Skin rapid detection and intelligent evaluation system based on mass spectrometry and its application

Abstract

The present invention provides a skin rapid detection and intelligent evaluation system and application based on mass spectrometry, which belongs to the field of biological detection technology. The skin rapid detection and intelligent evaluation system includes: sample pretreatment, UPLC-Q-TOF-MS separation and analysis, and intelligent analysis and identification system. The skin rapid detection and intelligent evaluation system provided by the present invention integrates the skin ceramide lipid tape stripping method specific extraction technique, a new ceramide analysis method based on UPLC-Q-TOF-MS, and a self-built ceramide intelligent analysis algorithm based on Python programming. The skin rapid detection and intelligent evaluation system provided by the present invention can quickly and accurately identify the ceramides in the skin, and by comparing the changes in the content of ceramides, the effects of cosmetics or drugs on skin ceramides can be effectively evaluated, providing technical support for the prevention and improvement of skin problems, and the realization of precise skin care and the prevention and treatment of skin diseases.

Description

Technical Field

[0001] The present invention belongs to the technical field of biological detection, and in particular relates to a rapid skin detection and intelligent evaluation system based on mass spectrometry and its application. Background Art

[0002] Due to the importance of ceramides in maintaining skin condition and exerting the efficacy of cosmetics, it has become a new hot spot in related research fields. On the technical level, liquid chromatography-mass spectrometry (LC-MS) technology is currently the best technology for simultaneous detection of ceramides in the body due to its high separation, high selectivity, high sensitivity and high throughput analysis. The development and application of skin ceramide LC-MS analysis related technologies and industries at home and abroad are summarized as follows.

[0003] First, the specific extraction of ceramides on the skin surface is the first step in the ceramide detection process. Zhang Xuejun et al. (Identification of types of ceramides and other lipids in the human stratum corneum using high-performance liquid chromatography-tandem high-resolution mass spectrometry, Chinese Journal of Dermatology and Venereology, June 2018, Vol. 32, No. 6, pp. 718-723) used cyanoacrylate glue to apply a thin layer on one end of the slide to allow the glue to fully contact the skin, and then used a methanol-chloroform mixture for extraction, but this method is obviously not as easy to operate as stickers, tapes, etc. The MTBE method of extraction places ceramides in the upper solvent, so it is easy to collect, has less protein interference, and is simple and fast to operate, but its strong solubility will dissolve the adhesive on the tape, thereby interfering with the mass spectrometry signal of ceramides.

[0004] Secondly, ultra-high performance liquid chromatography-triple quadrupole tandem mass spectrometry (UPLC-QQQ-MS / MS) is the current mainstream analysis technology for ceramides. Although UPLC has a higher separation ability than conventional LC, QQQ has low mass resolution and weak unknown qualitative ability, which makes the comprehensiveness and accuracy of ceramide identification insufficient, and it is heavily dependent on expensive commercial standards, which greatly increases the detection cost. In recent years, UPLC combined with high-resolution tandem mass spectrometry technologies such as quadrupole time-of-flight mass spectrometry (Q-TOF-MS) or electrostatic field orbital mass spectrometry (orbitrap-MS) has been gradually used for ceramide analysis. Compared with low-resolution QQQ, these high-resolution mass spectrometers are slightly less sensitive, but with their primary accurate mass determination and secondary fragment analysis capabilities, they can significantly improve the identification quantity and qualitative accuracy of ceramides in the body. Although ceramide analysis has gradually entered the era of UPLC and high-resolution tandem mass spectrometry, it is not widely used in the field of skin. Patent CN 116500180 A proposes a method for predicting and identifying ceramide chromatographic characteristics based on LC-MS / MS, based on the existing standard information and the prediction model of the target molecule XlogP, combined with MS / MS information to verify the qualitative nature of the structure. This patent relies on ceramide standards, otherwise it would be difficult to proceed, but the number of standards is small and cannot cover all ceramide categories that need to be identified; another drawback is that it is bound to the experimental method, which limits the universality of the method.

[0005] Finally, ceramides are of various types and complex structures. Currently, there are many lipid databases or related mass spectrometry software for their automatic mass spectrometry analysis. For example, the LipidBlast lipid database contains the most complete lipid mass spectrometry information among all open source databases, but the number of ceramide types and their secondary mass spectra is still far less than that of other common phospholipids and glycerides. Major mass spectrometry instrument manufacturers have also launched commercial lipid software, such as Lipid Search, Lipid View, and Lipid Annotator, but the coverage of ceramide information is still not complete. In addition, since most conventional analyses only use the above-mentioned lipid databases and software for matching, relying solely on the ceramide mass spectrometry fragmentation law for qualitative analysis will inevitably lead to a high false positive rate, reducing the reliability of the identification results. Therefore, it is necessary to consider adding multivariate parameters (such as retention time, carbon number of fatty acid chain, and degree of unsaturation) to assist in the qualitative analysis of ceramides, but this function is rarely seen in existing commercial or open source lipid software. Summary of the invention

[0006] In view of the above-mentioned deficiencies, the present invention provides a skin rapid detection and intelligent evaluation system and application based on mass spectrometry. The evaluation system includes: sample pretreatment, UPLC-Q-TOF-MS separation and analysis, and intelligent analysis and identification system. The skin rapid detection and intelligent evaluation system provided by the present invention integrates the skin ceramide lipid tape stripping method specific extraction, a new ceramide analysis method based on UPLC-Q-TOF-MS, and a self-built ceramide intelligent analysis algorithm based on Python programming. The skin rapid detection and intelligent evaluation system provided by the present invention can quickly and accurately identify the ceramides in the skin, and compare the changes in the ceramide content, thereby effectively evaluating the effects of cosmetics or drugs on skin ceramides, and providing technical support for the prevention and improvement of skin problems, the realization of precise skin care, and the prevention and treatment of skin diseases.

[0007] The present invention compares the extraction effect of ceramide on the tape used for skin adhesion to the stratum corneum on the market, obtains the best lipid tape, and optimizes the composition and ratio of the extraction solvent, realizes the specificity, non-destructive extraction and efficient enrichment of skin ceramide, and is the most complete pre-treatment technology for skin ceramide at present. The ceramide mass spectrometry database provided by the present invention contains the most comprehensive and complete ceramide molecular formula on the market and the most flexible and specific secondary characteristic fragment ion prediction function, and has good structural qualitative capabilities for all subclasses of skin ceramides, and can efficiently and quickly realize its comprehensive identification and content comparative analysis, which greatly improves the depth and breadth of ceramide research. The ceramide intelligent analysis algorithm provided by the present invention also has a retention time law auxiliary qualitative function that is not available in commercial / open source lipid databases or related mass spectrometry software on the market, which can greatly improve the depth and accuracy of skin ceramide identification, reaching the international advanced level.

[0008] To achieve the above-mentioned object of the invention, the technical solution of the present invention is as follows:

[0009] In a first aspect, the present invention provides a skin condition evaluation system, which includes: sample pretreatment, UPLC-Q-TOF-MS separation and analysis, and an intelligent analytical identification system.

[0010] Specifically, the intelligent analysis and identification system comprises the following steps:

[0011] (1) Predict fragment ions based on the target ion mass number to obtain predicted secondary mass spectrometry fragment information;

[0012] (2) Matching the predicted secondary mass spectrometry fragment information with the actual collected secondary mass spectrometry data using a fragment ion library to obtain data that meets the matching requirements;

[0013] (3) Structural annotation of the data with the highest credibility that meets the matching requirements;

[0014] (4) Construct a retention time prediction model based on the linear law between the carbon number of ceramides in the same series and their chromatographic retention time, and train the retention time prediction model to eliminate false detection results;

[0015] (5) Predict ceramide structure based on retention time.

[0016] Preferably, the secondary mass spectrometry fragment information in step (1) includes: [M+H]+, neutral loss of one water molecule, de-FA fatty acid, sphingosine skeleton fragment, and one or more of ceramide characteristic fragments.

[0017] Preferably, the standard of the data meeting the matching requirement in step (2) is that the difference between the predicted m / z value of the fragment and the actual predicted m / z value is within the range of ±10 ppm, which meets the matching requirement.

[0018] Preferably, the highest credibility standard in step (3) is: the highest credibility is achieved when the number of mass spectrum fragments is the largest and the ion current intensity is the largest.

[0019] Preferably, the training retention time prediction model in step (4) is as follows: import RANSACRegressor from the sklearn.linear_model module, and use robust regression to train the retention time prediction model.

[0020] Specifically, the sample pretreatment includes: adsorbing ceramide on a lipid tape, adding an extract, ultrasonicating, oscillating, collecting the upper phase by centrifugation, and re-dissolving after drying.

[0021] More specifically, the resin tape is Corneofix® F20 (Courage+Khazaka GmbH, Germany).

[0022] More specifically, the extract is methanol.

[0023] More specifically, the solution used for the re-dissolution is a mixture of isopropanol and methanol.

[0024] Preferably, in the mixed solution of isopropanol and methanol, the volume ratio of isopropanol to methanol is 1:1-3.

[0025] Further preferably, the volume ratio of isopropanol to methanol is 1:2.

[0026] More specifically, the ultrasound duration is 30-60s.

[0027] Preferably, the ultrasound duration is 30-40s, 40-50s, or 50-60s.

[0028] More specifically, the shaking conditions are: 35-40° C., shaking for 1-4 h.

[0029] Preferably, the shaking conditions are: 37° C., shaking for 2 h.

[0030] More specifically, the centrifugation is performed at 4000-6000 rpm for 5-15 min.

[0031] Preferably, the centrifugation is performed at 5000 rpm for 10 min.

[0032] More specifically, the drying is carried out by nitrogen blowing.

[0033] Preferably, the specific steps of the sample pretreatment are: sticking a first piece of lipid tape that absorbs ceramide on the skin of the subject, discarding it, sticking a second piece of lipid tape that absorbs ceramide on the same skin of the same subject, pressing for 30-60s, adding the extract, sonicating, oscillating, centrifuging to collect the upper phase, and re-dissolving after drying.

[0034] Preferably, the chromatographic parameters in the UPLC-Q-TOF-MS separation analysis are:

[0035] Chromatographic column: Agilent Eclipse Plus C18 reverse phase column, 100 mm × 2.1 mm, 1.8 µm;

[0036] Mobile phase A: methanol-water-formic acid solution containing 10 mM ammonium acetate, wherein the volume ratio of methanol, water and formic acid in the methanol-water-formic acid solution is 60:40:0.2;

[0037] Mobile phase B: methanol-isopropanol-formic acid solution containing 10 mM ammonium acetate, wherein the volume ratio of alcohol, isopropanol and formic acid in the methanol-isopropanol-formic acid solution is 60:40:0.2;

[0038] Flow rate: 0.35 mL / min;

[0039] Gradient elution program:

[0040] 0-3 min, 90%-100% mobile phase A, 0%-10% mobile phase B;

[0041] 3-5 min, 60%-80% mobile phase A, 10%-40% mobile phase B;

[0042] 5-5.3 min, 45%-60% mobile phase A, 40%-55% mobile phase B;

[0043] 5.3-8 min, 40%-45% mobile phase A, 55%-60% mobile phase B;

[0044] 8-8.5 min, 20%-40% mobile phase A, 60%-80% mobile phase B;

[0045] 8.5-10.5 min, 20% mobile phase A, 80% mobile phase B;

[0046] 10.5-16 min, 10%-20% mobile phase A, 80%-90% mobile phase B;

[0047] 16-19 min, 10% mobile phase A, 90% mobile phase B;

[0048] 19-22 min, 0%-10% mobile phase A, 90%-100% mobile phase B;

[0049] 22-26 min, 0% mobile phase A, 100% mobile phase B;

[0050] 26-34 min, 100% mobile phase A, 0% mobile phase B.

[0051] Preferably, the mass spectrometry parameters in the UPLC-Q-TOF-MS separation analysis are:

[0052] The primary mass spectrometer adopted TOF full scan mode and was detected in positive mode; the drying gas flow rate was 10 L / min, the drying gas temperature was 350℃, the nebulizer gas pressure was 30 psi, the capillary voltage was 3500 V, the skimmer voltage was 65 V, the octopole radio frequency voltage was 750 V, the capillary outlet voltage was 150 V, and the mass-to-charge ratio scanning range was m / z 50-1500; two internal standard ions m / z 121.0509 and 922.0098 in the reference solution were selected for real-time mass correction of the instrument measurement results;

[0053] The secondary mass spectrometry adopted the targeted secondary ion mode Targeted-MS / MS. According to the difficulty of target ion fragmentation, the collision voltage was selected as 20 or 40 eV, the acquisition time interval of Targeted-MS / MS was set to 0.4 min, and the quadrupole filter m / z range was set to Medium level.

[0054] In a second aspect, the present invention provides the use of the above-mentioned evaluation system in the identification of skin ceramides or the determination of skin ceramide content.

[0055] In a third aspect, the present invention provides a method for identifying skin ceramide or measuring skin ceramide content, wherein the method comprises using the above-mentioned evaluation system.

[0056] In a fourth aspect, the present invention provides a terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned evaluation system when executing the computer program.

[0057] In a fifth aspect, the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program implements the above-mentioned evaluation system when executed by a processor.

[0058] The beneficial effects of the present invention are:

[0059] (1) Ceramide is composed of a sphingosine skeleton and a fatty acid chain connected by an amide bond. Due to the length of the two chains, the diversity of unsaturation and the various functional groups connected, the types and quantity of ceramides are extremely large. According to the structural rules of different subclasses of ceramide, the present invention generates more than 10,000 discovered and potential theoretical molecular formulas of ceramides through the above-mentioned Python program. The LIPIDMAPS database only contains more than 500 ceramides, and the LipidBlast database only contains more than 1,000. In comparison, the mass spectrometry database we created covers the most complete number and types of ceramides, which will greatly help researchers discover new structures of ceramides.

[0060] (2) The mass spectrometric identification of ceramide is mainly based on the secondary characteristic fragments produced by its sphingosine skeleton. The ceramide secondary fragment ion prediction algorithm developed in the present invention will infer all theoretical fragment ions of each ceramide based on the fragmentation pattern of the reference compound; in the matching algorithm, the measured secondary fragment ion m / z of ceramide and the predicted theoretical fragment ion m / z are matched one by one, making the identification result more accurate and reliable.

[0061] (3) Although all major instrument manufacturers on the market have their own lipid identification software, such as Lipid Search, Lipid Annotator, etc., these software are only applicable to mass spectrometry data collected by our own instruments. The ceramide intelligent analysis algorithm of the present invention is not limited to the data acquisition format and is compatible with mass spectrometry data from all manufacturers' instruments, breaking down the barriers to data analysis between different manufacturers and greatly improving the versatility and convenience of research.

[0062] (4) The retention time parameter is crucial for preventing false positive identification results, missed detection (missing secondary fragment information, unable to identify) and distinguishing isomers, and is an important means to confirm the molecular structure of ceramide. The present invention uses the identified ceramide as a training set to fit the retention time prediction curve, which can be adapted to different chromatographic analysis methods. It has special functions that are not available in the lipid database or related mass spectrometry software on the market, greatly improving the depth and accuracy of skin ceramide identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 Flow chart of the study for the identification of skin ceramides.

[0064] Figure 2 This is the logic diagram of the intelligent analysis algorithm for ceramide mass spectrometry.

[0065] Figure 3 The fragmentation mechanism and theoretical fragments of monohydroxyceramide in the positive mode of mass spectrometry.

[0066] Figure 4 To explain the matching algorithm logic using Cer(m34:2) as an example.

[0067] Figure 5 This is the retention time prediction model training result for the Cer (d18:0 / x:0) series data.

[0068] Figure 6 Comparison of ceramide extraction effects of different tapes

[0069] Figure 7 To compare the extraction effects of ceramide using different extraction methods.

[0070] Figure 8 It is a retention time prediction model for the Cer (d18:2 / x:0) series ceramides.

[0071] Fig. 9 It is a retention time prediction model for the Cer (t18:0 / x:0) series ceramides. DETAILED DESCRIPTION

[0072] The following non-limiting examples can enable those skilled in the art to more fully understand the present invention, but do not limit the present invention in any way. The following content is merely an exemplary description of the scope of the present invention, and those skilled in the art can make various changes and modifications to the present invention based on the disclosed content, which should also fall within the scope of the present invention.

[0073] The present invention is further described below by way of specific examples. The various chemical reagents used in the examples of the present invention are obtained through conventional commercial channels unless otherwise specified. Unless otherwise specified, it is understood that the reaction is carried out at room temperature.

[0074] Example 1 Automatic skin condition evaluation system based on ceramide LC-MS analysis and intelligent analysis program

[0075] The research flow chart of the skin condition automatic evaluation system based on ceramide LC-MS analysis and intelligent parsing program provided by the present invention for skin ceramide identification is shown in Figure 1 The logic diagram of the intelligent analysis algorithm of ceramide mass spectrum of the present invention is shown in Figure 2 .

[0076] 1. Sample pretreatment

[0077] After the subjects washed their faces with water, they sat quietly in a room with a temperature of 22-25℃ and a humidity of 40%-60% for 10 minutes. The experimenter used tweezers to hold a piece of Corneofix® F20 tape and applied it to the middle of the right cheek of the subject. The tape was gently pressed for 30 seconds. The same part was sampled twice in a row. The first tape that may have skin surface contaminants was discarded, and the second tape was used to extract ceramide using the MeOH method. The second lipid tape sample was placed in a borosilicate sample bottle, and 1.2 mL of methanol was added to a lipid tape sample with an area of ​​2.8 cm x 1.9 cm. The sample was ultrasonicated for 30 seconds, the shaker temperature was set at 37℃, and the sample was shaken for 2 hours. The sample was centrifuged at 5000 rpm for 10 minutes, and the upper phase was collected; after nitrogen blowing, it was re-dissolved with 150 μL of solution, and the volume ratio of the solution was isopropanol: methanol = 1:2.

[0078] 2. Use UPLC-Q-TOF-MS method to separate and analyze ceramides on the skin surface

[0079] The present invention adopts UPLC-Q-TOF-MS method to separate and analyze ceramide on the skin surface.

[0080] Chromatographic parameters: the chromatographic column was an Agilent Eclipse Plus C18 reversed-phase column (100 mm × 2.1 mm, 1.8µm); the mobile phase A was methanol / water / formic acid (60:40:0.2, v / v / v) + 10 mmol / L ammonium acetate, and the mobile phase B was methanol / isopropanol / formic acid (60:40:0.2, v / v / v) + 10 mmol / L ammonium acetate; the sample chamber temperature was 4°C, the injection volume was 1 μL, the flow rate was 0.35 mL / min, and the column temperature was 40°C. Gradient elution program: 0-3 min, 0%-10% B; 3-5 min, 10%-40% B; 5-5.3 min, 40%-55% B; 5.3-8 min, 55%-60% B; 8-8.5 min, 60%-80% B; 8.5-10.5 min, 80% B; 10.5-16 min, 80%-90% B; 16-19 min, 90% B; 19-22 min, 90%-100% B; 22-26 min, 100% B; 26-34 min, 0% B.

[0081] Mass spectrometry parameters: The primary mass spectrometer adopted TOF full scan mode and was detected in positive mode; the drying gas flow rate was 10 L / min, the drying gas temperature was 350℃, the nebulizer gas pressure was 30 psi, the capillary voltage was 3 500 V, the skimmer voltage was 65 V, the octopole radio frequency voltage was 750 V, the capillary outlet voltage was 150 V, and the mass-to-charge ratio scanning range was m / z 50-1500; two internal standard ions m / z 121.0509 and 922.0098 in the reference solution were selected for real-time mass correction of the instrument measurement results; the secondary mass spectrometer adopted the targeted secondary ion mode (Targeted-MS / MS), and the collision voltage could be selected as 20 or 40 eV according to the difficulty of target ion fragmentation. The acquisition time interval of Targeted-MS / MS was set to 0.4 min, and the quadrupole filter m / z range was set to Medium level.

[0082] 3. Intelligent analysis and identification system

[0083] The obtained primary mass spectrometry data were imported into Agilent MassHunter Qualitative Analysis (version B.06.00) software. The self-built ceramide primary molecular formula library was selected as the database. The “Find by formula” function was used to match the quasi-molecular ions (i.e., parent ions) collected by the primary mass spectrometry, and they were organized into a targetlist.csv file for targeted-MS / MS analysis.

[0084] The structure of ceramide is complex and diverse, with up to 12 subclasses. Rapidly and accurately obtaining the mass spectrum fragments of each ceramide is the first step in spectrum interpretation. By consulting literature and experimental spectrum interpretation, the present invention summarizes the cleavage rules of each subclass of ceramide, and based on the data automatic processing function of the Python program, writes this chemical logic into a computer language to achieve the functions of secondary characteristic fragment ion prediction and automatic annotation of ceramide structure.

[0085] Figure 3 The fragmentation mechanism of monohydroxyceramide in the mass spectrometry positive mode is shown, and the four characteristic fragment ions shown in the figure are [M+H-H2O]+ (neutral loss of one water molecule), [M+H-FA]+ (de-FA fatty acid), [M+H-FA-H2O]+ (sphingosine skeleton fragment), and [C2H5NO+H]+ (ceramide characteristic fragment); the name of ceramide is highly correlated with its structure. The present invention converts the chemical logic of the fragmentation mechanism of ceramides with different structures into computer language and stores it in a function named after its abbreviation name, such as CerP is the abbreviation of 1-phosphoceramide, and the CerP(MH,name) function is the secondary fragment prediction function of CerP-type ceramides. The ceramide name "classy_value" is identified from the global namespace through func = globals().get(classy_value), and the mass spectrometry fragment information generated can be predicted.

[0086] The above is the operating principle of the ceramide fragment ion prediction system. The following takes monohydroxyceramide Cer (m34:2) as an example (e.g. Figure 4 ), explain the matching and annotation logic of the intelligent spectrum analysis algorithm:

[0087] First, through the above-mentioned fragment ion prediction system, we can reasonably predict the possible structures of ceramide Cer (m34:2) and the mass spectrometry secondary fragment information they theoretically generate, such as Cer (m18:1 / 16:1), Cer (m18:2 / 16:0), Cer (m20:1 / 14:1), etc. Cer (m18:1 / 16:1) may generate the following Figure 4 The five fragments shown in A are 520.5049 ([M+H]+), 502.5006 (neutral loss of 1 water molecule), 284.2983 (de-FA fatty acid), 266.2838 (sphingosine skeleton fragment), and 60.044 (ceramide characteristic fragment); Cer (m18:2 / 16:0) This structure may produce Figure 4The five fragments shown in B are 520.5049 ([M+H]+), 502.5006 (neutral loss of one water molecule), 282.2796 (de-FA fatty acid), 264.2625 (sphingosine backbone fragment) and 60.044 (ceramide characteristic fragment).

[0088] The second step, result = query(row, attr_array, attr_cols), calls the query function row by row. When the predicted m / z value of the fragment differs from the actual data within ±10ppm, the match is successful and the result is output to 'match result.csv' for easy data retrieval. The predicted m / z of Cer(m18:1 / 16:1) is matched with the actual mass spectrum data. 60.044, 520.5049, and 502.5006 are matched successfully, while 266.2838 and 284.2983 differ from the actual values ​​by more than 10ppm, indicating that the actual mass spectrum fragments do not contain these two fragment ions.

[0089] The third step is to infer the structure of ceramide based on the matched fragments. Call the find_max(row) function and, according to the principle of "largest number of fragments and greatest intensity", give the most credible result as the structural annotation of the ceramide. The predicted m / z of Cer(m18:1 / 16:1) and the predicted m / z of Cer(m18:2 / 16:0) have 3 and 5 fragment ions matched with the actual data respectively. According to the principle of "largest number of fragments and greatest intensity", Cer(m18:2 / 16:0) has higher credibility.

[0090] Finally, a retention time prediction model was constructed based on the linear law between the carbon number of ceramides in the same series and their chromatographic retention time to correct the results of the above-mentioned intelligent spectrum decomposition algorithm. At the same time, this prediction model can also be used to find some ceramides with low abundance or missing fragments. The results of the intelligent spectrum decomposition algorithm are used as training sets. Because it is not certain that all results are accurate, there will definitely be outliers (i.e., false detections) in the training set data. Therefore, the least squares method is not applicable to the current system. Instead, RANSACRegressor is imported from the sklearn.linear_model module, and robust regression is used to train the model. The number of carbon atoms on the ceramide fatty acid chain is x, and the retention time is y. Randomly select two points from M points, and calculate a and b to obtain a hypothetical linear model f(x) = ax + b. Set the minimum number of sample points required for random selection to estimate model parameters to 4, and the error ϵ to ±0.1. Calculate the distance between the y value of the hypothetical model and the true y value. If the distance is less than ϵ, that is, |ax +b - y|<ϵ, then the point is considered to be an inlier. If there are less than 4 inliers, resample points to generate a new hypothetical model for calculation. Figure 5 Taking the ceramide data of the Cer (d18:0 / x:0) series as an example, the false positives were successfully eliminated from the data of the training set (train_data), and a linear equation was fitted, R 2 The value of 0.9982 indicates a good fit.

[0091] Note: Some of the false positives in the training set are due to the structural isomerism of ceramide, such as double bond position isomerism. This problem does not mean that the accuracy of the intelligent analysis algorithm is low, but because a single mass spectrometry method cannot solve the problem of structural isomerism.

[0092] Example 2 Comparison of ceramide extraction effects of different stratum corneum adhesive tapes

[0093] The tapes used for skin stratum corneum adhesion were selected from: TERAOKA SEISAKUSHO 465 #40 (TERAOKASEISAKUSHO, Japan), Corneofix® F20 (Courage+Khazaka GmbH, Germany), and D-Squame D100 (CUDERM, USA). The extraction method used was the same pure methanol method, and the area of ​​each tape was cut to be roughly the same (5 ± 0.5 cm 2), put a piece of lipid tape with skin ceramide into a borosilicate sample bottle, then add 1.2mL methanol to extract, ultrasonicate for 30s, set the shaker temperature at 37℃, shake for 2 hours, collect the upper phase after centrifugation; re-dissolve with 150μL solution after nitrogen blowing, the volume ratio of the solution is isopropanol: methanol = 1:2. Pass through a 0.2 μm filter membrane into a liquid phase vial and store at -80℃ for testing.

[0094] The results of ceramide extraction by different tapes are shown in Figure 2. Figure 6 As shown, from the comparison of ceramide identification number and total abundance, Corneofix® F20>D-Squame D100>TERAOKA SEISAKUSHO 465 #40, the results show that Corneofix® F20 is the most suitable tape for skin ceramide extraction.

[0095] Example 3 Comparison of the effects of different extraction methods

[0096] Methanol + chloroform method (MeOH + CHCl3 method): Place a piece of lipid tape (Corneofix® F20) adhering to skin ceramide in a borosilicate sample bottle, add 1 mL of methanol and 0.5 mL of chloroform in turn, and vortex for 30 seconds; then place the sample in a shaker, shake at 37 °C and 120 rpm for 2 hours, centrifuge, take the supernatant into a new borosilicate sample bottle, blow with nitrogen and freeze-dry overnight; add 150 μL of solution to reconstitute, the volume ratio of the solution is isopropanol: methanol = 1:2. Pass through a 0.2 μm filter membrane into a liquid phase vial and store at -80 °C for testing.

[0097] Methyl tert-butyl ether method (MTBE method): Place a piece of lipid tape (Corneofix® F20) for skin ceramide adhesion in a borosilicate glass bottle, then add 225 μL MeOH and 1 ml MTBE, vortex for 30 seconds, centrifuge, take the supernatant, add it to a 4 mL glass vial, blow it with nitrogen and freeze-dry it overnight; add 150 μL of solution to reconstitute, the volume ratio of the solution is isopropanol: methanol = 1:2. Pass through a 0.2 μm filter membrane into a liquid phase vial and store at -80°C for testing.

[0098] Methanol method (MeOH method): Place a piece of lipid tape (Corneofix® F20) attached to skin ceramide in a borosilicate sample bottle, then add 1.2 mL of methanol for extraction, ultrasonicate for 30 seconds, set the shaker temperature at 37°C, shake for 2 hours, collect the upper phase after centrifugation; re-dissolve with 150 μL of solution after nitrogen blowing, the volume ratio of the solution is isopropanol: methanol = 1:2. Pass through a 0.2 μm filter membrane into a liquid phase vial and store at -80°C for testing.

[0099] The results of different extraction methods are as follows Figure 7As shown, from the comparison of the number of ceramide identification and total abundance, MeOH method > MTBE method > MeOH+CHCl3 method. The results show that the pure methanol method is the most suitable method for skin ceramide extraction.

[0100] Example 4 Comparison of the effects of different pre-treatment methods

[0101] MTBE method: Place a piece of lipid tape (D-Squame D100) with skin ceramides in a borosilicate glass bottle, then add 225 μL MeOH and 1 ml MTBE, vortex for 30 seconds, centrifuge, take the supernatant, add it to a 4 mL glass vial, blow it with nitrogen and freeze-dry it overnight; add 150 μL of solution to dissolve it, the volume ratio of the solution is isopropanol: methanol = 1:2. Pass through a 0.2 μm filter membrane into a liquid phase vial and store it at -80 °C for testing. The number of ceramides extracted by this treatment method is 126, and the total abundance is 6.78×10 6 .

[0102] Example 5 Application of the automatic skin condition assessment system

[0103] Sample sources: 3 female subjects from Shanghai, aged 25, 34, and 45 years old. Before sampling, the subjects cleaned their faces with clean water, sat quietly in a constant temperature and humidity room for 30 minutes, and applied laboratory-made lipid tape to the skin in the middle area of ​​the right cheek of the subjects. After pressing for 30 seconds, the tape was removed, and the same area was sampled twice.

[0104] Verification was performed according to the method described in Example 1. The verification results of subject 1 (female, 25 years old) are shown in Table 1, the verification results of subject 2 (female, 34 years old) are shown in Table 2, and the verification results of subject 3 (female, 45 years old) are shown in Table 3.

[0105] Table 1 Verification results of subject 1

[0106]

[0107] Table 2 Verification results of subject 2

[0108]

[0109] Table 3 Verification results of subject 3

[0110]

[0111] The results obtained by the intelligent analysis algorithm were verified and corrected by the retention time prediction model, which improved the accuracy of identification; for example, the retention time prediction model of Cer (d18:2 / x:0) and Cer (t18:0 / x:0) series ceramides was used as follows: Figure 8-Figure 9 As shown, R 2They are all greater than 0.99, indicating that there is a very strong linear relationship between the model and the data, and the degree of fit is very high.

[0112] The above detailed description is a specific description of one of the feasible embodiments of the present invention, and the embodiment is not intended to limit the scope of the present invention. It should be pointed out that any equivalent implementation or change that does not deviate from the present invention should be included in the scope of the technical solution of the present invention. Therefore, the protection scope of the patent of the present invention shall be based on the attached requirements.

Claims

1. A skin condition evaluation system, characterized in that: The evaluation system includes: sample pretreatment, UPLC-Q-TOF-MS separation and analysis, and intelligent analysis and identification system; The sample pretreatment includes: adsorbing ceramide with lipid tape, adding extracting solution, ultrasonication, oscillation, centrifugation to collect the upper phase, and re-dissolving after drying; the lipid tape is Corneofix® F20; the extracting solution is methanol, and the solution used for re-dissolving is a mixture of isopropanol and methanol; The intelligent analysis and identification system comprises the following steps: (1) Predict fragment ions based on the target ion mass number to obtain predicted secondary mass spectrometry fragment information; (2) Matching the predicted secondary mass spectrometry fragment information with the actual collected secondary mass spectrometry data using a fragment ion library to obtain data that meets the matching requirements; (3) Structural annotation of the data with the highest credibility that meets the matching requirements; (4) Construct a retention time prediction model based on the linear law between the carbon number of ceramides in the same series and their chromatographic retention time, and train the retention time prediction model to eliminate false detection results; (5) Predict ceramide structure based on retention time; The standard of the data meeting the matching requirement in step (2) of the intelligent analysis and identification system is: the difference between the predicted m / z value of the fragment and the actual predicted m / z value is within the range of ±10ppm, which meets the matching requirement; The highest credibility standard in step (3) of the intelligent analysis and identification system is: the highest credibility is achieved when the number of mass spectrum fragments is the largest and the ion current intensity is the largest; The training retention time prediction model described in step (4) of the intelligent analysis and identification system is: using robust regression to train the retention time prediction model.

2. The evaluation system according to claim 1, characterized in that: The secondary mass spectrum fragment information in step (1) of the intelligent analysis and identification system includes: [M+H]+, neutral loss of 1 water molecule, de-FA fatty acid, sphingosine skeleton fragment, ceramide characteristic fragment or more.

3. The evaluation system according to claim 1, characterized in that: In the mixed liquid of isopropanol and methanol, the volume ratio of isopropanol to methanol is 1:1-3.

4. The evaluation system according to claim 1, characterized in that: The chromatographic parameters in the UPLC-Q-TOF-MS separation analysis are: Chromatographic column: Agilent Eclipse Plus C18 reverse phase column, 100 mm × 2.1 mm, 1.8 µm; Mobile phase A: methanol-water-formic acid solution containing 10 mM ammonium acetate, wherein the volume ratio of methanol, water and formic acid in the methanol-water-formic acid solution is 60:40:0.2; Mobile phase B: methanol-isopropanol-formic acid solution containing 10 mM ammonium acetate, wherein the volume ratio of alcohol, isopropanol and formic acid in the methanol-isopropanol-formic acid solution is 60:40:0.2; Flow rate: 0.35 mL / min; Gradient elution program: 0-3 min, 90%-100% mobile phase A, 0%-10% mobile phase B; 3-5 min, 60%-80% mobile phase A, 10%-40% mobile phase B; 5-5.3 min, 45%-60% mobile phase A, 40%-55% mobile phase B; 5.3-8 min, 40%-45% mobile phase A, 55%-60% mobile phase B; 8-8.5 min, 20%-40% mobile phase A, 60%-80% mobile phase B; 8.5-10.5 min, 20% mobile phase A, 80% mobile phase B; 10.5-16 min, 10%-20% mobile phase A, 80%-90% mobile phase B; 16-19 min, 10% mobile phase A, 90% mobile phase B; 19-22 min, 0%-10% mobile phase A, 90%-100% mobile phase B; 22-26 min, 0% mobile phase A, 100% mobile phase B; 26-34 min, 100% mobile phase A, 0% mobile phase B.

5. The evaluation system according to claim 1, characterized in that: The mass spectrometry parameters in the UPLC-Q-TOF-MS separation analysis are: The primary mass spectrometer adopted TOF full scan mode and was detected in positive mode; the drying gas flow rate was 10 L / min, the drying gas temperature was 350℃, the nebulizer gas pressure was 30 psi, the capillary voltage was 3500 V, the skimmer voltage was 65 V, the octopole radio frequency voltage was 750 V, the capillary outlet voltage was 150 V, and the mass-to-charge ratio scanning range was m / z 50-1500; two internal standard ions m / z 121.0509 and 922.0098 in the reference solution were selected for real-time mass correction of the instrument measurement results; The secondary mass spectrometry adopted the targeted secondary ion mode Targeted-MS / MS. According to the difficulty of target ion fragmentation, the collision voltage was selected as 20 or 40 eV, the acquisition time interval of Targeted-MS / MS was set to 0.4 min, and the quadrupole filter m / z range was set to Medium level.

6. Use of the evaluation system according to any one of claims 1 to 5 in the identification of skin ceramide or the determination of skin ceramide content.

7. A method for identifying skin ceramide or determining the content of skin ceramide, characterized in that: The method comprises using the evaluation system according to any one of claims 1-5.

Images