresponding to the three-dimensional graph confirmed by the user. Through the above method, even a user without data analysis capability can conveniently determine the components of the Tween mixture compounds.

Description

Technical Field

[0001] This invention relates to the field of analytical methods for polymer solvent compounds, specifically to a computer-implemented analytical method for Tween mixtures. Background Technology

[0002] Polysorbates, also known as Tween, are a series of high molecular weight compounds formed from the partial fatty acid esters of polyoxyethylene sorbitol. Based on their molecular structure, polysorbates can be classified into Tween 20 and Tween 21, Tween 40, Tween 60 and Tween 61, Tween 80, Tween 81 and Tween 85, etc. As a type of nonionic surfactant, polysorbates are widely used in food, cosmetics, and pharmaceuticals. In the food industry, polysorbates are mainly used as stabilizers, emulsifiers, and defoamers (GB2760-2014 Standard for the Use of Food Additives) in bread, cakes, cream, ice cream, and baked goods. In the pharmaceutical industry, they are used as dissolving agents for antitumor drugs and for traditional Chinese medicine injections. The 2015 edition of the Chinese Pharmacopoeia, Part II, includes Tween 20, 40, 60, and 80 as pharmaceutical excipients. In the cosmetics industry, it is widely used as a dispersant, emulsifier, thickener, etc. in lotions, creams, and ointments.

[0003] There is currently a need for analysis of Tween mixtures, specifically to determine the type of Tween contained in mixtures containing multiple substances such as Tween 20 and Tween 40. Existing Tween analysis methods employ mass spectrometry to analyze samples. After sample pretreatment, data are collected using a cation linear model, and data analysis is performed using eMSTAT and BioNumberics.

[0004] While general-purpose software such as eMSTAT and BioNumberics can analyze the mass spectra of Tween mixtures, the complex composition of Tween mixtures, formed by the mixing of multiple Tween molecules with varying degrees of polymerization, esterification, and impurity reactions, necessitates additional component identification and similarity comparison. Furthermore, existing general-purpose software is not specifically designed for the analysis of Tween mixtures and struggles to adapt to their unique characteristics. Users lacking expertise in Tween compound data post-processing may find it difficult to perform data analysis using such general-purpose software. Summary of the Invention

[0005] To address the above problems, this invention provides a computer-implemented method for analyzing Tween mixtures. Even if the user lacks comprehensive expertise in analyzing relevant polymer mass spectrometry data, the computer can interact seamlessly with the user by executing this method, ultimately accurately determining the type of Tween contained in the Tween mixture. This solves the technical problems of existing Tween mixture analysis interfaces being unfriendly to users and having low analytical accuracy.

[0006] This invention provides a computer-implemented analytical method for Tween mixtures to determine the types of Tween components contained in the mixture, comprising the following steps:

[0007] Obtain multiple sets of mass spectrometry data for Tween mixture samples under different dissociation conditions;

[0008] Based on the user-defined mass-to-charge ratio range, multiple sets of three-dimensional feature components are extracted from multiple sets of mass spectrometry data, and cluster analysis is performed on the multiple sets of three-dimensional feature components corresponding to each Tween component.

[0009] The display shows a three-dimensional map constructed based on multiple sets of three-dimensional feature components. Each set of three-dimensional feature components corresponding to the same Tween component is displayed as a point in the three-dimensional map. The three-dimensional map distinguishes and displays the point clusters corresponding to each Tween component.

[0010] Based on the mass-to-charge ratio range corresponding to the 3D plot confirmed by the user, the types of Tween components contained in the Tween mixture sample are analyzed.

[0011] Because the process of extracting three-dimensional feature components is controlled or supervised by the user-defined mass-to-charge ratio range, and the mass-to-charge ratio range or mass number range of Tween components can usually be quickly determined in advance through other measurement methods, the inventors discovered that by pre-setting the mass-to-charge ratio range, the computational load required for extracting three-dimensional features or performing cluster analysis can be effectively reduced, significantly decreasing processing time and effectively improving the accuracy of the analysis results. Furthermore, presenting the results to the user in a three-dimensional graph not only makes the interaction more intuitive and convenient but also allows the user to observe the cluster analysis results of each point in the three-dimensional graph from various angles through dragging, rotating, translating, and zooming, facilitating a quick judgment of the cluster analysis results and improving the overall efficiency of the analysis method. In addition, combined with the type analysis of the mass-to-charge ratio range, the type judgment result of the Tween mixture sample can be given more accurately and quickly, significantly reducing the required computational load.

[0012] Through the above methods, this invention provides an analysis method that enables simpler and more efficient interaction with users. During operation, users only need to input the mass-to-charge ratio range and judge the clustering analysis results based on the dispersion of points. This provides an intuitive basis for the supervised clustering analysis algorithm, delivering more accurate analysis results with lower computational load. The entire interaction process is intuitive, user-friendly, and has high analysis accuracy.

[0013] The optional technical solution of the present invention further includes the following steps:

[0014] Display the clusters of impurity points corresponding to multiple sets of unclustered 3D feature components in a 3D plot;

[0015] In response to at least a portion of the impurity point clusters being marked as target compounds or target mixtures, multiple sets of three-dimensional feature components corresponding to the impurity point clusters are associated with the target compounds or target mixtures and saved as impurity point cluster data; subsequently...

[0016] When analyzing newly collected Tween mixture samples, multiple sets of un-clustered three-dimensional feature component data are compared with impurity point cluster data to determine the presence of the target compound or target mixture or to analyze its concentration.

[0017] The above methods can also be used to analyze impurities in Tween mixtures. Impurities in Tween mixtures will also show peaks in the spectrum, but they are difficult to accurately identify by cluster analysis and can easily interfere with the results. Furthermore, by pre-establishing and saving information on relevant impurities, these points can be treated as impurity cluster data, distinguishing them from other measurement points. This can reduce the interference of impurities on the qualitative or quantitative analysis of Tween components and, under specific conditions, enable qualitative or quantitative analysis of impurities in Tween mixtures.

[0018] The optional technical solution of the present invention further includes the following steps:

[0019] The accuracy of matching the point cluster with the Tween component is calculated based on the degree of dispersion of the point cluster.

[0020] According to this technical solution, a larger degree of dispersion indicates a lower probability that the cluster matches the Tween component, while a smaller degree of dispersion indicates a higher probability that the cluster matches the Tween component. This discrimination method can more accurately determine the type of Tween mixture sample corresponding to the cluster.

[0021] In an optional technical solution of the present invention, the Tween mixture sample is one or more of Tween 20, Tween 40, Tween 60 and Tween 80.

[0022] In an optional technical solution of the present invention, the mass spectrometry data are obtained by matrix-assisted laser dissociation-time-of-flight mass spectrometry.

[0023] The optional technical solution of the present invention further includes the following steps:

[0024] Provides a Tween standard mass spectrometry dataset constructed from various Tween 20, Tween 40, Tween 60 and Tween 80 standard samples.

[0025] According to this technical solution, by comparing the Tween standard mass spectrometry dataset established from multiple Tween standard samples with the Tween components to be analyzed, the Tween type can be determined, thus improving the convenience of Tween component analysis.

[0026] In the optional technical solution of the present invention, the step of extracting multiple sets of three-dimensional feature components from multiple sets of mass spectrometry data is implemented based on the principal component analysis algorithm.

[0027] According to this technical solution, the principal component analysis algorithm transforms many originally highly correlated data into mutually independent or uncorrelated data, replacing the originally complex mass spectrometry data with more simplified data, thereby improving the speed and accuracy of subsequent cluster analysis and facilitating intuitive display in the form of three-dimensional graphs.

[0028] In the optional technical solution of the present invention, the step of performing cluster analysis on multiple sets of three-dimensional feature components corresponding to each Tween component is implemented based on the K-means algorithm. Attached Figure Description

[0029] Figure 1 This is a flowchart of a computer-implemented analysis method for Tween mixtures according to the first embodiment of the present invention.

[0030] Figure 2 This is a schematic diagram of the cluster analysis results of Tween mixtures corresponding to the molecular weight range of 1000-2500 in this invention.

[0031] Figure 3 This is a schematic diagram of the cluster analysis results of Tween mixtures corresponding to the molecular weight range of 2000-2500 in this invention.

[0032] Figure 4 This is a schematic diagram of the cluster analysis results of Tween mixtures corresponding to the molecular weight range of 3000-5000 in this invention.

[0033] Figure 5 This is a flowchart of a computer-implemented analysis method for Tween mixtures according to the second embodiment of the present invention.

[0034] Figure 6 This is a flowchart of a computer-implemented analysis method for Tween mixtures according to the third embodiment of the present invention.

[0035] Figure 7 This is a flowchart of a computer-implemented analysis method for Tween mixtures according to the fourth embodiment of the present invention. Detailed Implementation

[0036] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0037] [First Implementation Method]

[0038] like Figure 1 The diagram shown is a flowchart of a computer-implemented analysis method for Tween mixtures according to the first embodiment of the present invention. Please refer to [link / reference]. Figure 1 This invention provides a computer-implemented method for analyzing Tween mixtures to determine the type of Tween mixture, comprising the following steps:

[0039] Step S1: Obtain multiple sets of mass spectrometry data of the Tween mixture sample under different dissociation conditions;

[0040] Step S2: Based on the mass-to-charge ratio range set by the user, extract multiple sets of three-dimensional feature components from multiple sets of mass spectrometry data, and perform cluster analysis on the multiple sets of three-dimensional feature components corresponding to each Tween component.

[0041] Step S3: Display a three-dimensional map constructed based on multiple sets of three-dimensional feature components. A set of three-dimensional feature components clustered to correspond to the same Tween component will be displayed as a point in the three-dimensional map. The point clusters corresponding to each Tween component will be displayed in the three-dimensional map.

[0042] Step S4: Analyze the type of Tween mixture sample based on the mass-to-charge ratio range corresponding to the 3D plot confirmed by the user.

[0043] This invention provides a computer-implemented method for analyzing Tween mixtures. After completing mass spectrometry analysis, users do not need to perform data analysis themselves. Through simple data post-processing operations implemented by the computer, the type of Tween mixture sample can be obtained. Even users who do not have professional knowledge of complex mass spectrometry data post-processing for Tween mixtures can perform Tween mixture analysis, thus reducing the difficulty of Tween mixture analysis.

[0044] Furthermore, since the extraction of three-dimensional feature components is controlled or supervised by the user-defined mass-to-charge ratio range, and the mass-to-charge ratio range (i.e., mass number range) of Tween components can usually be quickly determined in advance through other measurement methods, the inventors discovered that by pre-setting the mass-to-charge ratio range, the computational load required for three-dimensional feature extraction or cluster analysis can be effectively reduced, significantly decreasing processing time and effectively improving the accuracy of the analysis results. Additionally, presenting the results to the user in a three-dimensional graph not only makes the interaction more intuitive and convenient but also allows the user to observe the cluster analysis results of each point in the three-dimensional graph from various angles through dragging, rotating, translating, and zooming. This facilitates the user's rapid judgment of the cluster analysis results based on the dispersion of points, improving the overall efficiency of the analysis method. Furthermore, combined with the type analysis of the mass-to-charge ratio range, the type judgment result of the Tween mixture sample can be given more accurately and quickly, significantly reducing the required computational load.

[0045] Specifically, in this embodiment of the invention, the Tween mixture sample is one or more of Tween 20, Tween 40, Tween 60 and Tween 80.

[0046] The above methods enable component analysis of samples containing one or more of Tween 20, Tween 40, Tween 60, and Tween 80, thus expanding the applicability of Tween mixture sample analysis.

[0047] In this embodiment of the invention, the mass spectrometry data includes data such as mass-to-charge ratio, ion current intensity, and dissociation conditions. The dissociation conditions can be any physical quantity that can change the degree of sample dissociation, such as laser intensity or electric field intensity. By performing mass spectrometry analysis on the Tween mixture, multiple sets of mass spectrometry data corresponding to the Tween mixture under different dissociation conditions are obtained, resulting in multiple mass spectra.

[0048] In a preferred embodiment of the present invention, the mass spectrometry data are obtained by matrix-assisted laser dissociation-time-of-flight mass spectrometry.

[0049] In this embodiment of the invention, the user can set multiple mass-to-charge ratio ranges and select a suitable mass-to-charge ratio range based on the feedback from cluster analysis results under different mass-to-charge ratio ranges. In a specific embodiment of the invention, three mass-to-charge ratio ranges are set. Based on the one-to-one correspondence between mass-to-charge ratio and molecular weight (mass number), the molecular weight ranges corresponding to the three set mass-to-charge ratio ranges are 1000-2500, 2000-2500, and 3000-5000, respectively. Figure 2 , Figure 3 and Figure 4 These are schematic diagrams showing the cluster analysis results of Tween mixture samples corresponding to the above three molecular weight ranges. Figure 3As can be seen, within the molecular weight range of 1000–2500Th, the dispersion between different point clusters is relatively large, while the dispersion within the same point cluster is relatively small. This means that different point clusters represent different sample components, and the large dispersion between different point clusters indicates a low correlation between them. Conversely, the small dispersion within the same point cluster leads to a high accuracy rate in identifying it as the same sample component. In contrast,… Figure 3 In the three-dimensional graph corresponding to a molecular weight range of 3000-5000, the data points are not only highly discrete in the overall space, but also quite scattered in some relatively concentrated areas, making it impossible to achieve good clustering analysis results.

[0050] Furthermore, the clustering analysis of multiple sets of three-dimensional feature components corresponding to each Tween component is implemented based on the k-means algorithm. The k-means algorithm is a centroid-based method for partitioning data. Its basic idea is to randomly select k points as cluster centers; calculate the clusters from each point to the k cluster centers, and then assign the point to the nearest cluster center, thus forming k clusters; then recalculate the centroid (mean) of each cluster; repeat the above steps until the position of the centroid no longer changes or the set number of iterations is reached, thus obtaining the clustering analysis result. This invention, by providing multiple sets of three-dimensional feature components and the number of clusters to be divided (e.g., Tween 20, Tween 40, Tween 60, and Tween 80, four clusters), can use this algorithm to divide the dataset into k clusters, each cluster corresponding to a type of Tween component.

[0051] In a preferred embodiment of the present invention, the step of extracting multiple sets of three-dimensional feature components from multiple sets of mass spectrometry data is implemented based on principal component analysis (PCA) algorithm.

[0052] Principal component analysis (PCA) uses linear transformations to select a smaller number of important variables from multiple variables, attempting to recombine the original variables into a new set of independent composite variables. Simultaneously, depending on practical needs, a smaller number of composite variables can be extracted to reflect as much information as possible from the original variables. PCA can also achieve dimensionality reduction of multiple sets of mass spectrometry data, converting multiple two-dimensional mass spectra under different dissociation conditions into three-dimensional feature components. Given the complexity of Tween mixtures, the PCA algorithm can significantly reduce the amount of mass spectrometry data, decreasing the workload of data analysis. By extracting the three-dimensional feature components of the mass spectrometry data and excluding duplicate data, it helps reduce the workload of cluster analysis and improves the accuracy of cluster analysis results. In a specific embodiment of this invention, the specific formula for cluster analysis based on the PCA algorithm combined with the K-means algorithm is shown below.

[0053]

[0054] In the formula X np , T, μ i S represents the p-dimensional matrix eigencomponents, transformation matrix, clustering, and cluster combination of multiple sets of mass spectrometry data, respectively; I represents the identity matrix; and X is a matrix. T Let X be the transpose of X, w be the weight, λ be the eigenvalue, min be the minimum value, max be the maximum value, arg be the marker, det be the determinant, var be the variance, and tr be the trace.

[0055] In a specific embodiment of the present invention, the three-dimensional feature components and clustering results of the Tween mixture samples obtained based on principal component analysis and K-means algorithm are shown in Table 1 below.

[0056] Table 1. Examples of results based on PCA and K-means algorithms.

[0057]

[0058] In Table 1, columns d1, d2, and d3 represent the feature components transformed into a three-dimensional vector space using Principal Component Analysis (PCA) algorithm. The second column represents the clustering results based on the k-means algorithm. These data validate the effectiveness of the algorithm.

[0059] [Second Implementation Method]

[0060] Figure 5 This is a flowchart of a computer-implemented analysis method for Tween mixtures according to the second embodiment of the present invention. Please refer to... Figure 5 The second embodiment of the present invention provides a computer-implemented method for analyzing Tween mixtures, which differs from the first embodiment in that it further includes step S5:

[0061] Display the clusters of impurity points corresponding to multiple sets of unclustered 3D feature components in a 3D plot;

[0062] In response to at least a portion of the impurity point clusters being marked as target compounds or target mixtures, multiple sets of three-dimensional feature components corresponding to the impurity point clusters are associated with the target compounds or target mixtures and saved as impurity point cluster data; subsequently...

[0063] When analyzing newly collected Tween mixture samples, multiple sets of un-clustered three-dimensional feature component data are compared with impurity point cluster data to determine the presence of the target compound or target mixture or to analyze its concentration.

[0064] The above methods can also be used to analyze impurities in Tween mixtures. Impurities in Tween mixtures will also show peaks in the spectrum, but they are difficult to accurately identify by cluster analysis and can easily interfere with the results. Furthermore, by pre-establishing and saving information on relevant impurities, these points can be treated as impurity cluster data, distinguishing them from other measurement points. This can reduce the interference of impurities on the qualitative or quantitative analysis of Tween components and, under specific conditions, enable qualitative or quantitative analysis of impurities in Tween mixtures.

[0065] [Third Implementation Method]

[0066] Figure 6 This is a flowchart of a computer-implemented analysis method for Tween mixtures according to the third embodiment of the present invention. Please refer to... Figure 6 The second embodiment of the present invention provides a computer-implemented method for analyzing Tween mixtures, which differs from the first embodiment in that it further includes step S6:

[0067] The accuracy of matching the point cluster with the Tween component is calculated based on the degree of dispersion of the point cluster.

[0068] Using the above method, a larger degree of dispersion indicates a lower probability that the point cluster matches the Tween component, while a smaller degree of dispersion indicates a higher probability that the point cluster matches the Tween component. This discrimination method can more accurately determine the similarity calculation result of the Tween component corresponding to the point cluster.

[0069] Specifically, through the following formula:

[0070]

[0071] The accuracy of matching the point cluster with the Tween components can be calculated, where d and σ represent the distance from the sample (i.e., a point corresponding to a set of three-dimensional feature components displayed in a three-dimensional image) to the center of the point cluster and the maximum boundary radius of the point cluster data, respectively. In a specific embodiment of the present invention, the mass spectrometry data of 10 samples (T 80_1, T 80_2, T 20_1, T 20_2, AGTuningMix, NaI, T60_1, T 60_2, T 40_1, T 40_2) were clustered using PCA and K-means algorithms, and the classification results of the experimental data are shown in Table 2 below.

[0072] Table 2 Probability Table of Experimental Data

[0073] Sample Labels T20 T40 T60 T80 T 80_1 3.3372 0.0149 23.9157 81.2057 T 80_2 2.683 0.0036 7.8071 95.7658 T 20_1 95.3856 0 0.7389 0.6794 T 20_2 68.3187 0 0.688 0.8395 AGTuningMix 34.0147 9.7132 13.8437 11.5396 NaI 15.2838 31.9847 21.7744 20.447 T 60_1 15.3198 1.2262 60.6531 8.2995 T 60_2 11.4006 0.2046 74.9755 8.654 T 40_1 10.2451 84.4164 11.9117 4.8743 T 40_2 12.7078 72.3025 11.7222 4.9083

[0074] As shown in Table 2, the first column represents the label of the analyzed object, and the other four columns use percentages to represent the probability (relative value) of whether the analyzed object belongs to a specific Tween category. It can be seen that the probability of T80-1 being clustered as Tween 80 is 81.2057, the probability of T80-2 being clustered as Tween 80 is 95.7658, and the probability of T20-1 being clustered as Tween 20 is 95.3856. The results show that the model can classify Tween well. When inputting other samples completely different from Tween (such as NaI), the probability of the cluster matching the Tween component is low and will not exceed a certain threshold, such as a 60% threshold.

[0075] [Fourth Implementation Method]

[0076] Figure 7 This is a flowchart of a computer-implemented analysis method for Tween mixtures according to the fourth embodiment of the present invention. Please refer to... Figure 7 The fourth embodiment of the present invention provides a computer-implemented method for analyzing Tween mixtures, which differs from the first, second, and third embodiments in that it further includes the following step S7:

[0077] Provides a Tween standard mass spectrometry dataset constructed from Tween 20, Tween 40, Tween 60 and Tween 80 standard samples.

[0078] By comparing the Tween standard mass spectrometry dataset established from various Tween standard samples with the Tween components to be analyzed using the above method, the Tween type can be determined, improving the convenience and accuracy of Tween component analysis.

[0079] Industrial applicability

[0080] The analytical method for Tween mixtures provided by this invention can be used for qualitative identification and semi-quantitative analysis of Tween mixtures, and is particularly suitable for qualitative identification or semi-quantitative analysis of Tween mixtures used as pharmaceutical excipients.

[0081] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A computer-implemented method of analysis of a Tween mixture for determining the type of Tween component comprised by the Tween mixture, characterized in that, Includes the following steps: Obtain multiple sets of mass spectrometry data for Tween mixture samples under different dissociation conditions; Based on the mass-to-charge ratio range set by the user, multiple sets of three-dimensional feature components are extracted from the multiple sets of mass spectrometry data, and cluster analysis is performed on the multiple sets of three-dimensional feature components corresponding to each Tween component. The three-dimensional map is constructed based on the multiple sets of three-dimensional feature components. A point is displayed in the three-dimensional map corresponding to a set of three-dimensional feature components that are clustered with the same Tween component. The three-dimensional map distinguishes and displays the point clusters corresponding to each Tween component. Based on the mass-to-charge ratio range corresponding to the 3D plot confirmed by the user, the types of Tween components contained in the Tween mixture sample are analyzed. Display the clusters of impurity points corresponding to multiple sets of unclustered 3D feature components in a 3D plot; In response to at least a portion of the impurity point clusters being marked as target compounds or target mixtures, multiple sets of three-dimensional feature components corresponding to the impurity point clusters are associated with the target compounds or target mixtures and saved as impurity point cluster data. after, When analyzing a re-collected Tween mixture sample, multiple sets of un-clustered three-dimensional feature component data are compared with the impurity point cluster data to determine whether the target compound or the target mixture exists or to analyze its concentration.

2. The method of analysis of Tween mixtures according to claim 1, characterized in that, It also includes the following steps: The accuracy of matching the point cluster with the Tween component is calculated based on the degree of dispersion of the point cluster.

3. The method of analysis of Tween mixtures as claimed in claim 1, wherein, The Tween mixture sample contains one or more of Tween 20, Tween 40, Tween 60, and Tween 80.

4. The method of analysis of Tween mixtures as claimed in claim 1, wherein, The mass spectrometry data were obtained using a matrix-assisted laser dissociation-time-of-flight mass spectrometer.

5. The method of analysis of Tween mixtures as claimed in claim 1, wherein, It also includes the following steps: Provides a Tween standard mass spectrometry dataset constructed from Tween 20, Tween 40, Tween 60 and Tween 80 standard samples.

6. The method of analysis of Tween mixtures as claimed in claim 1, wherein, The step of extracting multiple sets of three-dimensional feature components from the multiple sets of mass spectrometry data is implemented based on the principal component analysis algorithm.

7. The method of analysis of Tween mixtures as claimed in claim 1, wherein, The step of performing cluster analysis on multiple sets of three-dimensional feature components corresponding to each Tween component is implemented based on the K-means algorithm.

Images