Method, model and marker for noninvasively judging endometrial receptivity

Abstract

The invention provides a method for predicting endometrial receptivity by using a sample obtained in a non-invasive sampling mode, a method for establishing a prediction model by using the sample, the established prediction model, a marker associated with the endometrial receptivity in the sample and a corresponding kit.

Description

[0001] This application is a divisional application of Chinese Patent Application No. 202210253498.7, filed on March 15, 2022, entitled "Method, Model and Marker for Non-invasive Determination of Endometrial Receptivity". TECHNICAL FIELD

[0002] The present application relates to the field of biology and clinical assisted reproduction, more particularly to a method for non-invasive determination of endometrial receptivity, and the markers, prediction models and corresponding kits and systems used in the method. BACKGROUND

[0003] In the process of human reproduction, the fertilized egg is positioned, adhered and implanted in the mother's uterus, and finally develops into a mature fetus. The implantation process has an important influence on successful pregnancy. In addition to the need for high-quality embryos, successful clinical pregnancy also requires good endometrial receptivity (ER) and synchronous development of the endometrium and the embryo. Endometrial receptivity is a physiological phenomenon, which refers to the acceptance ability of the endometrium to the embryo. Only in a short period of time, the endometrium allows the embryo to implant, which is called the "implantation window period". For adult women, it is equivalent to the 20th-24th day of the menstrual cycle or 6-8 days after ovulation.

[0004] The specific mechanism of endometrial receptivity is not fully understood at present, but it can be clearly stated that the endometrium in the non-receptive period is one of the important reasons for the failure of embryo implantation in in vitro fertilization-embryo transfer (IVF-ET). The success rate of embryo transfer is very important, which is directly related to whether all the costs and efforts paid before can be successful or not. However, the traditional method of using the menstrual cycle window period is not always feasible, and even for a considerable part of the subjects, such as women with repeated implantation failure or women suffering from other secondary infertility, the time node of the implantation window period is very inaccurate. If the implantation window period is still calculated according to the menstrual cycle or ovulation day, there will be a great risk of implantation failure.

[0005] There are several ways to determine endometrial receptivity in clinic, such as transvaginal ultrasound, serum estradiol and progesterone levels, endometrial biopsy, etc. to determine the time of endometrial implantation window, but there are certain errors. With the progress of molecular biology, especially the completion of the human genome project, the use of genes as molecular markers to assist in the determination of various physiological and pathological states in clinic is becoming more and more mature. However, in the prediction of endometrial receptivity, the existing methods all have a significant disadvantage, that is, the samples they use are all from endometrial biopsy sampling. This invasive method causes unnecessary damage to the mother and has potential instability, thus reducing the compliance of the subjects, and the damage caused by biopsy prevents embryo transfer, prolonging the time to pregnancy and ART cycle of the subjects. In addition, the additional operation itself is a way of wasting resources and increasing the pain of the subjects.

[0006] Researchers have been pursuing simpler and less invasive examination methods, and the prediction of endometrial receptivity is no exception. There is an urgent need for an efficient and non-invasive endometrial receptivity diagnostic method in clinic. Brief description of the invention

[0007] The present application provides a non-invasive sampling method for determining endometrial receptivity.

[0008] In a first aspect, the present application provides a method for obtaining a sample from a living body for analysis by a non-invasive method. In a specific embodiment, the non-invasive method is the adhesion of trace amounts of tissue to the wall of the transfer tube at the time of pre-embryo transfer. In another specific embodiment, the non-invasive method is the adhesion of trace amounts of tissue to the wall of the transfer tube at the time of embryo transfer.

[0009] In one embodiment, the number of cells in the sample obtained by the non-invasive method is less than 1000. In a more specific embodiment, the number of cells in the sample obtained by the non-invasive method is less than 900, 800, 700, 600, 500, 400, 300 or 200. In a preferred embodiment, the number of cells in the sample obtained by the non-invasive method is between 400 and 300.

[0010] In a specific embodiment, the method of the present application further comprises transferring the cells and tissues of the tube wall into the sample collection tube by the method of cutting and flushing. In a specific embodiment, in a preferred embodiment, a tissue sample preservation solution is pre-added to the sample collection tube,

[0011] In one embodiment, the method of the present application further comprises extracting total RNA from the sample. In one specific embodiment, the extraction procedure is performed using the Epicentre® MALBAC® Platinum Microprep RNA Amplification Kit (KT110700724) according to the manufacturer's instructions. In another specific embodiment, the extraction procedure comprises the following steps: a) centrifuge the sample at high speed (1600 g, 2 min; 3000 g, 3 min; 13000 rpm, 5 min) to pellet all cells. b) remove supernatant and resuspend cells in 200 μΐ^x2 Tissue Sample Preservative, centrifuge at 3000 g for 5 min; 200 μΐ^x1 time, retain approximately 50 μΐ^. c) add 300 μΐ^of cell lysis solution, vortex vigorously for 30 seconds to 1 minute; centrifuge briefly and discard supernatant. d) add 350 μΐ^of freshly prepared 70% ethanol, vortex to mix; centrifuge briefly and discard supernatant. e) add 650 μΐ^of the above mixture to the adsorption column, centrifuge at 13000 x g for 1 min, discard the centrifugation waste. f) add 500 μΐ^of wash buffer to the adsorption column membrane, centrifuge at 13000 x g for 1 min, discard the centrifugation waste. g) carefully pipette 100 μΐ^of DNA digestion enzyme onto the adsorption column membrane, incubate at 25°C for 15 min. h) add 500 μΐ^of wash buffer, centrifuge at 13000 x g for 1 min, discard the centrifugation waste; add another 500 μΐ^of wash buffer to the adsorption column, centrifuge at 13000 x g for 1 min, discard the centrifugation waste. i) add 700 μΐ^of 80% ethanol to the tube, centrifuge at 13000 x g for 2 min, discard the centrifugation waste; centrifuge at 13000 x g for another 1 min, discard the collection tube. j) insert the adsorption column into a 1.5 mL centrifuge tube, leave the cap off to dry for 1 min; add 20 μΐ^of DEPC water, incubate at 25°C for 1 min, centrifuge at 13000 x g for 2 min, repeat once more, collect a total of 40 μΐ^of eluate, which is the total RNA. In a more specific embodiment, almost all of the RNA in the micro-sample obtained by non-invasive sampling is extracted, thereby providing a sufficient amount of purified RNA nucleic acid sample for subsequent detection.

[0012] In one embodiment, the method of the present application further comprises reverse transcribing the mRNA in the RNA nucleic acid sample to generate a cDNA library. In one specific embodiment, the reverse transcription is performed using the Epicentre® MALBAC® Platinum Microprep RNA Amplification Kit (KT110700724) according to the manufacturer's instructions.

[0013] In one embodiment, the method of the present application further comprises further processing the cDNA library to construct a sequencing library. In one specific embodiment, the further processing is performed using the Epicentre® Fragmentation Kit (KT100804248) and the Gene Sequencing Library Kit (KT100804048) according to the manufacturer's instructions.

[0014] In one embodiment, the method of the application further comprises sequencing the resulting library. In one particular embodiment, the method of sequencing is Next-Generation Sequencing (NGS). In another particular embodiment, the method of sequencing is Sanger sequencing, or third generation sequencing (e.g. Nanopore sequencing). In a preferred embodiment of the application, the application uses the Illumina Nextseq sequencing system and the companion Next-seq High-Output Kit reagents.

[0015] In a second aspect, the application provides a method for building a receptivity prediction model. In one embodiment, the application provides a method for performing machine learning algorithm analysis on sequencing results and building a receptivity prediction model.

[0016] In one particular embodiment, the machine learning algorithm analysis comprises the following steps: I) dividing the samples into receptivity and non-receptivity according to the gold standard (actual clinical outcome), and preferably splitting into two sets, training set and test set, and taking the training set sample data for the next step; II) denoising and debiasing all the data of the training set samples, generating a fragment count alignment file for each gene respectively; III) screening and collecting differentially expressed genes as to be identified features; IV) converting the data into TPM as the normalized expression data of each gene, thereby eliminating the influence of sample data volume difference; V) inputting the expression data of the to-be-identified features (genes) into the machine learning algorithm, and performing supervised learning by the machine learning algorithm to obtain the threshold, sensitivity and importance of each feature for judging receptivity, thereby obtaining a receptivity prediction model. Preferably, the established model is verified for accuracy in the test set.

[0017] In one embodiment, the model comprises one or more sub-models, different sub-models containing all or different parts of the aforementioned differentially expressed genes.

[0018] In one particular embodiment, step II) specifically comprises:

[0019] a) for each sample's sequencing data, the following processing is performed respectively

[0020] i. the sequencing data is processed with trimmomatic (version 0.33) to remove adapter sequences and filter low-quality reads,

[0021] to generate clean sequencing data

[0022] ii. Clean sequencing data generated in the previous step is aligned to human reference genome (version GRCh38) using HISAT2 (version 2.0.5) to generate alignment files

[0023] iii. The alignment files generated in the previous step are sorted and duplicated sequences are marked to generate final alignment files

[0024] iv. The alignment files generated in the previous step are processed using htseq_count, which also requires a gene annotation file, Homo_sapiens.GRCh38.84.gtf file, which is downloaded from Ensembl, containing the name, location coordinates, etc. of each gene, htseq_count can count the number of sequences in the region of each gene according to this information (for paired-end sequencing, reads belonging to the same fragment are counted only once), to generate a genecount file

[0025] b) Combine the genecount files of all samples in a batch to generate a total genecount file, where each row is a sample and each column is a gene, and the data in it is the number of fragments belonging to the gene in the sample. In a specific embodiment, step III) specifically comprises the following selection method:

[0026] i. Exclude rRNA-related and mitochondrial-related genes that interfere with subsequent experiments.

[0027] ii. For the training set, perform gene differential expression analysis to screen differentially expressed genes in the two types of samples, obtaining 3193 genes

[0028] iii. The 3193 genes screened in the previous step are further screened by variable selection, and a total of 250 genes are obtained, which are used as candidate features for the next model training, and the detailed list is shown in Table 2.

[0029] In a more specific embodiment, the variable selection is to select genes with FDR < 0.001 from the differentially expressed genes, obtaining 1270 genes, using the cpm function of edgeR to output the logTMM value of each sample and each gene, calculating the pearson correlation between each other, and then selecting genes according to the following steps

[0030] 1) Sort the genes according to FDR from small to large;

[0031] 2) Mark the top gene in the list as selected;

[0032] 3) Mark the genes with a correlation greater than or equal to 0.6 with any of the selected genes as rejected;

[0033] 4) Repeat 2) and 3) until the end of the list;

[0034] 5) Use the genes selected in 4) to build a model, and use 10-fold cross-validation to evaluate the accuracy of the model, and remove the genes that have no effect on the accuracy of the model, and finally obtain 250 genes.

[0035] In a more specific embodiment, the above step ii. uses edgeR (version 3.34.0) for gene differential expression analysis, and the screening conditions are logFoldChange > 0, FDR < 0.05, including the following steps:

[0036] a) edgeR requires two input files, the first is the genecount file containing the number of reads of each gene in each sample, and the second is the sample information file containing the category to which the sample belongs;

[0037] b) After edgeR reads the input file, it first normalizes the number of reads of each gene in each sample, and selects genes that are stably expressed in all samples for normalization;

[0038] c) edgeR then finds genes with differential expression in different categories according to the category information of the sample, and the significance of the difference is represented by the FDR-adjusted p-value, the smaller the p-value, the more significant the difference.

[0039] In a specific embodiment, the step IV) specifically includes converting the genecount data to TPM to exclude the influence of the amount of data under the sample. The calculation formula is as follows:

[0040]

[0041] Wherein Ni is the read count on the ith gene, and Li is the length of the ith gene.

[0042] In a specific embodiment, the machine learning algorithm in the step V) is a random forest algorithm.

[0043] In an embodiment, the method of the present application further comprises evaluating and / or verifying the sensitivity, specificity, accuracy, ROC and other indicators of the established tolerance prediction model in the split independent test set.

[0044] In one specific embodiment, the method of the application further comprises validating the established receptivity prediction model in at least two test sets.

[0045] In a third aspect, the present application also provides a biomarker set for predicting endometrial receptivity to guide embryo transfer.

[0046] In one embodiment, the biomarker set of the application comprises some or all of the differentially expressed genes identified in step III) above, for example any number between more than 5 and less than or equal to 250. In a more specific embodiment, in order to accurately predict endometrial receptivity to guide embryo transfer, the biomarker set of the application comprises at least 10 gene members. In an even more specific embodiment, the biomarker set of the application comprises at least 11 gene members, at least 12 gene members, at least 13 gene members, at least 14 gene members, at least 15 gene members, at least 16 gene members, at least 17 gene members, at least 18 gene members, at least 19 gene members, at least 20 gene members, at least 25 gene members, at least 30 gene members, at least 35 gene members, at least 40 gene members, at least 45 gene members, at least 50 gene members, at least 55 gene members, at least 60 gene members, at least 65 gene members, at least 70 gene members, at least 80 gene members, at least 90 gene members, at least 100 gene members, at least 110 gene members, at least 120 gene members, at least 130 gene members, at least 140 gene members, at least 150 gene members, at least 160 gene members, at least 170 gene members, at least 180 gene members, at least 190 gene members, at least 200 gene members, at least 210 gene members, at least 220 gene members, at least 230 gene members, at least 240 gene members, at least 250 gene members. In one embodiment, the biomarker set of the application comprises the 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 40, 50, 60, 70, 80, 90, 100, 110, 120, 130, 140, 150, 160, 170, 180, 190, 200, 210, 220, 230, 240 or 250 genes with the highest weight in the receptivity prediction model established according to step V) above.

[0047] In one embodiment, the biomarker set of the present application comprises at least the following genes: ELK4, PRKACB, PHB2, GM2A, PPA1, NCEH1, CAPG, DDIT4, LDHB and TDRD6.

[0048] In a fourth aspect, the present application provides uses and methods based on the biomarker set.

[0049] In one embodiment, the present application provides uses of the above-mentioned biomarker set or specific detection tools thereof for the preparation of assay compositions, assay reagents and / or assay devices for detecting and determining the endometrial receptivity status of a subject, wherein the assay compositions, assay reagents and / or assay devices are capable of detecting the expression level of each member of the above-mentioned biomarker set in a sample from a subject; and the expression level is calculated by the receptivity prediction model of the present application, and the result is used to determine the diagnosis of the endometrial receptivity status.

[0050] In one embodiment, the present application provides a method for predicting endometrial receptivity, which comprises obtaining the expression level of all gene members of the biomarker set by using a substance or method for specifically determining the level or amount of a biomarker. In a specific embodiment, the substance for determining the level or amount of a biomarker is a composition, reagent or device. In a specific embodiment, the substance for determining the level or amount of a biomarker is a specific primer and / or probe molecule. In a more specific embodiment, the composition or reagent for determining the level or amount of a biomarker further comprises one or more of a tag sequence, a linker sequence, a fluorescent molecule, a pyrophosphate molecule and a label molecule. In a specific embodiment, the method for determining the level or amount of a biomarker is a PCR method. In a more specific embodiment, the PCR method is a real-time fluorescent quantitative PCR, isothermal amplification PCR and the like commonly used in the art. In a specific embodiment, the method for determining the level or amount of a biomarker is a sequencing method. In a more specific embodiment, the sequencing method is one or more of next-generation sequencing (NGS), Sanger sequencing and third-generation sequencing (e.g., nanopore sequencing).

[0051] In one embodiment, the reagent for determining the level or amount of a biomarker does not comprise a reagent for specifically detecting genes other than members of the biomarker set of the present application.

[0052] In one embodiment, the method for determining the level or amount of a biomarker does not determine the level or amount of a gene other than a member of the set of biomarkers of the application.

[0053] In one embodiment, the present application provides a method for predicting endometrial receptivity, wherein the sample used in the method is obtained by a non-invasive method, i.e. the sample is obtained without causing damage to the endometrium of the subject. In a preferred embodiment, the present application provides an integrated non-invasive method for predicting endometrial receptivity, wherein "integrated" means that the method is not performed separately and additionally, but is performed simultaneously with the pre-implantation of the embryo or the implantation of the embryo, and the sampling process is integrated in the above-mentioned operation, so that the sampling period is the implantation period. For example, when the cycle implantation fails and the detection result indicates that the endometrium is the cause, the detection result can be used to guide the implantation of the next cycle without waiting, which greatly shortens the time to pregnancy for the pregnant person.

[0054] In another aspect, the present application provides a kit for use in the above-mentioned method, which comprises reagents for determining the level or amount of each biomarker in the combination of biomarkers of the present application in a sample, such as specific reagents and general reagents. SUMMARY

[0055] Figure 1 The graph of Figure 10 shows the ROC curve obtained by the feature 10 sub-model in the independent test set.

[0056] Figure 2 The graph of Figure 30 shows the ROC curve obtained by the feature 30 sub-model in the independent test set.

[0057] Figure 3 The graph of Figure 50 shows the ROC curve obtained by the feature 50 sub-model in the independent test set.

[0058] Figure 4 The graph of Figure 70 shows the ROC curve obtained by the feature 70 sub-model in the independent test set.

[0059] Figure 5 The graph of Figure 90 shows the ROC curve obtained by the feature 90 sub-model in the independent test set.

[0060] Figure 6 The graph of Figure 110 shows the ROC curve obtained by the feature 110 sub-model in the independent test set.

[0061] Figure 7 The graph of Figure 130 shows the ROC curve obtained by the feature 130 sub-model in the independent test set.

[0062] Figure 8 The graph shows the ROC curves obtained by the feature150 sub-model on the independent test set.

[0063] Figure 9 The graph shows the ROC curves obtained by the feature170 sub-model on the independent test set.

[0064] Figure 10 The graph shows the ROC curves obtained by the feature190 sub-model on the independent test set.

[0065] Figure 11 The graph shows the ROC curves obtained by the feature210 sub-model on the independent test set.

[0066] Figure 12 The graph shows the ROC curves obtained by the feature230 sub-model on the independent test set.

[0067] Figure 13 The graph shows the ROC curves obtained by the feature250 sub-model on the independent test set. Invention Details

[0068] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art. For the purposes of this invention, the following terms are defined.

[0069] All numerical names, such as pH, temperature, time, concentration, and molecular weight, including ranges, are approximate values, varying in increments of 0.1 (+) or (-). It should be understood that, although not always explicitly stated, all numerical names are preceded by the term "approximately." The term "approximately," when used in conjunction with numerical values, means encompassing a range of numerical values ​​having a lower limit 5% less than the specified numerical value and an upper limit 5% greater than the specified numerical value.

[0070] When the term “and / or” is used to connect two or more options, it should be understood to mean any one of the options or any two or more of the options.

[0071] As used herein, the terms “comprising” or “including” mean to include the stated elements, integers, or steps, but do not exclude any other elements, integers, or steps. In this document, when the terms “comprising” or “including” are used, unless otherwise specified, they also cover situations consisting of the mentioned elements, integers, or steps. For example, when referring to an antibody variable region “comprising” a specific sequence, it is also intended to cover the antibody variable region consisting of that specific sequence.

[0072] It is also to be understood that the reagents described herein are merely exemplary and equivalents thereof are known in the art.

[0073] In the present context, the term "endometrial receptivity" refers to the ability of the endometrium to accept a blastocyst. The endometrium has its maximal ability to accept a blastocyst only during a short, specific period, which allows the blastocyst to implant. This period is referred to as the receptive phase, also known as the implantation window, receptive state, and the like. In contrast, when the endometrium is in a period in which it is not able to accept a blastocyst in most cases, it is referred to as the non-receptive phase, pre-receptive phase, or non-receptive state, and the like. These designations are used interchangeably herein, unless otherwise specified.

[0074] Currently, if the receptive state of the endometrium can be confirmed, in most cases it is possible to achieve successful implantation of a blastocyst by transferring blastocysts of different stages. If the endometrium is in the receptive state, then transfer of blastocysts of stage D5 (day 5 after fertilization) is likely to result in successful implantation, while if the endometrium is in the pre-receptive state, then transfer of blastocysts of stage D3 (day 3 after fertilization) is likely to result in successful implantation.

[0075] The terms "nucleic acid" and "polynucleotide" are used interchangeably and refer to a polymeric form of nucleotides of any length, either deoxyribonucleotides or ribonucleotides or their analogs. Polynucleotides can have any three-dimensional structure and can perform any function. The following are non-limiting examples of polynucleotides: a gene or gene fragment, exons, introns, messenger RNA (mRNA), transfer RNA, ribosomal RNA, ribozymes, cDNA, recombinant polynucleotides, branched polynucleotides, plasmids, vectors, isolated DNA of any sequence, isolated RNA of any sequence, nucleic acid probes, and primers. A polynucleotide can comprise modified nucleotides, such as methylated nucleotides and nucleotide analogs. If present, modifications to the nucleotide structure can be imparted before or after assembly of the polymer. The sequence of nucleotides can be interrupted by non-nucleotide components. A polynucleotide can be further modified, such as by conjugation with a labeling component. The term also refers to both double- and single-stranded molecules. Unless otherwise indicated, embodiments of the present application that include polynucleotides include both double-stranded and single-stranded forms of a polynucleotide, as

[0076] The terms "marker" or "biomarker" are used interchangeably herein.

[0077] The term "marker" as used herein refers to a class of substances derived from an in vivo or ex vivo sample and which are relatively easy to detect in the sample using existing or conventional experimental methods and tools in the art, and whose level of presence (e.g., high or low or absent) correlates with a certain physiological or pathological state, and thus can be used to infer or assist in inferring such a physiological or pathological state by obtaining the level of presence of the marker. A marker can be any substance present in the body, such as, but not limited to, a nucleic acid, a polysaccharide, a protein, an inorganic or organic small molecule, or a polymer or hybrid thereof (e.g., a glycoprotein, a phosphorylated protein, a methylated nucleic acid sequence, etc.), or a gene encoding the same (e.g., a protein). In some embodiments of the application, the marker is a specific gene whose expression level, preferably as reflected by the amount of mRNA, is changed. In some embodiments of the application, the specific physiological state to be inferred is endometrial receptivity, i.e., the likelihood of implantation of an embryo following an embryo transfer procedure. In some further embodiments, the sample is a non-invasive sample. In some preferred embodiments, the expression level of the gene is obtained by sequencing.

[0078] A "gene" refers to a polynucleotide comprising at least one open reading frame (ORF) that, when transcribed and translated, encodes a specific polypeptide or protein. A polynucleotide sequence can be used to identify larger fragments or full-length coding sequences associated therewith. Methods for isolating larger fragment sequences are known to those skilled in the art.

[0079] "Gene expression" or alternatively "gene product" refers to the nucleic acid or amino acid (e.g., peptide or polypeptide) produced upon transcription and translation of a gene.

[0080] A "probe" when used in the context of manipulation of polynucleotides refers to an oligonucleotide provided as a reagent for detecting a target that can be present in a sample of interest by hybridization to the target. Typically, the probe comprises a label or a means by which a label can be attached prior to or subsequent to the hybridization reaction. Suitable labels include, but are not limited to, radioisotopes, fluorescent dyes, chemiluminescent compounds, dyes and proteins including enzymes.

[0081] A "primer" is a short polynucleotide, usually with a free 3'-OH group, which binds to a target or "template" that can be present in a sample of interest by hybridizing to the target, thereby facilitating polymerization of a polynucleotide complementary to the target. A "polymerase chain reaction" ("PCR") is a reaction in which copies of a target polynucleotide are replicated using a "pair" or "set" of primers consisting of "upstream" and "downstream" primers, a catalyst for polymerization such as a DNA polymerase and, typically, a heat-stable polymerase. PCR methods are well known in the art and are taught, for example, in PCR: A Practical Approach, M. MacPherson et al., IRL Press at Oxford University Press (1991). The overall process of producing replicated copies of a polynucleotide, such as PCR or gene cloning, is referred to herein generally as "replication." Primers can also be used as probes in hybridization reactions, such as Southern or Northern blot analysis (Sambrook et al., Molecular Cloning: A Laboratory Manual, Second Edition (1989)).

[0082] As used herein, "expression" refers to the process by which DNA is transcribed into mRNA and / or the process by which the transcribed mRNA is subsequently translated into a peptide, polypeptide, or protein. If the polynucleotide is derived from genomic DNA, expression can include splicing of the mRNA in a eukaryotic cell.

[0083] When "differential expression" is applied to a gene, it refers to the differential production of mRNA transcribed and / or translated from the gene or protein product encoded by the gene. A differentially expressed gene can be overexpressed or underexpressed as compared to the level of expression in a control cell. However, as used herein, overexpression is an increase in gene expression and is typically at least 1.25 fold or, alternatively, at least 1.5 fold or, alternatively, at least 2 fold, or alternatively, at least 3 fold or alternatively, at least 4 fold expression detected in a control counterpart cell or tissue. As used herein, underexpression is a decrease in gene expression and is typically less than 1.25 fold or, alternatively, at least 1.5 fold or, alternatively, at least 2 fold or, alternatively, at least 3 fold or alternatively, at least 4 fold expression detected in a control counterpart cell or tissue. The term "differentially expressed" also refers to where expression is detected in a cancer cell or cancer tissue and the expression is present, but expression is not detectable in a control cell.

[0084] The term "cDNA" refers to complementary DNA, i.e., mRNA molecules present in a cell or organism are made into cDNA using enzymes such as reverse transcriptase. A "cDNA" library is a collection of all the mRNA molecules present in a cell or organism, which are all converted into cDNA molecules using reverse transcriptase, and then inserted into a "vector" (another DNA molecule that can continue to replicate after foreign DNA is added). Exemplary vectors for libraries include bacteriophage, viruses that infect bacteria, such as lambda phage. The library can then be probed for specific cDNAs (and thus mRNAs) of interest.

[0085] Measurement of gene expression

[0086] Gene expression can be detected by any suitable method, including, for example, detecting the amount of mRNA transcribed from the gene or the amount of cDNA produced by reverse transcription of mRNA transcribed from the gene or the amount of polypeptide or protein encoded by the gene. These methods can be performed on a sample, based on the sample or modified high-throughput analysis.

[0087] For example, any of gene copy number, transcription, or translation can be determined using known techniques. For example, amplification methods such as PCR can be used. General procedures for PCR are taught in MacPherson et al., PCR: A Practical Approach, (IRL Press at Oxford University Press (1991)). However, the PCR conditions for each application reaction are determined empirically. Many parameters influence the success of the reaction. In particular, annealing temperature and time, extension time, Mg 2 + and / or ATP concentration, pH, and relative concentrations of primers, template, and deoxyribonucleotides. Following amplification, the resulting DNA fragments can be detected by agarose gel electrophoresis followed by ethidium bromide staining and UV illumination.

[0088] In one embodiment, hybridized nucleic acids are detected by detecting one or more labels attached to the sample nucleic acids. Labels can be incorporated by any of a variety of methods well known to those of skill in the art. In one aspect, however, labels are incorporated simultaneously during an amplification step in which the sample nucleic acids are prepared. Thus, for example, a polymerase chain reaction (PCR) using labeled primers or labeled nucleotides can provide labeled amplification products. In a separate embodiment, labels are incorporated into transcribed nucleic acids using labeled nucleotides (e.g., fluorescein-labeled UTP and / or CTP) in a transcription amplification as described above.

[0089] Alternatively, the label can be added directly to the original nucleic acid sample (e.g., mRNA, poly A, mRNA, cDNA, etc.) or to the amplification product after amplification is complete. Methods of linking labels to nucleic acids are well known to those skilled in the art and include, for example, nick translation or end labeling by the action of a kinase of the nucleic acid (e.g., using labeled RNA), followed by binding (ligating) a nucleic acid linker that binds to the sample nucleic acid to the label (e.g., a fluorophore).

[0090] Detection of labels is well known to those skilled in the art. Thus, for example, radioactive labels can be detected using film or scintillation counters, fluorescent labels can be detected using light detectors that detect emitted light, enzyme labels are typically detected by providing an enzyme and a substrate and detecting the reaction product produced as a result of the action of the enzyme on the substrate, and calorimetric labels are detected by simple visualization of color.

[0091] Thus, "reagents for determining the level or amount of a biomarker" refers to reagents that can be used to quantify or measure the level or amount of a biomarker in a sample. Such reagents can be readily designed or obtained by routine methods known in the art based on the sequences of the biomarkers provided herein. For example, such reagents include, but are not limited to, PCR primers that can be used to quantify or measure the level or amount of a biomarker by, for example, real-time PCR; probes that can be used to quantify or measure the level or amount of a biomarker by, for example, quantitative Southern blotting; microarrays (e.g., gene chips) that can be used to quantify or measure the level or amount of a biomarker, etc. In addition, second-generation sequencing methods or third-generation sequencing methods can also be used to quantify or measure the level or amount of a biomarker, as is known in the art. Thus, such reagents can also be commercially available reagents for performing second-generation sequencing methods or third-generation sequencing methods.

[0092] The term human reference genome refers to a file that documents the base composition at each position on the human chromosomes, produced by the Human Genome Project, GRCh38 is the 38th version. Homo_sapiens.GRCh38.84.gtf is a gene annotation file matching the human reference genome GRCh38, documenting information such as genes and their locations on the reference genome.

[0093] The term Reads or called reads refers to the very large number of sequences included in the NGS machine output, each sequence is called a read.

[0094] Paired-end sequencing, one fragment has been sequenced at both ends, generating two reads, these two reads belong to the same DNA fragment

[0095] The term Trimmomatic refers to a tool to process NGS raw data, http: / / www.usadellab.org / cms / ?page=trimmomatic .

[0096] The term HISAT2 refers to a tool to align reads to human reference genome, see http: / / daehwankimlab.github.io / hisat2 / .

[0097] The term htseq_count refers to a tool to count reads on genes, https: / / htseq.readthedocs.io / en / master / index.html .

[0098] The term edgeR refers to a tool to perform gene differential expression analysis, https: / / www.bioconductor.org / packages / release / bioc / vignettes / edgeR / inst / doc / edgeRUsersGuide.pdf .

[0099] The term LogFoldChange refers to a measure of the difference of expression of a gene between two groups of samples in differential expression analysis, the calculation method is the mean (or median) of expression of group 1 divided by the mean (or median) of expression of group 2, then log2 transformed.

[0100] The term FDR refers to the value after p-value is adjusted by FDR in edgeR, the adjustment method is Benjamini-Hochberg method.

[0101] The term TPM or full name: Transcripts Per Million reads, is a relative quantification index when the sequencing sample size difference is too large to cause the sequencing raw data quantity difference too large. TMP first handles the reads number difference problem caused by gene length, and then handles the reads number difference problem caused by sequencing depth. When the accuracy or quantification of the expression of the target gene is emphasized, TPM is the most effective, and its application makes the relative expression between different sequencing samples comparable. The TPM calculation formula is as follows (Ni is the reads count on the ith gene, Li is the length of the ith gene):

[0102]

[0103] Machine learning

[0104] The term "predictive model" refers to a specific mathematical model obtained by applying a predictive method to a data set. In the embodiments detailed herein, such a data set is obtained from a set of measurements of gene expression in non-invasive samples from subjects, where the classification (either receptive or pre-receptive) of each sample is known. This model can be used to classify samples whose receptivity status is unknown as either receptive or pre-receptive, optionally, this classification is a probabilistic prediction (i.e. a proportion or percentage that is interpreted as a probability). The exact details of how these gene-specific measurements are combined to produce a classification and a probabilistic prediction depend on the specific mechanism used to construct the predictive model.

[0105] The term "machine learning algorithm" refers to a specific strategy that enables a computer to achieve some form of artificial intelligence, as it is commonly understood in the art. Common algorithmic strategies include inductive learning, analogical learning, analytical learning, etc. Depending on the form of learning, it can be supervised or unsupervised, the former referring to the process of adjusting the parameters of an algorithmic classifier using a set of samples with known outcomes, so that it gradually approaches and reaches the desired performance. Commonly used supervised learning algorithms in the art include Naive Bayes, Decision Tree, k-Nearest Neighbors, Neural Networks, Support Vector Machines, Random Forest, Logistic Regression, Least Squares, adboost algorithm, Hidden Markov Models, etc. Unless otherwise specified, machine learning algorithms in this document use supervised learning algorithms. The skilled person will appreciate that, since the present application seeks to discover underlying regularities, the machine learning algorithm does not have a determined requirement for the number of training samples, i.e. samples with known outcomes, and a regular, obtainable sample size is sufficient to train the algorithm.

[0106] Unless otherwise specified, the term "Random Forest" refers to the Random Forest model, which is an ensemble supervised learning algorithm that builds multiple decision tree models and then combines them to obtain a more accurate and stable model, and which also provides good results when using default parameters, being one of the most commonly used models in the art. An example of an implementation of the Random Forest model in R language can be found at: https: / / cran.r-project.org / web / packages / randomForest / randomForest.pdf In some preferred embodiments of the application, the machine learning algorithm is the Random Forest model.

[0107] In some specific embodiments of the application, the commonly used hyperparameters of the Random Forest model are as follows:

[0108] a) n_estimators, number of decision trees to build, usually the greater the number the better, but the computation time also increases, and when the number of decision trees reaches a certain value, the effect does not improve anymore.

[0109] b) max_features the number of features to consider when looking for the best split in a tree. If unspecific, the default value is "sqrt" of the number of features. The smaller the value, the less similar the trees will be.

[0110] c) max_sample the number or proportion of samples to use for building each tree. The smaller the value, the less similar the trees will be.

[0111] d) class_weight the proportion of samples for each class. If not specified, the default value is "None", which means the proportion of samples for each class is the same. If the samples are imbalanced, the value can be set to "balanced".

[0112] e) criterion the measure of how well a split is chosen when building the tree.

[0113] f) max_depth the maximum depth of the tree.

[0114] g) min_samples_split the minimum number of samples needed to split a node.

[0115] h) min_samples_leaf the minimum number of samples in a node.

[0116] In this document, the training of a model refers to the process of finding the optimal set of hyperparameters for the construction of the final model, unless otherwise specified. In some embodiments of the present application, the model training uses the following methods:

[0117] i. Random, randomly combine the hyperparameters, search for a period of time and select the optimal parameters. When tuning the parameters, the accuracy of the model is used to evaluate the model.

[0118] ii. Train the model with default parameters.

[0119] iii. Select the optimal method from the two methods, and the evaluation index is the average accuracy of 10-fold cross-validation (10-fold cv).

[0120] The term Scikit-learn refers to a python-based machine learning software, see https: / / scikit- learn.org / stable / .

[0121] The term Caret refers to a software for tuning the parameters of a machine learning model, see https: / / topepo.github.io / caret / index.html .

[0122] ROC curve is short for Receiver Operating Characteristic curve, which is a comprehensive index reflecting sensitivity and specificity of continuous variables, and showing the relationship between the two by plotting. AUC (area under the ROC curve) is the area under the ROC curve, the larger the AUC, the better, indicating that the classification accuracy of the test is higher.

[0123] In one aspect, the present application provides a method for obtaining a live sample for analysis by a non-invasive method. In one specific embodiment, the non-invasive method is the adhesion of trace amount of tissue on the wall of the transfer tube before the embryo is transferred. In another specific embodiment, the non-invasive method is the adhesion of trace amount of tissue on the wall of the transfer tube when the embryo is transferred.

[0124] In one embodiment, the number of cells in the sample obtained by the non-invasive method is less than 1000. In a more specific embodiment, the number of cells in the sample obtained by the non-invasive method is less than 900, 800, 700, 600, 500, 400, 300 or 200. In a preferred embodiment, the number of cells in the sample obtained by the non-invasive method is between 400 and 300.

[0125] In one specific embodiment, the method of the present application further comprises transferring the cells and tissues on the wall of the tube into the sample collection tube by the method of cutting and flushing. In one specific embodiment, in a preferred embodiment, a tissue sample preservative is pre-added into the sample collection tube,

[0126] In one embodiment, the method of the present invention further includes extracting total RNA from the sample. In one specific embodiment, the extraction process is performed using the Egon MALBAC Platinum MicroRNA Amplification Kit (KT110700724) according to the manufacturer's instructions. In another specific embodiment, the extraction process includes the following steps: a) centrifuging the sample at high speed (1600g, 2 min; 3000g, 3 min; 13000rpm, 5 min) to precipitate all cells. b) removing the supernatant, resuspending the cells in 200μL×2 tissue sample preservation solution, centrifuging at 3000g for 5 min; repeating 200μL×1 time, retaining approximately 50μL. c) adding 300μL of cell lysis buffer, vortexing vigorously for 30 seconds to 1 minute; briefly centrifuging, discarding the supernatant. d) adding 350μL of freshly prepared 70% ethanol, vortexing to mix; briefly centrifuging, discarding the supernatant. e) Add 650 μL of the above mixture to the adsorption column, centrifuge at 13000×g for 1 minute, and discard the centrifugation waste liquid. f) Add 500 μL of washing buffer to the adsorption column membrane, centrifuge at 13000×g for 1 minute, and discard the centrifugation waste liquid. g) Carefully drop 100 μL of DNA digesting enzyme onto the adsorption column membrane and incubate at 25°C for 15 minutes. h) Add 500 μL of washing buffer, centrifuge at 13000×g for 1 minute, and discard the centrifugation waste liquid; add another 500 μL of washing buffer to the adsorption column, centrifuge at 13000×g for 1 minute, and discard the centrifugation waste liquid. i) Add 700 μL of 80% ethanol to the tube, centrifuge at 13000×g for 2 minutes, and discard the centrifugation waste liquid; centrifuge again at 13000×g for 1 minute, and discard the collection tube. j) Insert the adsorption column into a 1.5 mL centrifuge tube, open the cap and let it air dry for 1 minute; add 20 μL of DEPC water, incubate at 25 degrees Celsius for 1 minute, centrifuge at 13000×g for 2 minutes, and repeat once more, collecting a total of 40 μL of eluent, which is the total RNA. In a more specific embodiment, almost all RNA is extracted from the trace sample obtained by non-invasive sampling, thus providing sufficient purified RNA nucleic acid sample for subsequent detection.

[0127] In one embodiment, the method of the present invention further includes reverse transcription of messenger RNA in an RNA nucleic acid sample to generate a cDNA library. In a specific embodiment, the reverse transcription is performed using the Egon MALBAC Platinum MicroRNA Amplification Kit (KT110700724) according to the manufacturer's instructions. In a specific embodiment, the concentration of the cDNA library obtained from the non-invasive sample of the present invention is approximately 10-20 ng / μL, approximately 20-30 ng / μL, approximately 30-40 ng / μL, approximately 40-50 ng / μL, approximately 50-60 ng / μL, approximately 60-70 ng / μL, approximately 70-80 ng / μL, or approximately 80-90 ng / μL.

[0128] In one embodiment, the method of the present application further comprises further processing the cDNA library to construct a sequencing library. In a specific embodiment, the further processing is performed using the Illumina Fragmentation Kit (KT100804248) and the Gene Sequencing Library Kit (KT100804048) according to the manufacturer’s instructions.

[0129] In another alternative embodiment, the cell lysis method, and / or the RNA extraction method, and / or the reverse transcription method, and / or the library amplification method, and / or the fragmentation method, and / or the sequencing library construction method used in the method of the present application are according to the prior application of the applicant’s application No. 201810019044.7 of Chinese patent application for invention, the whole content of which is incorporated herein.

[0130] In one embodiment, the method of the present application further comprises sequencing the resulting library. In a specific embodiment, the method of sequencing is Next-Generation Sequencing (NGS). In another specific embodiment, the method of sequencing is Sanger sequencing, or third-generation sequencing (e.g., such as Nanopore sequencing). In a preferred embodiment of the present application, the present application uses the Illumina Nextseq sequencing system and the accompanying Next-seq High-Output Kit.

[0131] In another aspect, the present application provides a method for establishing a receptivity prediction model. In one embodiment, the present application provides a method for machine learning algorithm analysis of sequencing results and establishing a receptivity prediction model.

[0132] In a specific embodiment, the machine learning algorithm analysis comprises the following steps: I) dividing the samples into receptivity and non-receptivity according to the gold standard (actual clinical outcome), and preferably splitting into two sets, training set and test set, and taking the training set sample data for the next step; II) denoising and debiasing all data of the training set samples to generate a fragment count alignment file for each gene; III) screening and collecting differentially expressed genes as to be identified features; IV) converting the data to TPM as the standardized expression data of each gene, thereby eliminating the influence of sample data volume difference; V) inputting the expression data of the to-be-identified features (genes) into the machine learning algorithm, and performing supervised learning by the machine learning algorithm to obtain the threshold, sensitivity and importance of each feature for judging receptivity, thereby obtaining a receptivity prediction model. Preferably, the established model is verified for accuracy in the test set.

[0133] In one embodiment, the model comprises one or more sub-models, different sub-models comprising all or different portions of the aforementioned differentially expressed genes.

[0134] In one specific embodiment, the step II) specifically comprises:

[0135] c) For each sample, the sequencing raw data is processed separately as follows

[0136] i. The raw data is processed with trimmomatic (version 0.33) to remove adapter sequences, filter low quality reads,

[0137] Clean sequencing data is generated

[0138] ii. The clean sequencing data generated in the previous step is aligned to the human reference genome (version GRCh38) with HISAT2 (version 2.0.5) to generate an alignment file

[0139] iii. The alignment file generated in the previous step is sorted and duplicate sequences are marked to generate a final alignment file

[0140] iv. The alignment file generated in the previous step is processed with htseq_count, which also requires the use of a gene annotation file, Homo_sapiens.GRCh38.84.gtf file, which is downloaded from Ensembl, containing the name, location coordinates, etc. of each gene, htseq_count can count the number of sequences in the region according to the information (for double-end sequencing, reads belonging to the same fragment are counted only once) to generate a genecount file

[0141] d) The genecount files of all samples in a batch are combined to generate a total genecount file, where each row is a sample and each column is a gene, and the data is the number of fragments belonging to the gene in the sample. In one specific embodiment, the step III) specifically comprises the following selection methods:

[0142] i. Exclude rRNA-related and mitochondrial-related genes that have greater subsequent interference.

[0143] ii. For the training set, perform gene differential expression analysis to screen differentially expressed genes in the two types of samples, obtaining 3193 genes

[0144] iii. The 3193 genes screened in the previous step are further screened by variable selection, and a total of 250 genes are obtained, which are used as candidate features for the next step of model training, and the detailed list is shown in Table 2.

[0145] In a more specific embodiment, the variable selection is selecting FDR<0.001 genes from the differentially expressed genes, resulting in 1270 genes, calculating the logTMM value of each gene in each sample using the cpm function of edgeR, calculating the pearson correlation between each pair of genes, and then selecting genes according to the following steps

[0146] 0.001, resulting in 1270 genes, calculating the logTMM value of each gene in each sample using the cpm function of edgeR, calculating the pearson correlation between each pair of genes, and then selecting genes according to the following steps

[0147] 1) Sort the genes according to FDR from small to large;

[0148] 2) Mark the top gene in the list as selected;

[0149] 3) Mark the genes with a correlation greater than or equal to 0.6 with any of the selected genes as eliminated;

[0150] 4) Repeat 2) and 3) until the end of the list;

[0151] 5) Use the selected genes to build a model, evaluate the accuracy using 10-fold cross-validation, and delete genes that have no effect on the accuracy of the model, resulting in 250 genes.

[0152] In a more specific embodiment, step ii. above uses edgeR (version 3.34.0) for gene differential expression analysis, with the following steps:

[0153] a) edgeR requires two input files, the first is the gene count file, which contains the number of reads for each gene in each sample, and the second is the sample information file, which contains information about which category the sample belongs to;

[0154] b) After edgeR reads the input files, it first normalizes the number of reads for each gene in each sample, and selects genes that are stably expressed in all samples for normalization;

[0155] c) edgeR then finds genes that are differentially expressed in different categories according to the sample category information, and the significance of the difference is represented by the FDR-adjusted p-value, with smaller p-values indicating more significant differences.

[0156] In a specific embodiment, step IV) specifically includes converting the gene count data to TPM to exclude the influence of sample machine data volume. The calculation formula is as follows:

[0157]

[0158] wherein Ni is the reads count on the ith gene, Li is the length of the ith gene.

[0159] In a particular embodiment, the machine learning algorithm in step V) is a random forest algorithm.

[0160] In one embodiment, the method of the application further comprises evaluating and / or validating the sensitivity, specificity, accuracy, ROC and the like of the established receptivity prediction model in a separate test set split out.

[0161] In a particular embodiment, the method of the application further comprises validating the established receptivity prediction model in at least two test sets.

[0162] In another aspect, the application also provides a biomarker set for predicting endometrial receptivity to guide embryo transfer.

[0163] In one embodiment, the biomarker set of the present application comprises some or all of the differentially expressed genes identified in step III) above, for example any number between more than 5 and less than or equal to 250. In a more particular embodiment, in order to accurately predict endometrial receptivity, to guide embryo transfer, the biomarker set of the present application comprises at least 10 gene members. In an even more particular embodiment, the biomarker set of the present application comprises at least 11 gene members, at least 12 gene members, at least 13 gene members, at least 14 gene members, at least 15 gene members, at least 16 gene members, at least 17 gene members, at least 18 gene members, at least 19 gene members, at least 20 gene members, at least 25 gene members, at least 30 gene members, at least 35 gene members, at least 40 gene members, at least 45 gene members, at least 50 gene members, at least 55 gene members, at least 60 gene members, at least 65 gene members, at least 70 gene members, at least 80 gene members, at least 90 gene members, at least 100 gene members, at least 110 gene members, at least 120 gene members, at least 130 gene members, at least 140 gene members, at least 150 gene members, at least 160 gene members, at least 170 gene members, at least 180 gene members, at least 190 gene members, at least 200 gene members, at least 210 gene members, at least 220 gene members, at least 230 gene members, at least 240 gene members, at least 250 gene members. In one embodiment, the biomarker set of the present application comprises the 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 40, 50, 60, 70, 80, 90, 100, 110, 120, 130, 140, 150, 160, 170, 180, 190, 200, 210, 220, 230, 240 or 250 genes with the highest weight in the receptivity prediction model established according to step V) above.

[0164] In one embodiment, the biomarker set of the present application comprises at least the following genes: ELK4, PRKACB, PHB2, GM2A, PPA1, NCEH1, CAPG, DDIT4, LDHB and TDRD6. In a preferred embodiment, the biomarker set of the present application comprises the following genes: ELK4, PRKACB, PHB2, GM2A, PPA1, NCEH1, CAPG, DDIT4, LDHB and TDRD6 and further comprises 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 40, 50, 60, 70, 80, 90, 100, 110, 120, 130, 140, 150, 160, 170, 180, 190, 200, 210, 220, 230 or 240 of the gene markers listed in Table 2, in sequence.

[0165] In a particular aspect, the endometrial receptivity prediction model of the present application comprises one or more sub-models, wherein different sub-models involve different biomarker sets as described above, respectively.

[0166] In yet another aspect, the present application provides uses and methods based on the biomarker sets.

[0167] In one embodiment, the present application provides uses of the above-mentioned biomarker sets or specific detection tools thereof for preparing assay compositions, assay reagents and / or assay devices for detecting and determining the endometrial receptivity status of a subject, wherein the assay compositions, assay reagents and / or assay devices are capable of detecting the expression level of each member of the above-mentioned biomarker sets in a sample from a subject; and the expression level is calculated by the receptivity prediction model of the present application, and the result is used to determine the diagnosis of the endometrial receptivity status.

[0168] In one embodiment, the present application provides a method for predicting endometrial receptivity, comprising obtaining the expression level of all gene members of the biomarker set using a substance or method for specifically determining the level or amount of the biomarker. In a specific embodiment, the substance for determining the level or amount of the biomarker is a composition, reagent or device. In a specific embodiment, the substance for determining the level or amount of the biomarker is a specific primer and / or probe molecule. In a more specific embodiment, the composition or reagent for determining the level or amount of the biomarker further comprises one or more of a tag sequence, a linker sequence, a fluorescent molecule, a pyrophosphate molecule and a label molecule. In a specific embodiment, the method for determining the level or amount of the biomarker is a PCR method. In a more specific embodiment, the PCR method is a real-time fluorescent quantitative PCR, isothermal amplification PCR and the like commonly used in the art. In a specific embodiment, the method for determining the level or amount of the biomarker is a sequencing method. In a more specific embodiment, the sequencing method is one or more of next-generation sequencing (NGS), Sanger sequencing and third-generation sequencing (e.g., nanopore sequencing).

[0169] In one embodiment, the reagent for determining the level or amount of the biomarker does not comprise a reagent for specifically detecting genes other than the members of the biomarker set of the present application.

[0170] In one embodiment, the method for determining the level or amount of the biomarker does not determine the level or amount of genes other than the members of the biomarker set of the present application.

[0171] In one embodiment, the present application provides a method for predicting endometrial receptivity, wherein the sample used in the method is obtained by a non-invasive method, i.e., the sample is obtained without causing damage to the endometrium of the pregnant subject. In a preferred embodiment, the present application provides an integrated non-invasive method for predicting endometrial receptivity, wherein "integrated" means that the method does not need to be performed separately and additionally, but is performed simultaneously with the operation of pre-implantation of the embryo or implantation of the embryo, and the sampling process is integrated into the above operation, so that the sampling period is the implantation period. For example, when the cycle implantation fails and the detection result suggests that the endometrium is the cause, the detection result can be referred to guide the implantation of the next cycle, without the need to wait, greatly shortening the time to pregnancy of the pregnant subject.

[0172] In yet another aspect, the present application provides a kit for use in the above-mentioned methods, said kit comprising reagents for determining the level or amount of each biomarker in the combination of biomarkers of the present application in a sample, such as specific and general reagents.

[0173] Embodiments of the present application, and the technical effects thereof, are illustrated by way of example and not limitation in the following examples. Other embodiments, methods, and products are within the skill of persons of ordinary skill in the art of molecular diagnostics, molecular biology, and need not be described in detail herein. Other embodiments falling within the scope of the art are considered part of the present application. Examples

[0174] Example 1 Method for establishing non-invasive sampling of endometrium in subjects awaiting embryo transfer

[0175] The inventors have incidentally discovered that, as one of the important guiding steps of embryo transfer, embryo pre-transfer, the trace amount of tissue cells adhered during its operation are sufficient for performing gene differential expression analysis related to endometrial receptivity.

[0176] In order to determine the stability and reliability of the above-mentioned discovery, the inventors collected 243 cases of trace amount of tissue cells adhered to the operation tube of the embryo pre-transfer from the subjects.

[0177] After the (pre) transfer operation, the wall of the transfer tube was cut and flushed, so that the extremely small amount of endometrial tissue or cells adhered thereto was transferred to the sample collection tube. The tissue sample preservation solution (Century Kangwei, CW0592M) in the sample collection tube can protect the tissue cell sample to ensure the subsequent total RNA extraction.

[0178] Clinical sampling scenario 1 (total of 12 cases): the transfer tube brought out visible tissue, the inner and outer walls of the transfer catheter were flushed in a culture dish using 4 degrees Celsius PBS, and the small piece of endometrial tissue was picked up and transferred to the ERT tissue preservation tube (1.5 mL) using a non-nuclease suction head (the tube contains 0.3 ml of preservation solution).

[0179] Clinical sampling scenario 2 (total of 231 cases): no visible tissue, a small amount of cells were recovered by the inner tube flushing method combined with the cutting method. Inner tube flushing method: replace a new 1 mL syringe, flush the transfer catheter with 100-150 uL of PBS, and inject the flushing liquid directly into the ERT tissue preservation tube (1.5 mL) (the tube contains 0.5 ml of preservation solution), and according to the actual situation, the above step can be repeated 1 time. Cutting method: cut about 5 mm of the outer tube and 8-10 mm of the inner tube of the transfer catheter, and directly place them in the ERT tissue preservation tube (1.5 mL) (the tube contains 0.3 ml of preservation solution).

[0180] ​Subsequently, total RNA is extracted as completely as possible using a trace cell RNA extraction method. The cDNA library is obtained by reverse transcription and amplification using the MALBAC platinum trace RNA amplification kit (KT110700724) developed by Yikang. The extraction method is performed according to the following steps:

[0181] a) High-speed centrifugation is performed on the sample (1600g for the first time, 2 minutes; 3000g for the second time, 3 minutes; 13000 rpm for the third time, 5 minutes) to precipitate all the cells.

[0182] b) The supernatant is removed, 200 μL x 2 tissue sample preservation solution is added to resuspend the cells, and centrifugation is performed at 3000g for 5 minutes; 200 μL x 1 time, and about 50 μL is left.

[0183] c) 300 μL of cell lysis solution is added, and vigorous shaking is performed for 30 seconds to 1 minute; instant centrifugation is performed, and the supernatant is discarded. d) 350 μL of freshly prepared 70% ethanol is added and shaken to mix; instant centrifugation is performed, and the supernatant is discarded.

[0184] e) 650 μL of the above mixture is added to the adsorption column, and centrifugation is performed at 13000 x g for 1 minute, and the centrifugation waste liquid is discarded.

[0185] f) 500 μL of the washing buffer is added to the adsorption column membrane, and centrifugation is performed at 13000 x g for 1 minute, and the centrifugation waste liquid is discarded.

[0186] g) 100 μL of DNA digestion enzyme is carefully dropped on the adsorption column membrane, and incubation is performed at 25 degrees for 15 minutes.

[0187] h) 500 μL of the washing buffer is added, and centrifugation is performed at 13000 x g for 1 minute, and the centrifugation waste liquid is discarded; 500 μL of the washing buffer is added to the adsorption column again, and centrifugation is performed at 13000 x g for 1 minute, and the centrifugation waste liquid is discarded.

[0188] i) 700 μL of 80% ethanol is added to the tube, and centrifugation is performed at 13000 x g for 2 minutes, and the centrifugation waste liquid is discarded; centrifugation is performed again at 13000 x g for 1 minute, and the collection tube is discarded.

[0189] j) The adsorption column is inserted into a 1.5 mL centrifuge tube, and the cover is opened to dry for 1 minute; 20 μL of DEPC water is added, and incubation is performed at 25 degrees for 1 minute, and centrifugation is performed at 13000 x g for 2 minutes, and the process is repeated once again, and a total of 40 μL of eluent is collected, which is the total RNA. The extraction can also be performed according to the instructions of the kit. Reverse transcription is performed according to the instructions of the kit (KT110700724) or according to the reverse transcription operation mode described in the specification of Chinese invention patent application No. 201810019044.7.

[0190] The cDNA library is obtained, as shown in Table 1 below:

[0191] Table 1

[0192] cDNA preparation Sequencing library preparation Final prediction result Number of successes 243 228 224 Total number of cases 243 243 243 Success rate 100.0％ 93.8％ 92.2％

[0193] The results show that 243 trace cells can all successfully obtain reliable cDNA library of quality. And in the subsequent method, 228 cases are successfully sequenced, and 224 cases finally obtain accurate receptivity prediction results according to the technical scheme of the application (see subsequent examples).

[0194] The inventors further measure the nucleic acid concentration of the obtained cDNA library, and deduce the number of cells in the non-invasive sample, and set up two groups of samples for comparison: the conventional method group and the non-invasive sample group.

[0195] The concentration of the cDNA library obtained from 243 non-invasive samples is 45.9±30.4 ng / μL on average;

[0196] The conventional method group, i.e. the samples obtained by clinical endometrial biopsy, endometrial exfoliated cell sampling and other invasive methods, a total of 7 cases, 10 ng of total RNA was taken for cDNA library construction, and the concentration of the obtained cDNA library was 114.5±12.4 ng / μL on average.

[0197] Based on the 10 ng of total RNA of the conventional method group, the cell amount is about 2300-2700 cells per sample, combined with the efficiency curve of cDNA amplification, it is calculated that the cell amount of the non-invasive sample group is about 300-1000 cells per sample.

[0198] Therefore, through this embodiment, the inventors establish a new method for obtaining endometrial receptivity prediction samples by non-invasive sampling method, and verify its stability and reliability.

[0199] For subjects who have performed embryo pre-implantation operation, samples can be obtained by the non-invasive sampling method of the application, and the endometrial receptivity can be predicted to guide the subsequent embryo transfer, greatly improving the success rate.

[0200] For subjects who have performed embryo transfer operation, samples can also be obtained by the non-invasive sampling method of the application, and the endometrial receptivity can be predicted. If this time of implantation fails, the prediction result can be referred to guide the subsequent embryo transfer, and more information can be obtained with as little operation as possible, reducing the pain of the subjects.

[0201] Due to the relatively scarce cell content in the sample and the different cell types from conventional samples, it brings new challenges to the establishment of predictive models, which will be established and improved through more subsequent examples.

[0202] Example 2: Amplification and sequencing of samples obtained by non-invasive sampling

[0203] For the cDNA library obtained in Example 1, the sequencing library construction was performed using the Illumina Fragmentation Kit (KT100804248) and the Gene Sequencing Library Kit (KT100804048). The specific operation steps were performed according to the instructions, or according to the reverse transcription operation mode described in the specification of Chinese Invention Patent Application No. 201810019044.7.

[0204] The obtained sequencing library was sequenced using the Illumina Nextseq sequencing system and the matching Next-seq High-Output Kit reagent kit, and the off-machine data was obtained. Those skilled in the art can also use any commonly used sequencing scheme in the art to sequence the obtained library.

[0205] Example 3 Selection and analysis of differentially expressed genes from sequencing data as a set of biomarkers of endometrial receptivity and establishment of a prediction model using a random forest algorithm. Example 4

[0206] The samples described in Example 1 were classified according to the final actual clinical outcome, and were divided into receptive phase and pre-receptive phase (i.e. non-receptive phase), and the specific division method is as follows:

[0207] If the sample is derived from the embryo of the subject at the D5 period of transplantation, and the subsequent embryo successfully implants, it indicates that the sample under test is in a receptive state at the time of transplantation; if the embryo does not successfully implant, the sample under test is in a pre-receptive state.

[0208] If the sample is derived from the embryo of the subject at the D3 period of transplantation, and the subsequent embryo successfully implants, it indicates that the sample under test is in a pre-receptive state at the time of transplantation; if the embryo does not successfully implant, the sample under test is in a receptive state.

[0209] The samples were numbered with D5+ serial number or D3+ serial number, and their receptive states were marked according to the clinical outcome gold standard. The 243 samples were divided into two groups: a training set for supervised training of machine learning algorithms to establish a prediction model, including 211 samples; and a test set for subsequent independent testing of the established model, including 32 samples.

[0210] 3.1 Identification of differentially expressed genes

[0211] The sequencing off-machine data obtained in Example 2 was used to analyze the relative expression amount of each gene, and then the gene expression amount between the two types of samples was compared, so as to identify the differentially expressed genes.

[0212] Specifically, the off-machine data of each sample was processed as follows:

[0213] ​i. The data from the machine is processed using Trimmomatic (version 0.33) to remove connector sequences and filter low-quality reads.

[0214] This produces cleaned sequencing data.

[0215] ii. Align the clean sequencing data generated in the previous step to the human reference genome (version GRCh38) using HISAT2 (version 2.0.5) to generate an alignment file.

[0216] iii. Sort the alignment files generated in the previous step and mark the duplicate sequences to generate the final alignment files.

[0217] iv. Process the alignment file generated in the previous step using htseq_count. This step also requires a gene annotation file, namely the Homo_sapiens.GRCh38.84.gtf file, which can be downloaded from Ensembl. It contains information such as the name and location coordinates of each gene. htseq_count can use this information to count how many sequences are in the region where each gene is located (for paired-end sequencing, reads belonging to the same fragment are only counted once during the statistics) and generate a genecount file.

[0218] v. Merge the genecount files of all samples in a batch to generate a total genecount file, where each row represents a sample and each column represents a gene, with the data being the number of fragments belonging to that gene in that sample.

[0219] Thus, the relative expression level of each gene in each specific sample was obtained. Then, the differentially expressed genes with application significance were identified through the following specific steps:

[0220] i. Exclude rRNA-related and mitochondrial-related genes that cause significant subsequent interference.

[0221] ii. For the training set, edgeR (version 3.34.0) was used for differential gene expression analysis:

[0222] a) edgeR requires two input files: the first is a genecount file, which contains the number of reads for each gene in each sample; the second is a sample information file, which contains the category to which the sample belongs.

[0223] (i.e., the acceptance period or the pre-acceptance period);

[0224] b) After edgeR reads the input file, it first normalizes the number of reads for each gene in each sample. During normalization, genes that are stably expressed in all samples are selected.

[0225] c)edgeR then according to the category information of the samples, find the genes that are differentially expressed in different categories, the significance of the difference is represented by the FDR corrected p-value, the smaller the p-value, the more significant the difference. The screening condition is set to logFoldChange>0, FDR<0.05, and the differentially expressed genes in the two samples are screened out, and 3193 genes are obtained.

[0226] iii. The 3193 genes screened in the last step are further subjected to variable screening, which is to select genes with FDR<0.001 from differentially expressed genes, and 1270 genes are obtained. The logTMM value of each sample and each gene is output by the cpm function of edgeR, and the pearson correlation between each other is calculated. Then the genes are screened according to the following steps:

[0227] 1) The genes are sorted according to FDR from small to large;

[0228] 2) Mark the top of the list as selected;

[0229] 3) Mark the genes with correlation greater than or equal to 0.6 with any of the selected genes as rejected;

[0230] 4) Repeat 2) and 3) until the end of the list;

[0231] 5) Use the screened genes to build a model, and use 10-fold cross-validation method to evaluate the accuracy, and delete the genes that have no effect on the accuracy of the model one by one, and finally obtain 250 genes.

[0232] 3.2 Use differentially expressed genes to establish a prediction model

[0233] The expression data of 250 differentially expressed genes obtained in the last step are used as features to further use machine learning algorithm to establish a prediction model for judging the endometrial receptivity state from gene expression data.

[0234] 1. Data processing:

[0235] Convert genecount data to TPM to exclude the influence of sample down-machine data volume. The calculation formula is as follows:

[0236]

[0237] Where Ni is the reads count on the ith gene, and Li is the length of the ith gene.

[0238] 2. Model selection: Random forest model is an ensemble learning method, which builds multiple decision tree models and then combines them to get a more accurate and stable model. Moreover, the model can also get good results when using default parameters, which is one of the commonly used models at present.

[0239] (1). The commonly used hyperparameters of random forest model are as follows:

[0240] a) n_estimators, the number of decision trees constructed, usually the larger the better, but the calculation time will also increase, when the number of decision trees reaches a certain value, the effect will not be better

[0241] b) max_features, the number of randomly selected features when constructing each decision tree, the smaller the value, the lower the similarity between decision trees

[0242] c) max_sample, the number or proportion of samples used when constructing each decision tree, the lower the value, the lower the similarity between decision trees

[0243] d) class_weight, the proportion of samples of each class, by default, the proportion of samples of all classes is the same, for unbalanced samples, balanced can be used

[0244] e) criterion, the index used to evaluate the quality of each partition when constructing a decision tree

[0245] f) max_depth, the maximum depth of each decision tree

[0246] g) min_samples_split, the minimum number of samples required for decision tree splitting

[0247] h) min_samples_leaf, the minimum number of samples for each node of the decision tree

[0248] (2). Model training: the process of model training is to find a set of optimal hyperparameters for the construction of the final model. The methods used in this training are as follows

[0249] i. Random method, randomly combine hyperparameters, search for a period of time and select the optimal parameters, when tuning, use accuracy to evaluate the model respectively;

[0250] ii. Train the model with default parameters;

[0251] iii. Select the optimal method from the two methods, the evaluation index is the average accuracy of 10-fold cross-validation (10-fold cv).

[0252] The following software was used to train the model: report caret (6.0_88), randomForest (4.6_14) and scikit-learn (0.24.1).

[0253] The importance of individual genes was ranked by the random forest algorithm in the training set split for machine learning as shown in Table 2 below:

[0254] Table 2 (values in the table are TPM)

[0255]

[0256]

[0257]

[0258]

[0259]

[0260]

[0261]

[0262]

[0263] As shown in the above table, the TPM values of the differentially expressed genes were used to train the algorithm by implementing the random forest algorithm, by randomly generating multiple decision trees and obtaining the prediction results, and comparing the prediction results with the actual results of the samples. After a sufficient machine learning process, the present application successfully established a prediction model for endometrial receptivity of non-invasive samples. The model can obtain correct prediction results for all samples after training with samples in the training set. Moreover, by evaluating in an independent test set, its high accuracy, i.e. the universality to other samples, was confirmed, which has clinical application significance.

[0264] As shown in the results in Table 2, the model gives a high-to-low ranking according to the importance of the identified individual genes in prediction. The higher the ranking of the gene, the higher the accuracy of the judgment when it is used as a marker alone. For example, the gene ELK4 (ENSG00000158711) ranked first, which can obtain an accuracy of 60% when judging according to its expression in the resulting model.

[0265] Obviously, the complex physiological process in vivo must involve the participation of multiple genes or multiple pathways, so the 60% prediction accuracy obtained by using a single marker is obviously insufficient. By sequentially adding genes features to build sub-models, the inventors found that a sub-model including at least the top 10 gene members (i.e., ELK4, PRKACB, PHB2, GM2A, PPA1, NCEH1, CAPG, DDIT4, LDHB, and TDRD6) can achieve higher accuracy and has application value.

[0266] In other words, in the samples obtained by the non-invasive sampling method discovered and adopted by the present application, the differential expression of the above-mentioned 250 genes reflects the receptivity status of the sample. Among them, the expression of the 10 genes ELK4, PRKACB, PHB2, GM2A, PPA1, NCEH1, CAPG, DDIT4, LDHB, and TDRD6 has the most correlation with the receptivity status and can be used as a molecular marker of the receptivity status in non-invasive samples. If more markers from the 250 genes are included on the basis of the 10 genes for comprehensive consideration, more accurate prediction results can be obtained.

[0267] In order to verify this conclusion identified in the training set, the model of the present application was used for the separately divided test set according to the previous experimental design to test the accuracy of the model of the present application in samples different from the training set (i.e., further verify the universality of the marker set and the prediction model).

[0268] Validation of the set of biomarkers of endometrial receptivity and the prediction model in a test set Figures 1-13

[0269] The independent test set was used to evaluate the sensitivity, specificity, accuracy, ROC, and other indicators of the model trained in Example 3.

[0270] ​​Feature 10 refers to a sub-model consisting of the differentially expressed genes ELK4, PRKACB, PHB2, GM2A, PPA1, NCEH1, CAPG, DDIT4, LDHB and TDRD6 ranked 1-10 in Table 2, and its predictive accuracy was evaluated. Similarly, Feature 30 refers to a model consisting of genes ranked 1-30 in Table 2, and so on. In this example, Feature 10, Feature 30, Feature 50, Feature 70, Feature 90, Feature 110, Feature 130, Feature 150, Feature 170, Feature 190, Feature 210, Feature 230, Feature 250 (Feature 250 contains all the markers) were calculated.

[0271] The results are shown below and in ​

[0272] feature 10

[0273] Model output results

[0274]

[0275] Model accuracy

[0276]

[0277] feature 30

[0278] Model output results

[0279]

[0280]

[0281] Model accuracy

[0282]

[0283] feature 50

[0284] Model output results

[0285]

[0286]

[0287] Model accuracy

[0288] ​

[0289] feature 70

[0290] model output

[0291]

[0292]

[0293] model accuracy

[0294]

[0295] feature 90

[0296] model output

[0297]

[0298]

[0299] model accuracy

[0300]

[0301] feature 110

[0302] model output

[0303]

[0304]

[0305] model accuracy

[0306]

[0307] feature 130

[0308] model output

[0309]

[0310]

[0311] model accuracy

[0312]

[0313] feature 150

[0314] model output

[0315]

[0316]

[0317] model accuracy

[0318]

[0319] feature 170

[0320] model output

[0321]

[0322]

[0323] model accuracy

[0324]

[0325] feature 190

[0326] model output

[0327]

[0328] model accuracy

[0329]

[0330] feature 210

[0331] model output

[0332]

[0333]

[0334] model accuracy

[0335]

[0336] feature 230

[0337] model output

[0338]

[0339]

[0340] model accuracy

[0341]

[0342] feature 250

[0343] model output

[0344]

[0345]

[0346] Model accuracy

[0347]

[0348] As the above results show, the sub-model consisting of the top 10 markers has a prediction sensitivity as high as 89% and a prediction specificity of 85%, with a prediction accuracy of 87.5% (i.e., accurately predicting the receptivity status of 28 out of 32 samples), which is of sufficient clinical significance.

[0349] On the basis of the 10 markers, further markers ranked later can also be incorporated into the model to improve the detection accuracy to a certain extent.

[0350] It can be seen that the differentially expressed gene markers identified in Example 3 using the training set also reflect the correlation with endometrial receptivity in the test set, and a stable and reliable prediction model is established by evaluating the weight of each gene using a machine learning algorithm.

[0351] In summary, the present application first discovers that the gene expression information carried by trace amounts of non-invasive samples is sufficient for predicting endometrial receptivity, and proposes a method for predicting endometrial receptivity using non-invasive samples. The present application also first establishes a marker set and a prediction model for predicting endometrial receptivity using non-invasive samples. Under the condition of almost no increase in the pain of the subjects, the success rate of subsequent embryo transfer of the subjects is greatly improved, the time to pregnancy is shortened, and the present application has significant clinical significance.

Claims

1. A method for obtaining a sample for predicting endometrial receptivity in a subject, characterized in that, The method is non-invasive, and the number of cells in the resulting sample is preferably less than 500.

2. The method of claim 1, wherein the method comprises obtaining the sample from trace amounts of tissue cells adhering to the wall of the transfer tube during embryo transfer.

3. The method of claim 1, wherein the method comprises obtaining the sample from trace amounts of tissue cells adhering to the wall of the transplant tube during embryo transfer.

4. A method for predicting endometrial receptivity in a subject, characterized in that, Includes the following steps: (1) Obtain a sample by the method of any one of claims 1-3; (2) Detect the expression level of endometrial receptivity-related markers in the sample.

5. The method of claim 4, wherein the endometrial receptivity-related markers are genes numbered 1-10 in Table 2, namely ELK4, PRKACB, PHB2, GM2A, PPA1, NCEH1, CAPG, DDIT4, LDHB, and TDRD6, and optionally further include one or more genes numbered 11-250 in Table 2.

6. A combination of biomarkers, characterized in that, It includes genes numbered 1-10 in Table 2, namely ELK4, PRKACB, PHB2, GM2A, PPA1, NCEH1, CAPG, DDIT4, LDHB, and TDRD6, and optionally also includes one or more genes numbered 11-250 in Table 2.

7. Use of the biomarker combination of claim 6 in predicting endometrial receptivity in a subject.

8. Use of the biomarker combination of claim 6 in the preparation of a kit for predicting endometrial receptivity in a subject.

9. Use of a reagent for detecting the level or amount of the combination of biomarkers of claim 7 or 8 in predicting endometrial receptivity in a subject.

10. A method for establishing a non-invasive tolerance prediction model, comprising the following steps: i. Obtaining samples from subjects by means of any one of claims 1-3; ii. RNA was extracted from the samples and reverse transcribed to obtain cDNA libraries, and the cDNA libraries of all samples were sequenced; iii. Use the sequencing data from step ii. for machine learning algorithms.

11. The method for establishing an acceptability prediction model as described in claim 10, wherein step iii further includes the following specific steps: I) labeling the samples as either acceptable or non-acceptable based on the embryo transfer outcome of the subjects from which the samples originated, and preferably splitting them into two sets, a training set and a test set, and taking the training set sample data for the next step; II) removing noise and bias from all data of the training set samples to generate fragment count alignment files for each gene; III) screening and collecting genes with differential expression as features to be identified; IV) converting the sequencing data from raw data to TPM as standardized expression data for each gene, thereby eliminating the influence of differences in the amount of sequencing data from the samples; V) inputting the expression data of the features (genes) to be identified into a machine learning algorithm, and having the machine learning algorithm perform supervised learning to obtain the threshold, sensitivity, and importance of each feature for judging acceptability, thereby obtaining an acceptability prediction model.

12. The method for establishing an acceptable prediction model as described in claim 10 or 11, wherein the machine learning algorithm is a random forest.

13. A combination product of polynucleotides, characterized in that, The product comprises polynucleotides of genes numbered 1-10 in Table 2, and optionally further includes polynucleotides of one or more genes numbered 11-250 in Table 2, which are differentially expressed in non-invasive samples of different endometrial receptivity states and are endometrial receptivity gene markers.

14. A kit for predicting endometrial receptivity in subjects, comprising reagents for determining the level or amount of each biomarker in a sample from a biomarker group comprising the following biomarkers: genes numbered 1-10 in Table 2, and optionally one or more genes numbered 11-250 in Table 2.

Images