WO2023102786A1 - 基因标志物在预测孕妇早产风险中的应用 - Google Patents
基因标志物在预测孕妇早产风险中的应用 Download PDFInfo
- Publication number
- WO2023102786A1 WO2023102786A1 PCT/CN2021/136566 CN2021136566W WO2023102786A1 WO 2023102786 A1 WO2023102786 A1 WO 2023102786A1 CN 2021136566 W CN2021136566 W CN 2021136566W WO 2023102786 A1 WO2023102786 A1 WO 2023102786A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- pregnant women
- premature
- risk
- gene markers
- membranes
- Prior art date
Links
Images
Classifications
-
- C—CHEMISTRY; METALLURGY
- C12—BIOCHEMISTRY; BEER; SPIRITS; WINE; VINEGAR; MICROBIOLOGY; ENZYMOLOGY; MUTATION OR GENETIC ENGINEERING
- C12Q—MEASURING OR TESTING PROCESSES INVOLVING ENZYMES, NUCLEIC ACIDS OR MICROORGANISMS; COMPOSITIONS OR TEST PAPERS THEREFOR; PROCESSES OF PREPARING SUCH COMPOSITIONS; CONDITION-RESPONSIVE CONTROL IN MICROBIOLOGICAL OR ENZYMOLOGICAL PROCESSES
- C12Q1/00—Measuring or testing processes involving enzymes, nucleic acids or microorganisms; Compositions therefor; Processes of preparing such compositions
- C12Q1/68—Measuring or testing processes involving enzymes, nucleic acids or microorganisms; Compositions therefor; Processes of preparing such compositions involving nucleic acids
Definitions
- the present invention relates to the field of premature delivery of pregnant women, in particular to the application of gene markers in predicting the risk of premature rupture of membranes and unexplained spontaneous premature delivery.
- Preterm birth is defined as birth before 37 weeks of gestation. Globally, preterm birth is the leading cause of death for children under five, and rates are increasing in almost all countries with reliable data. Preterm birth is an important issue in the field of mothers and babies.
- Premature rupture of membranes and unexplained causes can lead to premature labor.
- Premature rupture of membranes refers to the spontaneous rupture of membranes before labor, and premature rupture of membranes at a gestational age less than 37 weeks is called premature rupture of membranes.
- Preventing deaths and complications from preterm birth starts with a healthy pregnancy. Early prediction and early intervention can improve pregnancy outcomes.
- cervical length detection and fFN fetal fibronectin detection in vaginal secretions are clinically used to assess the risk of preterm birth for high-risk groups, but they are mainly aimed at high-risk groups, and the sensitivity and specificity are limited.
- Several studies and patent applications have involved the use of gene expression, metabolites, proteins/peptides, and microbes for preterm birth prediction and diagnosis, but the main problem still lies in the low sensitivity and specificity of these methods for preterm birth risk prediction.
- the main purpose of the present invention is to provide the application of gene markers in predicting the risk of premature rupture of membranes or unexplained spontaneous premature birth, so as to provide a high specificity and high sensitivity prediction scheme for the risk of premature birth.
- a method for predicting the risk of premature rupture of membranes and premature delivery in pregnant women comprising:
- Step S1 Obtain the expression profile of gene markers in biological samples from pregnant women.
- Gene markers include one or more of the following genes: CCNB1IP1, COL9A2, DNAJC13, FAM45A, FBXO38, FZD3, HCK, KIAA1257, KIF2A, LARP1B, LRRC56, PLD5, PROS1, SEPT10, SLC41A3, SPIN1, TMUB1, TUBB4A, UPF1, WDR34, ZBTB10, AC004803.1, AC009779.2, AC011461.1, AC015878.2, AC016727.1, AC022568.1, AC08475 9.3, AC092338.2, AC093249.2, AC103876.1, AC105020.6, AC108099.1, AL031733.2, AL451074.2, AP000688.4, NORAD, PINK1-AS, REV3L-IT1;
- Step S2 Based on the expression profile of the gene markers, the risk of premature rupture of membranes of pregnant women is identified.
- step S2 identifying the risk of premature rupture of membranes and premature delivery of pregnant women is implemented by using the risk prediction model of premature rupture of membranes and premature delivery of pregnant women, and the risk prediction model of premature rupture of membranes and premature delivery of pregnant women is implemented by using Computer-generated expression profiles of gene markers in biological samples from pregnant women with premature rupture of membranes.
- the computer training is implemented by a machine learning method, preferably, the machine learning method includes one or more of the following: generalized linear model, gradient boosting machine, random forest, and support vector machine.
- the biological sample is one or more of the following: plasma, serum, whole blood, urine, amniotic fluid; preferably, the biological sample is collected from the 11th to 25th gestational week of the pregnant woman.
- step S1 the expression profile of gene markers is obtained by quantitatively analyzing the free extracellular RNA in the biological sample
- the extracellular free RNA in the biological sample is quantitatively analyzed by high-throughput sequencing or RT-PCR;
- a high-throughput sequencing method is used to quantitatively analyze the extracellular free RNA in the biological sample.
- kits for predicting the risk of premature rupture of membranes in pregnant women includes detection reagents for gene markers, and the gene markers include one or more of the following genes: CCNB1IP1, COL9A2, DNAJC13, FAM45A, FBXO38, FZD3, HCK, KIAA1257, KIF2A, LARP1B, LRRC56, PLD5, PROS1, SEPT10, SLC41A3, SPIN1, TMUB1, TUBB4A, UPF1, WDR34, ZBTB10, AC004803.1, AC0097 79.2, AC011461. 1.
- the detection reagents for gene markers include probes and/or primers for detecting gene markers; preferably, they are related reagents for preparing RNA of gene markers into high-throughput sequencing libraries.
- the application of the detection reagent of the gene marker in the preparation of the kit for predicting the risk of premature rupture of membranes and preterm birth in pregnant women is provided, the gene marker includes one or more of the following genes: CCNB1IP1, COL9A2, DNAJC13 , FAM45A, FBXO38, FZD3, HCK, KIAA1257, KIF2A, LARP1B, LRRC56, PLD5, PROS1, SEPT10, SLC41A3, SPIN1, TMUB1, TUBB4A, UPF1, WDR34, ZBTB10, AC004803.1, AC009779.2, AC0114 61.1, AC015878 .2 ⁇ AC016727.1 ⁇ AC022568.1 ⁇ AC084759.3 ⁇ AC092338.2 ⁇ AC093249.2 ⁇ AC103876.1 ⁇ AC105020.6 ⁇ AC108099.1 ⁇ AL031733.2 ⁇ AL451074.2 ⁇ AP000688.4 ⁇ NORAD ⁇ PINK1
- the detection reagents for gene markers include probes and/or primers for detecting gene markers; preferably, they are related reagents for preparing RNA of gene markers into high-throughput sequencing libraries.
- a device for predicting the risk of premature rupture of membranes and premature delivery in pregnant women has a built-in risk prediction model for premature rupture of membranes and premature delivery in pregnant women.
- the expression profiles of gene markers in the biological samples of preterm pregnant women are trained by computer, and the gene markers include one or more of the following genes: CCNB1IP1, COL9A2, DNAJC13, FAM45A, FBXO38, FZD3, HCK, KIAA1257, KIF2A, LARP1B , LRRC56, PLD5, PROS1, SEPT10, SLC41A3, SPIN1, TMUB1, TUBB4A, UPF1, WDR34, ZBTB10, AC004803.1, AC009779.2, AC011461.1, AC015878.2, AC016727.1, AC022568.1, AC084 759.3 , AC092338.2, AC093249.2, AC103876.1, AC105020.6, AC108099.1,
- a method for constructing a risk prediction model for premature rupture of membranes in pregnant women comprising:
- the remaining part of the group of pregnant women with premature rupture of membranes and the remaining group of pregnant women with full-term delivery are used as a verification set, and the verification set is used to verify the risk prediction model of premature rupture of membranes in pregnant women;
- the best gene markers include one or more of the following genes: CCNB1IP1, COL9A2, DNAJC13, FAM45A, FBXO38, FZD3, HCK, KIAA1257, KIF2A, LARP1B, LRRC56, PLD5, PROS1, SEPT10, SLC41A3, SPIN1, TMUB1, TUBB4A, UPF1, WDR34, ZBTB10, AC004803.1, AC009779.2, AC011461.1, AC015878.2, AC016727.1, AC022568.1, AC084759.3, AC092338.2, AC093249.2, AC103876.1, AC 105020. 6. AC108099.1, AL031733.2, AL451074.2, AP000688.4, NORAD, PINK1-AS, REV3L-IT1.
- the biological sample is one or more of the following: plasma, serum, whole blood, urine, amniotic fluid; preferably, the biological sample is collected from the 11th to 25th gestational week of the pregnant woman.
- the computer training is implemented by a machine learning method, preferably, the machine learning method includes one or more of the following: generalized linear model, gradient boosting machine, random forest, and support vector machine.
- a computer-readable storage medium includes a stored program, wherein, when the program is running, the device where the storage medium is located is controlled to execute the method for predicting premature fetal membranes in pregnant women according to the first aspect of the present invention.
- a processor is provided, and the processor is used to run a program, wherein, when the program is running, the method for predicting the risk of premature rupture of membranes in a pregnant woman according to the first aspect of the present invention or the fifth method of the present invention is executed.
- a method for predicting the risk of unexplained spontaneous premature birth in pregnant women comprising:
- Step S1 Obtain the expression profile of gene markers in biological samples from pregnant women.
- Gene markers include one or more of the following genes: AKAP2, CCNB1IP1, CEACAM19, EMP3, FAR1, FOXN3, GSAP, GTF3C2, HPS3, MTURN, NR1D2, PIK3CG, TMUB1, UPF1, WDR34, ZFR, AC005332.6, AC016727.1, AC018716.2, AC021087.2, AC022613.3, AC084759.3, AC092338.2, AC093525.9, AC099689.1, AC10 5020. 6.
- Step S2 Based on the expression profile of the gene markers, the risk of unexplained spontaneous preterm birth of pregnant women is identified.
- step S2 identifying the risk of unexplained spontaneous preterm birth of pregnant women is implemented by using the risk prediction model of unexplained spontaneous preterm birth for pregnant women, and the risk prediction model of pregnant women’s unexplained spontaneous preterm Expression profiles of gene markers in biological samples from pregnant women were trained to generate a computer.
- the computer training is implemented by a machine learning method, preferably, the machine learning method includes one or more of the following: generalized linear model, gradient boosting machine, random forest, and support vector machine.
- the biological sample is one or more of the following: plasma, serum, whole blood, urine, amniotic fluid; preferably, the biological sample is collected from the 11th to 25th gestational week of the pregnant woman.
- step S1 the expression profile of gene markers is obtained by quantitatively analyzing the free extracellular RNA in the biological sample
- the extracellular free RNA in the biological sample is quantitatively analyzed by high-throughput sequencing or RT-PCR;
- a high-throughput sequencing method is used to quantitatively analyze the extracellular free RNA in the biological sample.
- kits for predicting the risk of unexplained spontaneous premature birth in pregnant women includes detection reagents for genetic markers, and the genetic markers include one or more of the following genes: AKAP2, CCNB1IP1 , CEACAM19, EMP3, FAR1, FOXN3, GSAP, GTF3C2, HPS3, MTURN, NR1D2, PIK3CG, TMUB1, UPF1, WDR34, ZFR, AC005332.6, AC016727.1, AC018716.2, AC021087.2, AC022613.3, AC0847 59 .3, AC092338.2, AC093525.9, AC099689.1, AC105020.6, AL109936.2, AL138921.1, AL606760.3, AP000688.4, FP671120.4, LINC00221, LINC00511, LINC00689, LINC02 076.TTLL10-AS1 .
- the detection reagents for gene markers include probes and/or primers for detecting gene markers; preferably, they are related reagents for preparing RNA of gene markers into high-throughput sequencing libraries.
- the application of detection reagents for gene markers in the preparation of kits for predicting the risk of unexplained spontaneous premature birth in pregnant women is provided.
- the gene markers include one or more of the following genes: AKAP2, CCNB1IP1, CEACAM19, EMP3, FAR1, FOXN3, GSAP, GTF3C2, HPS3, MTURN, NR1D2, PIK3CG, TMUB1, UPF1, WDR34, ZFR, AC005332.6, AC016727.1, AC018716.2, AC021087.2, AC022613.3, AC084759.3, AC092338.2, AC093525.9, AC099689.1, AC105020.6, AL109936.2, AL138921.1, AL606760.3, AP000688.4, FP671120.4, LINC00221, LINC00511, LINC00689, LINC0207 6. TTLL10-AS1.
- the detection reagents for gene markers include probes and/or primers for detecting gene markers; preferably, they are related reagents for preparing RNA of gene markers into high-throughput sequencing libraries.
- a device for predicting the risk of unexplained spontaneous preterm birth in pregnant women has a built-in risk prediction model for unexplained spontaneous preterm birth in pregnant women.
- the expression profiles of gene markers in the biological samples of pregnant women were trained to generate computer-generated gene markers, including one or more of the following genes: AKAP2, CCNB1IP1, CEACAM19, EMP3, FAR1, FOXN3, GSAP, GTF3C2, HPS3, MTURN, NR1D2 , PIK3CG, TMUB1, UPF1, WDR34, ZFR, AC005332.6, AC016727.1, AC018716.2, AC021087.2, AC022613.3, AC084759.3, AC092338.2, AC093525.9, AC099689.1, AC105020 .6 , AL109936.2, AL138921.1, AL606760.3, AP000688.4, FP671120.4,
- a method for constructing a risk prediction model for pregnant women with unexplained spontaneous premature birth includes:
- the best gene markers include one or more of the following genes: AKAP2, CCNB1IP1, CEACAM19, EMP3, FAR1, FOXN3, GSAP, GTF3C2, HPS3, MTURN, NR1D2, PIK3CG, TMUB1, UPF1, WDR34, ZFR, AC005332. 6.
- the biological sample is one or more of the following: plasma, serum, whole blood, urine, amniotic fluid; preferably, the biological sample is collected from the 11th to 25th gestational week of the pregnant woman.
- the computer training is implemented by a machine learning method, preferably, the machine learning method includes one or more of the following: generalized linear model, gradient boosting machine, random forest, and support vector machine.
- a computer-readable storage medium includes a stored program, wherein, when the program is running, the device where the storage medium is located is controlled to execute the eighth aspect of the present invention for predicting unknown causes of pregnant women A method for the risk of spontaneous premature birth or a method for constructing a risk prediction model for spontaneous premature birth of unknown cause in pregnant women according to the twelfth aspect of the present invention.
- a fourteenth aspect of the present invention there is provided a processor, wherein the processor is used to run a program, wherein, when the program is running, the method for predicting the risk of unexplained spontaneous premature birth in pregnant women or
- the twelfth aspect of the present invention relates to a method for constructing a risk prediction model for unexplained spontaneous premature birth in pregnant women.
- the present invention aims at the low prediction accuracy of the risk of preterm birth in the prior art, and proposes to use the gene marker of the present application as the detection target, through the expression profile of the gene marker and the risk of preterm birth due to premature rupture of membranes and unexplained spontaneous premature birth
- the correlation between the two methods has achieved high specificity and high sensitivity risk prediction for the risk of premature rupture of membranes and unexplained spontaneous premature birth.
- Fig. 1 shows a histogram of gestational weeks of collection of biological samples from pregnant women according to a preferred embodiment of the present invention
- Fig. 2 shows the histogram of the interval of gestational weeks between delivery and collection of biological samples among pregnant women according to a preferred embodiment of the present invention
- Fig. 3 shows a flow chart of screening gene markers according to a preferred embodiment of the present invention
- Fig. 4 shows the construction flowchart of the premature birth risk prediction model according to the preferred embodiment of the present invention
- Fig. 5 shows the importance sorting diagram of the best gene markers for predicting premature rupture of membranes and premature labor according to a preferred embodiment of the present invention and the AUC curve diagram predicted by the model;
- Fig. 6 shows the importance ranking diagram of the best gene markers for predicting unexplained spontaneous premature birth and the AUC curve diagram predicted by the model in a preferred embodiment of the present invention.
- this application compares the gene expression differences between the preterm group and the full-term group in the first and second trimesters, combined with machine learning algorithms, screens out the genetic markers that predict the risk of preterm birth, and realizes this by building a model High-accuracy prediction of preterm birth in the second trimester.
- the gene markers and prediction model of the present invention have high specificity and sensitivity for the prediction of premature birth risk, especially premature rupture of membranes and unexplained spontaneous premature birth, and can detect premature birth of pregnant women with high accuracy in the second trimester risk, enabling early intervention.
- a method for predicting the risk of premature rupture of membranes and premature delivery in pregnant women comprising:
- Step S1 Obtain the expression profile of gene markers in the biological sample from the pregnant woman, the gene markers include one or more of the following genes: CCNB1IP1, COL9A2, DNAJC13, FAM45A, FBXO38, FZD3, HCK, KIAA1257, KIF2A, LARP1B, LRRC56, PLD5, PROS1, SEPT10, SLC41A3, SPIN1, TMUB1, TUBB4A, UPF1, WDR34, ZBTB10, AC004803.1, AC009779.2, AC011461.1, AC015878.2, AC016727.1, AC022568.1, AC084759. 3.
- genes include one or more of the following genes: CCNB1IP1, COL9A2, DNAJC13, FAM45A, FBXO38, FZD3, HCK, KIAA1257, KIF2A, LARP1B, LRRC56, PLD5, PROS1, SEPT10, SLC41A3, SPIN1, TMUB1,
- Step S2 Based on the expression profile of the gene markers, the risk of premature rupture of membranes of pregnant women is identified.
- This application is the first to discover that the gene markers in biological samples of pregnant women have a significant correlation with premature rupture of membranes and preterm birth disease in pregnant women, and thus can be used as markers for predicting premature rupture of membranes and premature birth in pregnant women.
- These gene markers include 21 mRNA genes and 18 lncRNA genes, among which mRNA gene markers include CCNB1IP1, COL9A2, DNAJC13, FAM45A, FBXO38, FZD3, HCK, KIAA1257, KIF2A, LARP1B, LRRC56, PLD5, PROS1, SEPT10, SLC41A3 , SPIN1, TMUB1, TUBB4A, UPF1, WDR34, ZBTB10; lncRNA gene markers include AC004803.1, AC009779.2, AC011461.1, AC015878.2, AC016727.1, AC022568.1, AC084759.3, AC092338.2, AC093249.2, AC103876.1, AC105020.6, AC10
- the gene markers preferably include DNAJC13, CCNB1IP1, AC022568.1, PLD5, WDR34, UPF1, KIF2A, SEPT10, FAM45A, ZBTB10, PROS1, COL9A2, LARP1B, AC009779.2, TMUB1, HCK, AC015878. 2.
- genes listed above can be used alone or in combination.
- a combination of all the following genes can be used as a gene marker: DNAJC13, CCNB1IP1, AC022568.1, PLD5, WDR34, UPF1, KIF2A, SEPT10, FAM45A, ZBTB10, PROS1, COL9A2, LARP1B, AC009779.2, TMUB1, HCK, AC015878.2, AC084759.3, AP000688.4, AC092338.2, so as to realize the risk prediction of premature rupture of membranes and premature birth.
- identifying the risk of premature rupture of membranes and premature delivery of pregnant women can be implemented by using the risk prediction model of premature rupture of membranes and premature delivery of pregnant women, by using the above-mentioned genetic markers in biological samples from pregnant women who have experienced premature rupture of membranes and premature delivery
- the expression profiling of the drug trains a computer to generate a predictive model of preterm birth risk in pregnant women with premature rupture of membranes.
- Training the computer can be implemented by machine learning methods.
- the machine learning method is selected from regression, classification or a combination thereof.
- Machine learning generally refers to algorithms that give computers the ability to learn without being explicitly programmed, including algorithms that learn from data and make predictions about that data.
- the machine learning methods used in the present invention may include random forest, least absolute shrinkage and selection operator logistic regression, regularized logistic regression, XGBoost, decision tree learning, artificial neural network, deep neural network, support vector machine, rule-based machine Learning, Generalized Linear Models, Gradient Boosting Machines, etc.
- Preferred machine learning methods include one or more of the following: generalized linear models, gradient boosting machines, random forests, and support vector machines.
- the risk score automatically calculated by the model can be used to evaluate and predict the risk of premature rupture of membranes and premature delivery. For example, if the risk score is greater than 0.5, the risk of premature rupture of membranes is considered high, and if the risk score is less than 0.5, the risk of premature rupture of membranes is considered low.
- the biological sample derived from a pregnant woman can be one or more of the following: plasma, serum, whole blood, urine, amniotic fluid. Plasma, serum or whole blood derived from pregnant women are preferably used for the detection and identification steps of the present invention.
- the biological sample is most preferably plasma, for example, peripheral blood can be obtained from a pregnant woman and subjected to plasma separation to obtain a plasma biological sample to be used.
- plasma, serum or whole blood other bodily fluid samples such as urine, amniotic fluid, etc. can also be used.
- Biological samples can be obtained by conventional methods in the art.
- the collection of biological samples can be carried out during the 11th to 25th gestational weeks of pregnant women.
- the application population of the present invention does not need to distinguish whether pregnant women are at high risk of premature delivery, and can be applied to general pregnant populations.
- the present invention can realize the prediction of premature rupture of membranes and premature delivery in the second trimester.
- the present invention can achieve preterm birth prediction up to 23 weeks in advance. Therefore, the method of the present invention is applicable to a wider population and has more clinical applicability.
- the expression profile of the gene markers is obtained by quantitative analysis of free extracellular RNA (cfRNA) in the biological sample; preferably, high-throughput sequencing or RT-PCR method is used Quantitative analysis of free extracellular RNA in the biological sample; more preferably, quantitative analysis of free extracellular RNA in the biological sample by next-generation sequencing.
- cfRNA free extracellular RNA
- the free extracellular RNA in the biological sample can be extracted by a method or a kit commonly used in the art or a combination of the two.
- cell-free extracellular RNA can be isolated from plasma biological samples using TRIzol LS standard RNA extraction procedures.
- the quantitative analysis of free extracellular RNA preferably includes sequencing the free extracellular RNA in biological samples (preferably plasma samples) of pregnant women using next-generation sequencing by whole transcriptome sequencing.
- This method can simultaneously sequence plasma free mRNA and free lncRNA.
- RT-PCR method can also be used for analysis.
- the expression profile of extracellular free RNA can also be quantitatively analyzed by other methods known in the art such as qPCR.
- the quantitative analysis of extracellular free RNA also includes the step of quality control of the original extracellular free RNA sequencing data, preferably including cutting adapters, removing low-quality reads, removing ⁇ 17bp reads, and removing rRNA sequence, value RNA and Y RNA sequence, and the remaining read lengths are first compared to the human transcriptome (the sequence is miRNA, tRNA and piRNA, mRNA and lncRNA, and finally other RNA).
- RNA alignment is performed using bowtie software, and quantification is performed using RSEM.
- a prediction kit for the gene markers of the present invention can be prepared. Detection probes, chips, etc. for predicting the risk of premature rupture of membranes and premature delivery in pregnant women can also be prepared for these gene markers.
- the present invention provides a kit for predicting the risk of premature rupture of membranes in pregnant women, the kit includes detection reagents for genetic markers, and the genetic markers include one or more of the following: Genes: CCNB1IP1, COL9A2, DNAJC13, FAM45A, FBXO38, FZD3, HCK, KIAA1257, KIF2A, LARP1B, LRRC56, PLD5, PROS1, SEPT10, SLC41A3, SPIN1, TMUB1, TUBB4A, UPF1, WDR34, ZBTB10, AC004 803.1, AC009779 .2, AC011461.1, AC015878.2, AC016727.1, AC022568.1, AC084759.3, AC092338.2, AC093249.2, AC103876.1, AC105020.6, AC108099.1, AL031733.2, AL451074.
- the genetic markers include one or more of the following: Genes: CCNB1IP1, COL9A2, DNAJC13, F
- the above gene markers include one or more of the following genes: DNAJC13, CCNB1IP1, AC022568.1, PLD5, WDR34, UPF1, KIF2A, SEPT10, FAM45A, ZBTB10, PROS1, COL9A2, LARP1B, AC009779.2, TMUB1, HCK, AC015878.2, AC084759.3, AP000688.4, AC092338.2.
- the detection reagents for gene markers may include probes and/or primers for detecting gene markers, specifically one or more probes and/or primers that specifically bind (hybridize) to gene markers One or more primers that specifically amplify a genetic marker.
- the kit of the present invention may also include reagents for converting RNA in a biological sample into a library of cDNA fragments.
- the application of detection reagents for gene markers in the preparation of kits for predicting the risk of premature rupture of membranes and preterm birth in pregnant women is provided.
- the gene markers include one or more of the following genes: CCNB1IP1, COL9A2 , DNAJC13, FAM45A, FBXO38, FZD3, HCK, KIAA1257, KIF2A, LARP1B, LRRC56, PLD5, PROS1, SEPT10, SLC41A3, SPIN1, TMUB1, TUBB4A, UPF1, WDR34, ZBTB10, AC004803.1, AC009779.2 ⁇ AC011461.1 , AC015878.2, AC016727.1, AC022568.1, AC084759.3, AC092338.2, AC093249.2, AC103876.1, AC105020.6, AC108099.1, AL031733.2, AL451074.2, AP000688.4, NORAD , PINK1-AS, REV3L-
- the gene markers include one or more of the following genes: DNAJC13, CCNB1IP1, AC022568.1, PLD5, WDR34, UPF1, KIF2A, SEPT10, FAM45A, ZBTB10, PROS1, COL9A2, LARP1B, AC009779.2, TMUB1, HCK , AC015878.2, AC084759.3, AP000688.4, AC092338.2.
- the detection reagents of gene markers include probes and/or primers for detecting gene markers, specifically one or more probes that specifically bind (hybridize) to gene markers and/or one or more Primers that specifically amplify gene markers.
- the present invention provides a device for predicting the risk of premature rupture of membranes and premature delivery in pregnant women.
- the device has a built-in risk prediction model for premature rupture of membranes and premature delivery in pregnant women.
- the prediction model is obtained by using sources In the biological samples of pregnant women with premature rupture of membranes, the computer is trained to generate expression profiles of gene markers, which include one or more of the following genes: CCNB1IP1, COL9A2, DNAJC13, FAM45A, FBXO38, FZD3 , HCK, KIAA1257, KIF2A, LARP1B, LRRC56, PLD5, PROS1, SEPT10, SLC41A3, SPIN1, TMUB1, TUBB4A, UPF1, WDR34, ZBTB10, AC004803.1, AC009779.2, AC011461.1, AC015878.2, AC01 6727.1 , AC022568.1, AC084759.3, AC092338.2, AC093249.2, AC103876.1, AC10502
- the gene markers include one or more of the following genes: DNAJC13, CCNB1IP1, AC022568.1, PLD5, WDR34, UPF1, KIF2A, SEPT10, FAM45A, ZBTB10, PROS1, COL9A2, LARP1B, AC009779.2, TMUB1, HCK , AC015878.2, AC084759.3, AP000688.4, AC092338.2.
- the prediction model is a generalized linear model, a gradient boosting machine, a random forest or a support vector machine model.
- a method for constructing a risk prediction model for premature rupture of membranes and premature birth in pregnant women comprising: detecting the group of pregnant women with premature rupture of membranes and premature delivery and the group of pregnant women with full-term delivery Differential expression of gene markers in biological samples; Part of the group of pregnant women with premature rupture of membranes and part of the group of pregnant women with full-term delivery were used as training sets, and the best gene markers were screened out using the training set; in the training set, using The optimal genetic markers train the computer to obtain a risk prediction model for premature rupture of membranes and premature birth; the remaining group of pregnant women with premature rupture of membranes and premature delivery and the remaining group of pregnant women with full-term delivery are used as the verification set, and the verification set is used to verify Preterm birth risk prediction model for pregnant women with premature rupture of membranes; among them, the best genetic markers include one or more of the following genes: CCNB1IP1, COL9A2, DNAJC13, FAM45A, FBXO38, FZD
- the above optimal gene markers include one or more of the following genes: DNAJC13, CCNB1IP1, AC022568.1, PLD5, WDR34, UPF1, KIF2A, SEPT10, FAM45A, ZBTB10, PROS1, COL9A2, LARP1B, AC009779.2, TMUB1, HCK, AC015878.2, AC084759.3, AP000688.4, AC092338.2.
- the biological sample used in the model construction method of the present invention is preferably one or more of the following: plasma, serum, whole blood, urine, amniotic fluid; particularly preferably plasma, serum, whole blood; most preferably plasma. Also, biological samples can be collected from the 11th to 25th week of pregnancy.
- a machine learning method can be used, preferably the machine learning method includes one or more of the following: generalized linear model, gradient boosting machine, random forest and support vector machine.
- the training set and the verification set can be split according to a certain ratio according to needs.
- all pregnant women with premature rupture of membranes are randomly split into the training set and the verification set according to the ratio of 7:3.
- the verification set all pregnant women who gave birth at full term were randomly split into a training set and a verification set according to the ratio of 7:3.
- the screening of the best gene markers is done in the training set, and the validation set is used to test the prediction effect of the best gene markers and models.
- candidate gene markers are preliminarily screened by comparing gene expression profile differences between premature rupture of membranes preterm pregnant women and term pregnant women.
- the gene markers may include mRNA genes and lncRNA genes.
- This step can be performed, for example, using the DESeq2 package (R package).
- R package the DESeq2 package
- the difference and stability of the average expression level in the two populations will be considered in this step (preferably the average expression level difference is greater than or equal to 2, and the corrected p value is less than 0.2), and finally the genes that pass the screening become candidates genetic markers.
- two models can be used to filter based on feature importance. The joint use of the two models is beneficial to ensure the stability of the features.
- generalized linear models and random forests can be used to screen according to the importance of features. For example, 30 most important molecules can be screened out from each screen. The screening process is performed 20 times, and the gene markers with higher frequency of occurrence are selected as the most important molecules. Good genetic markers.
- each algorithm adopts a 7-fold cross-validation method to select the optimal parameters for prediction model construction.
- the resulting model can be validated against the validation set.
- the model with the best effect can be selected and the feature importance can be calculated through the effect verification of the verification set.
- the mRNA gene and the lncRNA gene can be used together as gene markers for effect verification, so as to construct a risk prediction model.
- the prediction model constructed by the method of the present invention can be used in the second trimester and up to 23 weeks in advance, and the risk of premature rupture of membranes can be predicted in a non-invasive way only by taking peripheral blood from pregnant women.
- the predicted The sensitivity can reach 75%
- the specificity can reach 83%
- the area under the receiver operating characteristic curve (AUC) is 0.94 in the training set, and 0.82 in the verification set, both of which are higher than the state of the art.
- a method for predicting the risk of unexplained spontaneous premature birth in a pregnant woman comprising:
- Step S1 Obtain the expression profile of gene markers in the biological sample from the pregnant woman, the gene markers include one or more of the following genes: AKAP2, CCNB1IP1, CEACAM19, EMP3, FAR1, FOXN3, GSAP, GTF3C2, HPS3, MTURN, NR1D2, PIK3CG, TMUB1, UPF1, WDR34, ZFR, AC005332.6, AC016727.1, AC018716.2, AC021087.2, AC022613.3, AC084759.3, AC092338.2, AC093525.9, AC099689. 1.
- genes include one or more of the following genes: AKAP2, CCNB1IP1, CEACAM19, EMP3, FAR1, FOXN3, GSAP, GTF3C2, HPS3, MTURN, NR1D2, PIK3CG, TMUB1, UPF1, WDR34, ZFR, AC005332.6, AC016727.1, AC018716.2, AC021087.2, AC
- Step S2 Based on the expression profile of the gene markers, the risk of unexplained spontaneous preterm birth of pregnant women is identified.
- gene markers in biological samples of pregnant women have a significant correlation with unexplained spontaneous premature birth diseases in pregnant women, and thus can be used as markers for predicting unexplained spontaneous premature birth in pregnant women.
- gene markers include 16 mRNA genes and 20 lncRNA genes, among which mRNA gene markers include AKAP2, CCNB1IP1, CEACAM19, EMP3, FAR1, FOXN3, GSAP, GTF3C2, HPS3, MTURN, NR1D2, PIK3CG, TMUB1, UPF1, WDR34 , ZFR; lncRNA gene markers include AC005332.6, AC016727.1, AC018716.2, AC021087.2, AC022613.3, AC084759.3, AC092338.2, AC093525.9, AC099689.1, AC105020.6, AL109936. 2. AL138921.1, AL606760.3, AP000688.4, FP671120.4, LINC00221,
- the gene markers preferably include FP671120.4, TTLL10-AS1, AL109936.2, LINC02076, AC021087.2, AL606760.3, AC018716.2, LINC00221, LINC00511, AC099689.1, AC005332.6, One or more of AL138921.1, AC093525.9, LINC00689, AP000688.4, AC022613.3, AC105020.6, AC084759.3, AC016727.1, AC092338.2.
- genes listed above can be used alone or in combination.
- a combination of all the following genes can be used as a gene marker: FP671120.4, TTLL10-AS1, AL109936.2, LINC02076, AC021087.2, AL606760.3, AC018716.2, LINC00221, LINC00511, AC099689.1, AC005332. 6.
- identifying the risk of pregnant women with unexplained spontaneous preterm birth can be implemented by using a risk prediction model for pregnant women with unexplained spontaneous preterm birth, by using the expression of the above gene markers in biological samples from pregnant women who have experienced unexplained spontaneous preterm birth Spectrum trains a computer to generate a predictive model for pregnant women's risk of unexplained spontaneous preterm birth.
- Training the computer can be implemented by machine learning methods.
- the machine learning method is selected from regression, classification or a combination thereof.
- Machine learning generally refers to algorithms that give computers the ability to learn without being explicitly programmed, including algorithms that learn from data and make predictions about that data.
- the machine learning methods used in the present invention may include random forest, least absolute shrinkage and selection operator logistic regression, regularized logistic regression, XGBoost, decision tree learning, artificial neural network, deep neural network, support vector machine, rule-based machine Learning, Generalized Linear Models, Gradient Boosting Machines, etc.
- Preferred machine learning methods include one or more of the following: generalized linear models, gradient boosting machines, random forests, and support vector machines.
- the risk score automatically calculated by the model can be used to evaluate and predict the risk of unexplained spontaneous preterm birth. For example, if the risk score is greater than 0.5, the risk of unexplained spontaneous preterm birth is considered high, and if the risk score is less than 0.5, the risk of unexplained spontaneous preterm birth is considered low.
- the biological sample derived from a pregnant woman can be one or more of the following: plasma, serum, whole blood, urine, amniotic fluid. Plasma, serum or whole blood derived from pregnant women are preferably used for the detection and identification steps of the present invention.
- the biological sample is most preferably plasma, for example, peripheral blood can be obtained from a pregnant woman and subjected to plasma separation to obtain a plasma biological sample to be used.
- plasma, serum or whole blood other bodily fluid samples such as urine, amniotic fluid, etc. can also be used.
- Biological samples can be obtained by conventional methods in the art.
- the collection of biological samples can be carried out during the 11th to 25th gestational weeks of pregnant women.
- the application population of the present invention does not need to distinguish whether pregnant women are at high risk of premature delivery, and can be applied to general pregnant populations.
- the present invention can realize the prediction of unexplained spontaneous premature birth in the second trimester.
- the present invention can achieve preterm birth prediction up to 23 weeks in advance. Therefore, the method of the present invention is applicable to a wider population and has more clinical applicability.
- the expression profile of the gene markers is obtained by quantitative analysis of free extracellular RNA (cfRNA) in the biological sample; preferably, high-throughput sequencing or RT-PCR method is used Quantitative analysis of free extracellular RNA in the biological sample; more preferably, quantitative analysis of free extracellular RNA in the biological sample by next-generation sequencing.
- cfRNA free extracellular RNA
- the free extracellular RNA in the biological sample can be extracted by a method or a kit commonly used in the art or a combination of the two.
- cell-free extracellular RNA can be isolated from plasma biological samples using TRIzol LS standard RNA extraction procedures.
- the quantitative analysis of free extracellular RNA preferably includes sequencing the free extracellular RNA in biological samples (preferably plasma samples) of pregnant women using next-generation sequencing by whole transcriptome sequencing.
- This method can simultaneously sequence plasma free mRNA and free lncRNA.
- RT-PCR method can also be used for analysis.
- the expression profile of extracellular free RNA can also be quantitatively analyzed by other methods known in the art such as qPCR.
- the quantitative analysis of extracellular free RNA also includes the step of quality control of the original extracellular free RNA sequencing data, preferably including cutting adapters, removing low-quality reads, removing ⁇ 17bp reads, and removing rRNA sequence, value RNA and Y RNA sequence, and the remaining read lengths are first compared to the human transcriptome (the sequence is miRNA, tRNA and piRNA, mRNA and lncRNA, and finally other RNA).
- RNA alignment is performed using bowtie software, and quantification is performed using RSEM.
- a prediction kit for the gene markers of the present invention can be prepared according to the existing kit preparation principles. It is also possible to prepare detection probes, chips, etc. for predicting the risk of spontaneous premature birth of pregnant women with unknown reasons for these gene markers.
- the present invention uses a specific gene marker as a detection target, and based on the correlation between the expression profile of the gene marker and the unexplained spontaneous premature birth disease of pregnant women, realizes the high specificity and high sensitivity risk prediction of unexplained spontaneous premature birth in pregnant women .
- the present invention provides a kit for predicting the risk of unexplained spontaneous premature birth in pregnant women, the kit includes detection reagents for genetic markers, and the genetic markers include one or more of the following: Genes: AKAP2, CCNB1IP1, CEACAM19, EMP3, FAR1, FOXN3, GSAP, GTF3C2, HPS3, MTURN, NR1D2, PIK3CG, TMUB1, UPF1, WDR34, ZFR, AC005332.6, AC016727.1, AC018716.2, AC021087.2, AC022613.3, AC084759.3, AC092338.2, AC093525.9, AC099689.1, AC105020.6, AL109936.2, AL138921.1, AL606760.3, AP000688.4, FP671120.4, LINC00221, LINC0 0511, LINC00689, LINC02076, TTLL10-AS1.
- the genetic markers include one or more of the following
- the kit is used for prediction, which makes the prediction more convenient, simple and fast.
- the above gene markers include one or more of the following genes: FP671120.4, TTLL10-AS1, AL109936.2, LINC02076, AC021087.2, AL606760.3, AC018716.2, LINC00221, LINC00511, AC099689.1, AC005332. 6. AL138921.1, AC093525.9, LINC00689, AP000688.4, AC022613.3, AC105020.6, AC084759.3, AC016727.1, AC092338.2.
- the detection reagents for gene markers may include probes and/or primers for detecting gene markers, specifically one or more probes and/or primers that specifically bind (hybridize) to gene markers One or more primers that specifically amplify a genetic marker.
- the kit of the present invention may also include reagents for converting RNA in a biological sample into a library of cDNA fragments.
- the application of detection reagents for gene markers in the preparation of kits for predicting the risk of unexplained spontaneous premature birth in pregnant women is provided.
- the gene markers include one or more of the following genes: AKAP2, CCNB1IP1, CEACAM19, EMP3, FAR1, FOXN3, GSAP, GTF3C2, HPS3, MTURN, NR1D2, PIK3CG, TMUB1, UPF1, WDR34, ZFR, AC005332.6, AC016727.1, AC018716.2, AC021087.2, AC022613.3, AC08475 9. 3.
- the gene markers include one or more of the following genes: FP671120.4, TTLL10-AS1, AL109936.2, LINC02076, AC021087.2, AL606760.3, AC018716.2, LINC00221, LINC00511, AC099689.1, AC005332 .6, AL138921.1, AC093525.9, LINC00689, AP000688.4, AC022613.3, AC105020.6, AC084759.3, AC016727.1, AC092338.2.
- the detection reagents of gene markers include probes and/or primers for detecting gene markers, specifically one or more probes that specifically bind (hybridize) to gene markers and/or one or more Primers that specifically amplify gene markers.
- the present invention provides a device for predicting the risk of pregnant women with unexplained spontaneous preterm birth.
- the device has a built-in risk prediction model for pregnant women with unexplained spontaneous Computer-trained expression profiles of gene markers in biological samples from pregnant women with unexplained spontaneous preterm birth, including one or more of the following genes: AKAP2, CCNB1IP1, CEACAM19, EMP3, FAR1, FOXN3, GSAP, GTF3C2, HPS3, MTURN, NR1D2, PIK3CG, TMUB1, UPF1, WDR34, ZFR, AC005332.6, AC016727.1, AC018716.2, AC021087.2, AC022613.3, AC084759.3, AC092338.2, AC093525.9 , AC099689.1, AC105020.6, AL109936.2, AL138921.1, AL606760.3, AP000688.4, FP671120.4, LINC00221, LINC0051
- the gene markers include one or more of the following genes: FP671120.4, TTLL10-AS1, AL109936.2, LINC02076, AC021087.2, AL606760.3, AC018716.2, LINC00221, LINC00511, AC099689.1, AC005332 .6, AL138921.1, AC093525.9, LINC00689, AP000688.4, AC022613.3, AC105020.6, AC084759.3, AC016727.1, AC092338.2.
- the prediction model is a generalized linear model, a gradient boosting machine, a random forest or a support vector machine model.
- a method for constructing a risk prediction model for pregnant women with unexplained spontaneous preterm birth includes: detecting biological Differential expression of gene markers in samples; some pregnant women with unexplained spontaneous premature birth and some pregnant women with full-term delivery are used as training sets, and the best gene markers are screened out using the training set; in the training set, the best gene markers are used
- the marker trains the computer to obtain a risk prediction model for pregnant women with unexplained spontaneous preterm birth; the remaining part of the group of pregnant women with unexplained spontaneous preterm birth and the remaining part of the group of pregnant women with full-term delivery are used as the verification set, and the verification set is used to verify the unexplained spontaneous preterm birth of pregnant women Risk prediction model; wherein the optimal genetic markers include one or more of the following genes: AKAP2, CCNB1IP1, CEACAM19, EMP3, FAR1, FOXN3, GSAP, GTF3C2, HPS3, MTURN,
- the above optimal gene markers include one or more of the following genes: FP671120.4, TTLL10-AS1, AL109936.2, LINC02076, AC021087.2, AL606760.3, AC018716.2, LINC00221, LINC00511, AC099689.1, AC005332.6, AL138921.1, AC093525.9, LINC00689, AP000688.4, AC022613.3, AC105020.6, AC084759.3, AC016727.1, AC092338.2.
- the biological sample used in the model construction method of the present invention is preferably one or more of the following: plasma, serum, whole blood, urine, amniotic fluid; particularly preferably plasma, serum, whole blood; most preferably plasma. Also, biological samples can be collected from the 11th to 25th week of pregnancy.
- a machine learning method can be used, preferably the machine learning method includes one or more of the following: generalized linear model, gradient boosting machine, random forest and support vector machine.
- the training set and the verification set can be split according to a certain ratio according to the needs.
- all pregnant women with unexplained spontaneous premature delivery are randomly split into the training set and the verification set according to the ratio of 7:3.
- Set, all pregnant women who gave birth at full term were randomly split into a training set and a validation set according to the ratio of 7:3.
- the screening of the best gene markers is done in the training set, and the validation set is used to test the prediction effect of the best gene markers and models.
- candidate gene markers are preliminarily screened by comparing the difference in gene expression profile between a group of pregnant women with unexplained spontaneous premature labor and a group of pregnant women who gave birth at term.
- the gene markers may include mRNA genes and lncRNA genes.
- This step can be performed, for example, using the DESeq2 package (R package).
- R package the DESeq2 package
- the difference and stability of the average expression level in the two populations will be considered in this step (preferably the average expression level difference is greater than or equal to 2, and the corrected p value is less than 0.2), and finally the genes that pass the screening become candidates genetic markers.
- generalized linear models and random forests can be used to screen according to the importance of features. For example, 30 most important molecules can be screened out of each screen, and the screening process is performed 20 times, and the gene markers with higher frequency are selected as the best Gene markers.
- each algorithm adopts a 7-fold cross-validation method to select the optimal parameters for prediction model construction.
- the resulting model can be validated against the validation set.
- the model with the best effect can be selected and the feature importance can be calculated through the effect verification of the verification set.
- the mRNA gene and the lncRNA gene can be used together as gene markers for effect verification, so as to construct a risk prediction model.
- the prediction model constructed by the method of the present invention can be used in the second trimester and up to 23 weeks in advance, and only need to collect peripheral blood from pregnant women to use a non-invasive method to predict the risk of unexplained spontaneous premature birth, and the prediction is sensitive
- the accuracy can reach 74%
- the specificity can reach 90%
- the area under the receiver operating characteristic curve (AUC) is 0.96 in the training set, and 0.91 in the verification set, both of which are higher than the state of the art.
- the present application can be realized by means of software plus necessary detection instruments and other hardware devices.
- the data processing part in the technical solution of the present application can be embodied in the form of software products, and the computer software products can be stored in storage media, such as ROM/RAM, magnetic disks, optical disks, etc., including several instructions.
- a computer device which may be a personal computer, a server, or a network device, etc. executes the methods of various embodiments or some parts of the embodiments of the present application.
- the application can be used in numerous general purpose or special purpose computing system environments or configurations. Examples: personal computers, server computers, handheld or portable devices, tablet-type devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, including A distributed computing environment for any of the above systems or devices, etc.
- modules or steps of the above-mentioned application can be implemented on general-purpose computing devices, and they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices , alternatively, they can be implemented with executable program codes of the computing device, thus, they can be stored in the storage device and executed by the computing device, or they can be made into individual integrated circuit modules respectively, or the Multiple modules or steps are implemented as a single integrated circuit module.
- the present application is not limited to any specific combination of hardware and software.
- a storage medium is provided, and the storage medium includes a stored program, wherein, when the program is running, the device where the storage medium is located is controlled to execute the above-mentioned method for predicting the risk of premature rupture of membranes in pregnant women or The method for constructing the risk prediction model for premature rupture of membranes in pregnant women is implemented above.
- a processor is provided, and the processor is used to run a program, wherein, when the program is running, the above-mentioned method for predicting the risk of premature rupture of membranes in pregnant women or the above-mentioned method for predicting the risk of premature rupture of membranes in pregnant women is executed.
- the construction method of risk prediction model is provided, and the processor is used to run a program, wherein, when the program is running, the above-mentioned method for predicting the risk of premature rupture of membranes in pregnant women or the above-mentioned method for predicting the risk of premature rupture of membranes in pregnant women is executed.
- a storage medium includes a stored program, wherein, when the program is running, the device where the storage medium is located is controlled to execute the above-mentioned method for predicting the risk of unexplained spontaneous premature birth in a pregnant woman or execute A method for constructing a risk prediction model for unexplained spontaneous preterm birth in pregnant women mentioned above.
- a processor is provided, and the processor is used to run a program, wherein, when the program is running, the above-mentioned method for predicting the risk of pregnant women with unexplained spontaneous premature birth or the above-mentioned prediction of the risk of pregnant women with unexplained spontaneous premature birth is performed. How the model was built.
- gene markers of the present invention may be effective in predicting the gestational age of pregnant women.
- the peripheral blood of 277 cases of singleton pregnant women was obtained from the hospital, and the blood collection was from 11 to 25 gestational weeks, as shown in Figure 1.
- Blood came from premature and full-term pregnant women, including 104 cases of premature rupture of membranes, 74 cases of unexplained spontaneous premature birth, and 99 cases of full-term pregnant women.
- the gestational weeks of premature pregnant women from blood collection to delivery range from 6 to 23 weeks, as shown in Figure 2. All blood samples were immediately stored at 4°C and plasma separation was performed within 8 hours. Plasma was separated by a two-step centrifugation method, centrifuged at 1,600g for 10 minutes at 4°C, and then centrifuged at 12,000g for 10 minutes. Immediately after separation, the plasma was stored at -80°C pending further processing.
- Trizol LS Add Trizol LS to the plasma and vortex immediately to mix.
- the subsequent cfRNA extraction steps are performed using the standard RNA extraction method of TRIzol LS.
- Sequencing of cfRNA utilized whole-transcriptome sequencing of plasma samples from preterm (premature rupture of membranes preterm and unexplained spontaneous preterm birth, respectively) and term pregnant women using next-generation sequencing. This method can simultaneously sequence plasma free mRNA and free lncRNA.
- RNA and Y RNA sequences Quality control was performed on the original cfRNA sequencing data, including cutting adapters, removing low-quality reads, removing reads ⁇ 17bp in length, removing rRNA sequences, value RNA and Y RNA sequences. Align the remaining reads to the human transcriptome (in the order of miRNA, tRNA and piRNA, mRNA and lncRNA, and finally other RNAs), and then align the remaining reads to the human genome.
- the expression level of long RNA is corrected to TPM, the formula is as follows:
- TPM (Ni/Li)*1000000/(sum(N1/L1+N2/L2+N3/L3+...+Nn/Ln))
- Ni is the number of reads aligned to the i-th gene; Li is the length of the i-th gene; sum(N1/L1+N2/L2+...+Nn/Ln) is the length of all (n) genes The sum of values after normalization.
- TotalMappingReads is the sum of the read lengths on all alignments.
- the group of pregnant women with premature rupture of membranes, the group of pregnant women with unexplained spontaneous premature delivery and the group of pregnant women with full-term delivery were randomly divided into a training set and a verification set according to the ratio of 7:3.
- the training set contained 72 samples of premature rupture of membranes and premature delivery.
- the validation set contains 32 premature rupture of membranes preterm samples, 23 unexplained spontaneous preterm samples and 30 full-term samples.
- the screening of gene markers is completed in the training set, and the verification set is used to test the prediction effect of gene markers and models. Please refer to Table 1 for the relevant data of the pregnant women group.
- Table 1 Relevant data of the group of pregnant women with premature delivery and the group of pregnant women with full-term delivery in Example 1
- Candidate gene markers were preliminarily screened by comparing the expression profile differences among pregnant women with premature rupture of membranes, unexplained spontaneous premature labor, and full-term labor. This step was implemented using the DESeq2 package (R software package). For each gene, the difference and stability of the average expression level between the two groups are considered in this step (the average expression level difference is greater than or equal to 2, and the corrected p value is less than 0.2), and finally the genes that pass the screening become candidate gene markers things. Screening based on feature importance was performed using generalized linear models and random forests, from which the 30 most important molecules were selected for each screening. This process was performed 20 times, and the gene marker with higher frequency was selected as the best gene marker. The flow chart of the screening of gene markers is shown in Figure 3.
- 21 mRNA gene markers (CCNB1IP1, COL9A2, DNAJC13, FAM45A, FBXO38, FZD3, HCK, KIAA1257, KIF2A, LARP1B, LRRC56, PLD5, PROS1, SEPT10, SLC41A3, SPIN1, TMUB1, TUBB4A, UPF1, WDR34, ZBTB10) and 18 lncRNA gene markers (AC004803.1, AC009779.2, AC011461.1, AC015878.2, AC016727.1, AC022568.1, AC084759.3, AC092338.2, AC093249.2, AC103876.1, AC105020.6, AC108099.1, AL031733.2, AL451074.2, AP000688.4, NORAD, PINK1-AS, REV3L-IT1), as the best genetic markers for premature rupture of membranes 16 mRNA gene markers (AKAP2, CCNB1IP1, CE
- Table 2 Gene and transcript information of the best genetic markers for premature rupture of membranes and preterm birth obtained through screening in Example 1
- Table 3 Gene and transcript information of the best gene markers for unexplained spontaneous premature birth screened in Example 1
- the above-mentioned embodiments of the present invention have achieved the following technical effects: using the combination of multiple gene markers of the present invention in plasma, combined with the machine learning model, can predict premature rupture of membranes and premature labor up to 23 weeks earlier. Unexplained spontaneous premature birth.
- the present invention can predict the risk of premature birth in a non-invasive way only by taking peripheral blood from pregnant women.
- the gene markers of the present invention can be used alone or in combination. When used alone, the predictive sensitivity and specificity of the premature rupture of membranes and preterm gene markers of the present invention can reach at least 44% and 57% respectively, and the predictive sensitivity and specificity of the unexplained preterm gene markers can respectively reach 44% and 57%.
- the gene markers of the present invention can achieve a prediction sensitivity of more than 63% and a prediction specificity of more than 83% for premature rupture of membranes, and a prediction of more than 74% for unexplained premature birth
- the sensitivity and the prediction specificity of more than 80% are both higher than the state of the art.
- the sensitivity of premature rupture of membranes and preterm birth prediction can reach 75%, the specificity can reach 83%, and the area under the receiver operating characteristic curve (AUC) reaches 0.94 in the training set and 0.82 in the validation set , are higher than the existing technical level; the sensitivity of unexplained spontaneous preterm birth prediction can reach 74%, the specificity can reach 90%, the area under the receiver operating characteristic curve reaches 0.96 in the training set, and 0.91 in the verification set, which is much higher at the current level of technology.
- the method of the present invention is applicable to asymptomatic general pregnant women, regardless of high-risk or not, it can be predicted in the second trimester, and premature birth can be predicted up to 23 weeks earlier, which is 15 weeks earlier than the prior art.
- the method of the invention is applicable to a wider population and has more clinical applicability.
- the prediction model of the present invention has relatively high accuracy, and is suitable for early prediction of the premature birth risk of pregnant women, so as to achieve early intervention.
Landscapes
- Chemical & Material Sciences (AREA)
- Organic Chemistry (AREA)
- Life Sciences & Earth Sciences (AREA)
- Zoology (AREA)
- Wood Science & Technology (AREA)
- Proteomics, Peptides & Aminoacids (AREA)
- Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Microbiology (AREA)
- Immunology (AREA)
- Physics & Mathematics (AREA)
- Molecular Biology (AREA)
- Biotechnology (AREA)
- Biophysics (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Genetics & Genomics (AREA)
- Measuring Or Testing Involving Enzymes Or Micro-Organisms (AREA)
Abstract
本发明提供了基因标志物在预测孕妇早产风险中的应用。本发明提供了一种用于预测孕妇胎膜早破早产及不明原因自发早产风险的方法,包括:获取来源于所述孕妇的生物样品中基因标志物的表达谱;基于基因标志物的表达谱,鉴别孕妇的胎膜早破早产风险及不明原因自发早产风险。本发明还提供了用于预测孕妇胎膜早破早产及不明原因自发早产风险的试剂盒和装置、以及孕妇早产风险预测模型的构建方法,还提供了涉及用于执行早产风险预测方法和模型构建方法的程序的存储介质和处理器。本发明通过基因标志物的表达谱与胎膜早破早产风险及不明原因自发早产的关联性,实现了对胎膜早破早产风险及不明原因自发早产的高特异性和高灵敏性的风险预测。
Description
本发明涉及孕妇早产领域,具体而言,涉及基因标志物在预测胎膜早破早产及不明原因自发早产风险中的应用。
早产是指妊娠不足37周的生产。在全球,早产是五岁以下儿童的主要死亡原因,在几乎所有具有可靠数据的国家,早产率都在日益增加。早产是母婴领域的重要问题。
胎膜早破及不明原因可导致早产。胎膜早破是指在临产前胎膜自然破裂,孕龄小于37周的胎膜早破称为早产胎膜早破。预防早产带来的死亡和并发症要从健康妊娠做起。早预测早干预可改善妊娠结局。
目前,临床上有针对早产高风险人群进行宫颈长度检测以及阴道分泌物的fFN胎儿纤维连接蛋白检测用于评估早产风险,但主要针对高危人群,且灵敏度、特异性有限。一些研究和专利申请涉及利用基因表达、代谢物、蛋白/多肽、微生物进行早产预测和诊断,但是主要问题仍然在于这些方法对于早产风险预测的灵敏度和特异性较低。
到目前为止,还没有任何一种可以对胎膜早破早产或不明原因自发早产进行高特异性和灵敏性预测的基因标志物。所以,迫切需要开发一种可以高特异性和高灵敏性地对胎膜早破早产或不明原因自发早产进行预测的基因标志物。
发明内容
本发明的主要目的在于提供基因标志物在预测胎膜早破早产风险或不明原因自发早产风险中的应用,以提供一种对早产风险的高特异性和高灵敏性的预测方案。
为了实现上述目的,根据本发明的第一方面,提供了一种用于预测孕妇胎膜早破早产风险的方法,该方法包括:
步骤S1:获取来源于孕妇的生物样品中基因标志物的表达谱,基因标志物包括以下一种或多种基因:CCNB1IP1、COL9A2、DNAJC13、FAM45A、FBXO38、FZD3、HCK、KIAA1257、KIF2A、LARP1B、LRRC56、PLD5、PROS1、SEPT10、SLC41A3、SPIN1、TMUB1、TUBB4A、UPF1、WDR34、ZBTB10、AC004803.1、AC009779.2、AC011461.1、AC015878.2、AC016727.1、AC022568.1、AC084759.3、AC092338.2、AC093249.2、AC103876.1、AC105020.6、AC108099.1、AL031733.2、AL451074.2、AP000688.4、NORAD、PINK1-AS、REV3L-IT1;
步骤S2:基于基因标志物的表达谱,鉴别孕妇的胎膜早破早产风险。
进一步地,在步骤S2中,鉴别孕妇的胎膜早破早产风险是通过利用孕妇胎膜早破早产风险预测模型来实施的,孕妇胎膜早破早产风险预测模型是通过利用来源于已发生胎膜早破早产的孕妇的生物样品中基因标志物的表达谱训练计算机而产生。
进一步地,训练计算机是通过机器学习方法来实施,优选机器学习方法包括以下一种或多种:广义线性模型、梯度提升机、随机森林、支持向量机。
进一步地,生物样品为以下一种或多种:血浆、血清、全血、尿液、羊水;优选生物样品在孕妇第11至25孕周时采集获得。
进一步地,在步骤S1中,通过对生物样品中的胞外游离RNA进行定量分析,从而获取基因标志物的表达谱;
优选地,采用高通量测序法或RT-PCR法对生物样品中的胞外游离RNA进行定量分析;
更优选地,采用高通量测序法对生物样品中的胞外游离RNA进行定量分析。
根据本发明的第二方面,提供了一种用于预测孕妇胎膜早破早产风险的试剂盒,试剂盒包括基因标志物的检测试剂,基因标志物包括以下一种或多种基因:CCNB1IP1、COL9A2、DNAJC13、FAM45A、FBXO38、FZD3、HCK、KIAA1257、KIF2A、LARP1B、LRRC56、PLD5、PROS1、SEPT10、SLC41A3、SPIN1、TMUB1、TUBB4A、UPF1、WDR34、ZBTB10、AC004803.1、AC009779.2、AC011461.1、AC015878.2、AC016727.1、AC022568.1、AC084759.3、AC092338.2、AC093249.2、AC103876.1、AC105020.6、AC108099.1、AL031733.2、AL451074.2、AP000688.4、NORAD、PINK1-AS、REV3L-IT1。
进一步地,基因标志物的检测试剂包括用于检测基因标志物的探针和/或引物;优选为将基因标志物的RNA制备成高通量测序文库的相关试剂。
根据本发明的第三方面,提供了基因标志物的检测试剂在制备预测孕妇胎膜早破早产风险的试剂盒中的应用,基因标志物包括以下一种或多种基因:CCNB1IP1、COL9A2、DNAJC13、FAM45A、FBXO38、FZD3、HCK、KIAA1257、KIF2A、LARP1B、LRRC56、PLD5、PROS1、SEPT10、SLC41A3、SPIN1、TMUB1、TUBB4A、UPF1、WDR34、ZBTB10、AC004803.1、AC009779.2、AC011461.1、AC015878.2、AC016727.1、AC022568.1、AC084759.3、AC092338.2、AC093249.2、AC103876.1、AC105020.6、AC108099.1、AL031733.2、AL451074.2、AP000688.4、NORAD、PINK1-AS、REV3L-IT1。
进一步地,基因标志物的检测试剂包括用于检测基因标志物的探针和/或引物;优选为将基因标志物的RNA制备成高通量测序文库的相关试剂。
根据本发明的第四方面,提供了一种用于预测孕妇胎膜早破早产风险的装置,装置内置有孕妇胎膜早破早产风险预测模型,预测模型是通过利用来源于已发生胎膜早破早产的孕妇的生物样品中基因标志物的表达谱训练计算机而产生,基因标志物包括以下一种或多种基因:CCNB1IP1、COL9A2、DNAJC13、FAM45A、FBXO38、FZD3、HCK、KIAA1257、KIF2A、 LARP1B、LRRC56、PLD5、PROS1、SEPT10、SLC41A3、SPIN1、TMUB1、TUBB4A、UPF1、WDR34、ZBTB10、AC004803.1、AC009779.2、AC011461.1、AC015878.2、AC016727.1、AC022568.1、AC084759.3、AC092338.2、AC093249.2、AC103876.1、AC105020.6、AC108099.1、AL031733.2、AL451074.2、AP000688.4、NORAD、PINK1-AS、REV3L-IT1。
根据本发明的第五方面,提供了一种孕妇胎膜早破早产风险预测模型的构建方法,构建方法包括:
检测来源于胎膜早破早产的孕妇群体和足月分娩的孕妇群体的生物样品中的基因标志物的差异表达;
将部分胎膜早破早产的孕妇群体和部分足月分娩的孕妇群体作为训练集,利用训练集筛选出最佳基因标志物;
在训练集中,利用最佳基因标志物训练计算机,从而得到孕妇胎膜早破早产风险预测模型;
将剩余部分的胎膜早破早产的孕妇群体和剩余部分的足月分娩的孕妇群体作为验证集,利用验证集验证孕妇胎膜早破早产风险预测模型;
其中,最佳基因标志物包括以下一种或多种基因:CCNB1IP1、COL9A2、DNAJC13、FAM45A、FBXO38、FZD3、HCK、KIAA1257、KIF2A、LARP1B、LRRC56、PLD5、PROS1、SEPT10、SLC41A3、SPIN1、TMUB1、TUBB4A、UPF1、WDR34、ZBTB10、AC004803.1、AC009779.2、AC011461.1、AC015878.2、AC016727.1、AC022568.1、AC084759.3、AC092338.2、AC093249.2、AC103876.1、AC105020.6、AC108099.1、AL031733.2、AL451074.2、AP000688.4、NORAD、PINK1-AS、REV3L-IT1。
进一步地,生物样品为以下一种或多种:血浆、血清、全血、尿液、羊水;优选生物样品在孕妇第11至25孕周时采集获得。
进一步地,训练计算机是通过机器学习方法来实施,优选机器学习方法包括以下一种或多种:广义线性模型、梯度提升机、随机森林、支持向量机。
根据本发明的第六方面,提供了一种计算机可读存储介质,存储介质包括存储的程序,其中,在程序运行时控制存储介质所在设备执行本发明第一方面的用于预测孕妇胎膜早破早产风险的方法或本发明第五方面的孕妇胎膜早破早产风险预测模型的构建方法。
根据本发明的第七方面,提供了一种处理器,处理器用于运行程序,其中,程序运行时执行本发明第一方面的用于预测孕妇胎膜早破早产风险的方法或本发明第五方面的孕妇胎膜早破早产风险预测模型的构建方法。
根据本发明的第八方面,提供了一种用于预测孕妇不明原因自发早产风险的方法,方法包括:
步骤S1:获取来源于孕妇的生物样品中基因标志物的表达谱,基因标志物包括以下一种或多种基因:AKAP2、CCNB1IP1、CEACAM19、EMP3、FAR1、FOXN3、GSAP、GTF3C2、HPS3、MTURN、NR1D2、PIK3CG、TMUB1、UPF1、WDR34、ZFR、AC005332.6、AC016727.1、AC018716.2、AC021087.2、AC022613.3、AC084759.3、AC092338.2、AC093525.9、AC099689.1、AC105020.6、AL109936.2、AL138921.1、AL606760.3、AP000688.4、FP671120.4、LINC00221、LINC00511、LINC00689、LINC02076、TTLL10-AS1;
步骤S2:基于基因标志物的表达谱,鉴别孕妇的不明原因自发早产风险。
进一步地,在步骤S2中,鉴别孕妇的不明原因自发早产风险是通过利用孕妇不明原因自发早产风险预测模型来实施的,孕妇不明原因自发早产风险预测模型是通过利用来源于已发生不明原因自发早产的孕妇的生物样品中基因标志物的表达谱训练计算机而产生。
进一步地,训练计算机是通过机器学习方法来实施,优选机器学习方法包括以下一种或多种:广义线性模型、梯度提升机、随机森林、支持向量机。
进一步地,生物样品为以下一种或多种:血浆、血清、全血、尿液、羊水;优选生物样品在孕妇第11至25孕周时采集获得。
进一步地,在步骤S1中,通过对生物样品中的胞外游离RNA进行定量分析,从而获取基因标志物的表达谱;
优选地,采用高通量测序法或RT-PCR法对生物样品中的胞外游离RNA进行定量分析;
更优选地,采用高通量测序法对生物样品中的胞外游离RNA进行定量分析。
根据本发明的第九方面,提供了一种用于预测孕妇不明原因自发早产风险的试剂盒,试剂盒包括基因标志物的检测试剂,基因标志物包括以下一种或多种基因:AKAP2、CCNB1IP1、CEACAM19、EMP3、FAR1、FOXN3、GSAP、GTF3C2、HPS3、MTURN、NR1D2、PIK3CG、TMUB1、UPF1、WDR34、ZFR、AC005332.6、AC016727.1、AC018716.2、AC021087.2、AC022613.3、AC084759.3、AC092338.2、AC093525.9、AC099689.1、AC105020.6、AL109936.2、AL138921.1、AL606760.3、AP000688.4、FP671120.4、LINC00221、LINC00511、LINC00689、LINC02076、TTLL10-AS1。
进一步地,基因标志物的检测试剂包括用于检测基因标志物的探针和/或引物;优选为将基因标志物的RNA制备成高通量测序文库的相关试剂。
根据本发明的第十方面,提供了基因标志物的检测试剂在制备预测孕妇不明原因自发早产风险的试剂盒中的应用,基因标志物包括以下一种或多种基因:AKAP2、CCNB1IP1、CEACAM19、EMP3、FAR1、FOXN3、GSAP、GTF3C2、HPS3、MTURN、NR1D2、PIK3CG、TMUB1、UPF1、WDR34、ZFR、AC005332.6、AC016727.1、AC018716.2、AC021087.2、AC022613.3、AC084759.3、AC092338.2、AC093525.9、AC099689.1、AC105020.6、AL109936.2、 AL138921.1、AL606760.3、AP000688.4、FP671120.4、LINC00221、LINC00511、LINC00689、LINC02076、TTLL10-AS1。
进一步地,基因标志物的检测试剂包括用于检测基因标志物的探针和/或引物;优选为将基因标志物的RNA制备成高通量测序文库的相关试剂。
根据本发明的第十一方面,提供了一种用于预测孕妇不明原因自发早产风险的装置,装置内置有孕妇不明原因自发早产风险预测模型,预测模型是通过利用来源于已发生不明原因自发早产的孕妇的生物样品中基因标志物的表达谱训练计算机而产生,基因标志物包括以下一种或多种基因:AKAP2、CCNB1IP1、CEACAM19、EMP3、FAR1、FOXN3、GSAP、GTF3C2、HPS3、MTURN、NR1D2、PIK3CG、TMUB1、UPF1、WDR34、ZFR、AC005332.6、AC016727.1、AC018716.2、AC021087.2、AC022613.3、AC084759.3、AC092338.2、AC093525.9、AC099689.1、AC105020.6、AL109936.2、AL138921.1、AL606760.3、AP000688.4、FP671120.4、LINC00221、LINC00511、LINC00689、LINC02076、TTLL10-AS1。
根据本发明的第十二方面,提供了一种孕妇不明原因自发早产风险预测模型的构建方法,构建方法包括:
检测来源于不明原因自发早产的孕妇群体和足月的孕妇群体的生物样品中的基因标志物的差异表达;
将部分不明原因自发早产的孕妇群体和部分足月的孕妇群体作为训练集,利用训练集筛选出最佳基因标志物;
在训练集中,利用最佳基因标志物训练计算机,从而得到孕妇不明原因自发早产风险预测模型;
将剩余部分的不明原因自发早产的孕妇群体和剩余部分的足月的孕妇群体作为验证集,利用验证集验证孕妇不明原因自发早产风险预测模型;
其中,最佳基因标志物包括以下一种或多种基因:AKAP2、CCNB1IP1、CEACAM19、EMP3、FAR1、FOXN3、GSAP、GTF3C2、HPS3、MTURN、NR1D2、PIK3CG、TMUB1、UPF1、WDR34、ZFR、AC005332.6、AC016727.1、AC018716.2、AC021087.2、AC022613.3、AC084759.3、AC092338.2、AC093525.9、AC099689.1、AC105020.6、AL109936.2、AL138921.1、AL606760.3、AP000688.4、FP671120.4、LINC00221、LINC00511、LINC00689、LINC02076、TTLL10-AS1。
进一步地,生物样品为以下一种或多种:血浆、血清、全血、尿液、羊水;优选生物样品在孕妇第11至25孕周时采集获得。
进一步地,训练计算机是通过机器学习方法来实施,优选机器学习方法包括以下一种或多种:广义线性模型、梯度提升机、随机森林、支持向量机。
根据本发明的第十三方面,提供了一种计算机可读存储介质,存储介质包括存储的程序,其中,在程序运行时控制存储介质所在设备执行本发明第八方面的用于预测孕妇不明原因自发早产风险的方法或本发明第十二方面的孕妇不明原因自发早产风险预测模型的构建方法。
根据本发明的第十四方面,提供了一种处理器,其特征在于,处理器用于运行程序,其中,程序运行时执行本发明第八方面的用于预测孕妇不明原因自发早产风险的方法或本发明第十二方面的孕妇不明原因自发早产风险预测模型的构建方法。
本发明针对现有技术中早产风险的预测准确性较低的问题,提出了采用本申请的基因标志物作为检测靶标,通过基因标志物的表达谱与胎膜早破早产风险及不明原因自发早产的关联性,实现了对胎膜早破早产风险及不明原因自发早产的高特异性和高灵敏性的风险预测。
构成本申请的一部分的说明书附图用来提供对本发明的进一步理解,本发明的示意性实施例及其说明用于解释本发明,并不构成对本发明的不当限定。在附图中:
图1示出了根据本发明的优选实施例中孕妇群体生物样品采集孕周的柱形图;
图2示出了根据本发明的优选实施例中孕妇群体分娩与生物样品采集间隔孕周的柱形图;
图3示出了根据本发明的优选实施例中基因标志物的筛选流程图;
图4示出了根据本发明的优选实施例中早产风险预测模型的构建流程图;
图5示出了根据本发明的优选实施例中预测胎膜早破早产的最佳基因标志物的重要性排序图以及模型预测的AUC曲线图;
图6示出了根据本发明的优选实施例中预测不明原因自发早产的最佳基因标志物的重要性排序图以及模型预测的AUC曲线图。
需要说明的是,在不冲突的情况下,本申请中的实施例及实施例中的特征可以相互组合。下面将结合实施例来详细说明本发明。
如背景技术部分所提到的,目前存在着对孕妇早产进行临床早期预测的需求。本申请基于来源于孕妇的生物样品,通过比较早产组及足月组在孕早、中期的基因表达量差异,结合机器学习算法,筛选出预测早产风险的基因标志物,并通过构建模型实现了在孕中期对早产的高准确度预测。本发明的基因标志物和预测模型对于早产风险、特别是胎膜早破早产和不明原因自发早产的预测具有较高的特异性和灵敏性,可在孕中期以较高准确度发现孕妇的早产风险,实现尽早干预。
在该研究结果的基础上,申请人提出了本申请的技术方案。在一种典型的实施方式中,提供了一种用于预测孕妇胎膜早破早产风险的方法,该方法包括:
步骤S1:获取来源于所述孕妇的生物样品中基因标志物的表达谱,基因标志物包括以下一种或多种基因:CCNB1IP1、COL9A2、DNAJC13、FAM45A、FBXO38、FZD3、HCK、KIAA1257、KIF2A、LARP1B、LRRC56、PLD5、PROS1、SEPT10、SLC41A3、SPIN1、TMUB1、TUBB4A、UPF1、WDR34、ZBTB10、AC004803.1、AC009779.2、AC011461.1、AC015878.2、AC016727.1、AC022568.1、AC084759.3、AC092338.2、AC093249.2、AC103876.1、AC105020.6、AC108099.1、AL031733.2、AL451074.2、AP000688.4、NORAD、PINK1-AS、REV3L-IT1;
步骤S2:基于基因标志物的表达谱,鉴别孕妇的胎膜早破早产风险。
本申请首次发现孕妇生物样品中的基因标志物与孕妇胎膜早破早产疾病有着显著的相关性,因而可以作为预测孕妇胎膜早破早产的标志物。这些基因标志物包括21个mRNA基因和18个lncRNA基因,其中mRNA基因标志物包括CCNB1IP1、COL9A2、DNAJC13、FAM45A、FBXO38、FZD3、HCK、KIAA1257、KIF2A、LARP1B、LRRC56、PLD5、PROS1、SEPT10、SLC41A3、SPIN1、TMUB1、TUBB4A、UPF1、WDR34、ZBTB10;lncRNA基因标志物包括AC004803.1、AC009779.2、AC011461.1、AC015878.2、AC016727.1、AC022568.1、AC084759.3、AC092338.2、AC093249.2、AC103876.1、AC105020.6、AC108099.1、AL031733.2、AL451074.2、AP000688.4、NORAD、PINK1-AS、REV3L-IT1。
在本发明的方法中,基因标志物优选包括DNAJC13、CCNB1IP1、AC022568.1、PLD5、WDR34、UPF1、KIF2A、SEPT10、FAM45A、ZBTB10、PROS1、COL9A2、LARP1B、AC009779.2、TMUB1、HCK、AC015878.2、AC084759.3、AP000688.4、AC092338.2中的一种或多种。
上面列出的各基因可单独或组合使用。例如,可以采用以下全部基因的组合作为基因标志物:DNAJC13、CCNB1IP1、AC022568.1、PLD5、WDR34、UPF1、KIF2A、SEPT10、FAM45A、ZBTB10、PROS1、COL9A2、LARP1B、AC009779.2、TMUB1、HCK、AC015878.2、AC084759.3、AP000688.4、AC092338.2,从而实现胎膜早破早产的风险预测。
在上述步骤S2中,鉴别孕妇的胎膜早破早产风险可以通过利用孕妇胎膜早破早产风险预测模型来实施,通过利用来源于已发生胎膜早破早产的孕妇的生物样品中上述基因标志物的表达谱训练计算机来产生孕妇胎膜早破早产风险预测模型。
训练计算机可通过机器学习方法来实施。机器学习方法选自回归法、分类法或其组合。“机器学习”一般表示在未明确编程的情况下,给予计算机学习能力的算法,包括从数据学习并对数据做出预测的算法。本发明所使用的机器学习方法可以包括随机森林、最小绝对收缩和选择算子逻辑回归、正则化逻辑回归、XGBoost、决策树学习、人工神经网络、深度神经网络、支持向量机、基于规则的机器学习、广义线性模型、梯度提升机等。优选机器学习方法包括以下一种或多种:广义线性模型、梯度提升机、随机森林、支持向量机。
在预测模型中,可通过模型自动计算得出的风险分数,来评价和预测胎膜早破早产风险高低。例如,若风险分数大于0.5,认为胎膜早破早产高风险,若风险分数小于0.5,则认为胎膜早破早产低风险。
来源于孕妇的生物样品可以为以下一种或多种:血浆、血清、全血、尿液、羊水。优选采用来源于孕妇的血浆、血清或全血,用于本发明的检测和鉴别步骤。该生物样品最优选为血浆,例如,可以从孕妇获取外周血并实施血浆分离,从而获得待使用的血浆生物样品。除了血浆、血清或全血,还可以使用其他体液样品,如尿液、羊水等。生物样品的获取可以采用本领域常规的方法实施。
在本发明中,生物样品的采集可以在孕妇第11至25孕周时进行。通过采用上述特定的基因标志物作为预测因子,本发明的应用群体不必区分孕妇是否早产高危,可以适用于一般孕妇群体。利用上述基因标志物,本发明在孕中期可以实现胎膜早破早产的预测。本发明最高可以提早23周实现早产预测。因此,本发明的方法适用人群更广,更具有临床应用性。
在上述方法的步骤S1中,通过对生物样品中的胞外游离RNA(cfRNA)进行定量分析,从而获取所述基因标志物的表达谱;优选地,采用高通量测序法或RT-PCR法对生物样品中的胞外游离RNA进行定量分析;更优选地,采用下一代测序法对生物样品中的胞外游离RNA进行定量分析。
具体来说,生物样品中的胞外游离RNA可采用本领域常用的方法或试剂盒或两者组合提取获得。例如,可以使用TRIzol LS标准的RNA提取步骤,从血浆生物样品中提取胞外游离RNA。
在一种具体的实施方式中,对胞外游离RNA进行定量分析,优选包括利用全转录组测序,使用下一代测序法对孕妇生物样品(优选血浆样品)中的胞外游离RNA进行测序。该方法能同时对血浆游离mRNA和游离lncRNA进行测序。也可以采用RT-PCR的方法进行分析。还可以采用本领域已知的其他方法如qPCR法对胞外游离RNA的表达谱进行定量分析。
优选地,对胞外游离RNA进行定量分析,还包括将原始的胞外游离RNA测序数据进行质控的步骤,优选包括剪切接头,去除低质量读长,去除<17bp长度的读长,去除rRNA序列和value RNA及Y RNA序列,将剩余读长先比对到人源转录组(顺序为miRNA、tRNA和piRNA,mRNA和lncRNA,最后为其他RNA)。在优选的实施方式中,RNA比对用bowtie软件,定量用RSEM进行。
通过将本发明的基因标志物作为预测孕妇胎膜早破早产风险的标志物,根据现有试剂盒的制备原则,可以制备出针对本发明所述基因标志物的预测试剂盒。还可以针对这些基因标志物,制备出用于预测孕妇胎膜早破早产风险的检测探针、芯片等。
本发明通过采用特定的基因标志物作为检测靶标,基于基因标志物的表达谱与孕妇胎膜早破早产疾病的关联性,实现了对孕妇胎膜早破早产的高特异性和高灵敏性的风险预测。
在第二种典型的实施方式中,本发明提供了一种用于预测孕妇胎膜早破早产风险的试剂盒,该试剂盒包括基因标志物的检测试剂,基因标志物包括以下一种或多种基因:CCNB1IP1、COL9A2、DNAJC13、FAM45A、FBXO38、FZD3、HCK、KIAA1257、KIF2A、LARP1B、LRRC56、PLD5、PROS1、SEPT10、SLC41A3、SPIN1、TMUB1、TUBB4A、UPF1、WDR34、ZBTB10、AC004803.1、AC009779.2、AC011461.1、AC015878.2、AC016727.1、AC022568.1、AC084759.3、AC092338.2、AC093249.2、AC103876.1、AC105020.6、AC108099.1、AL031733.2、AL451074.2、AP000688.4、NORAD、PINK1-AS、REV3L-IT1。采用试剂盒进行预测,使得预测更加方便、简单、快速。优选上述基因标志物包括以下一种或多种基因:DNAJC13、CCNB1IP1、AC022568.1、PLD5、WDR34、UPF1、KIF2A、SEPT10、FAM45A、ZBTB10、PROS1、COL9A2、LARP1B、AC009779.2、TMUB1、HCK、AC015878.2、AC084759.3、AP000688.4、AC092338.2。
在试剂盒中,基因标志物的检测试剂可包括用于检测基因标志物的探针和/或引物,具体为一种或多种特异性结合(杂交)至基因标志物的探针和/或一种或多种特异性扩增基因标志物的引物。
由于RNA测序通常包括产生用于测序的cDNA分子的反转录步骤,因而在采用RNA测序时,本发明的试剂盒还可以包含将生物样品中的RNA转化为cDNA片段文库的试剂。
在第三种典型的实施方式中,提供了基因标志物的检测试剂在制备预测孕妇胎膜早破早产风险的试剂盒中的应用,基因标志物包括以下一种或多种基因:CCNB1IP1、COL9A2、DNAJC13、FAM45A、FBXO38、FZD3、HCK、KIAA1257、KIF2A、LARP1B、LRRC56、PLD5、PROS1、SEPT10、SLC41A3、SPIN1、TMUB1、TUBB4A、UPF1、WDR34、ZBTB10、AC004803.1、AC009779.2、AC011461.1、AC015878.2、AC016727.1、AC022568.1、AC084759.3、AC092338.2、AC093249.2、AC103876.1、AC105020.6、AC108099.1、AL031733.2、AL451074.2、AP000688.4、NORAD、PINK1-AS、REV3L-IT1。优选地,基因标志物包括以下一种或多种基因:DNAJC13、CCNB1IP1、AC022568.1、PLD5、WDR34、UPF1、KIF2A、SEPT10、FAM45A、ZBTB10、PROS1、COL9A2、LARP1B、AC009779.2、TMUB1、HCK、AC015878.2、AC084759.3、AP000688.4、AC092338.2。基因标志物的检测试剂包括用于检测基因标志物的探针和/或引物,具体地是一种或多种特异性结合(杂交)至基因标志物的探针和/或一种或多种特异性扩增基因标志物的引物。
在第四种典型的实施方式中,本发明提供了一种用于预测孕妇胎膜早破早产风险的装置,该装置内置有孕妇胎膜早破早产风险预测模型,该预测模型是通过利用来源于已发生胎膜早破早产的孕妇的生物样品中基因标志物的表达谱训练计算机而产生,所述基因标志物包括以下一种或多种基因:CCNB1IP1、COL9A2、DNAJC13、FAM45A、FBXO38、FZD3、HCK、KIAA1257、KIF2A、LARP1B、LRRC56、PLD5、PROS1、SEPT10、SLC41A3、SPIN1、TMUB1、TUBB4A、UPF1、WDR34、ZBTB10、AC004803.1、AC009779.2、AC011461.1、AC015878.2、AC016727.1、AC022568.1、AC084759.3、AC092338.2、AC093249.2、AC103876.1、AC105020.6、AC108099.1、AL031733.2、AL451074.2、AP000688.4、NORAD、PINK1-AS、REV3L-IT1。 优选地,基因标志物包括以下一种或多种基因:DNAJC13、CCNB1IP1、AC022568.1、PLD5、WDR34、UPF1、KIF2A、SEPT10、FAM45A、ZBTB10、PROS1、COL9A2、LARP1B、AC009779.2、TMUB1、HCK、AC015878.2、AC084759.3、AP000688.4、AC092338.2。在一种优选的实施方式中,该预测模型为广义线性模型、梯度提升机、随机森林或支持向量机模型。
在第五种典型的实施方式中,提供了一种孕妇胎膜早破早产风险预测模型的构建方法,该构建方法包括:检测来源于胎膜早破早产的孕妇群体和足月分娩的孕妇群体的生物样品中的基因标志物的差异表达;将部分胎膜早破早产的孕妇群体和部分足月分娩的孕妇群体作为训练集,利用训练集筛选出最佳基因标志物;在训练集中,利用最佳基因标志物训练计算机,从而得到孕妇胎膜早破早产风险预测模型;将剩余部分的胎膜早破早产的孕妇群体和剩余部分的足月分娩的孕妇群体作为验证集,利用验证集验证孕妇胎膜早破早产风险预测模型;其中,最佳基因标志物包括以下一种或多种基因:CCNB1IP1、COL9A2、DNAJC13、FAM45A、FBXO38、FZD3、HCK、KIAA1257、KIF2A、LARP1B、LRRC56、PLD5、PROS1、SEPT10、SLC41A3、SPIN1、TMUB1、TUBB4A、UPF1、WDR34、ZBTB10、AC004803.1、AC009779.2、AC011461.1、AC015878.2、AC016727.1、AC022568.1、AC084759.3、AC092338.2、AC093249.2、AC103876.1、AC105020.6、AC108099.1、AL031733.2、AL451074.2、AP000688.4、NORAD、PINK1-AS、REV3L-IT1。优选上述最佳基因标志物包括以下一种或多种基因:DNAJC13、CCNB1IP1、AC022568.1、PLD5、WDR34、UPF1、KIF2A、SEPT10、FAM45A、ZBTB10、PROS1、COL9A2、LARP1B、AC009779.2、TMUB1、HCK、AC015878.2、AC084759.3、AP000688.4、AC092338.2。
本发明的模型构建方法所采用的生物样品优选为以下一种或多种:血浆、血清、全血、尿液、羊水;特别优选血浆、血清、全血;最优选血浆。并且,生物样品可在孕妇第11至25孕周时采集获得。
本发明训练计算机时可采用机器学习方法,优选机器学习方法包括以下一种或多种:广义线性模型、梯度提升机、随机森林和支持向量机。
在本发明的模型构建方法中,训练集和验证集可以根据需要按照一定比例进行拆分,优选地,将所有胎膜早破早产的孕妇按照7:3的人数比例随机拆分为训练集和验证集,将所有足月分娩的孕妇按照7:3的人数比例随机拆分为训练集和验证集。最佳基因标志物的筛选在训练集完成,验证集则用于检验最佳基因标志物及模型的预测效果。
在优选的实施方式中,通过比较胎膜早破早产孕妇群体和足月分娩孕妇群体的基因表达谱差异来初步筛选候选的基因标志物,基因标志物可包括mRNA基因和lncRNA基因。该步骤例如可使用DESeq2包(R软件包)实施。对于每一个基因,在两群体中的平均表达量的差异和稳定性会在该步骤中考虑(优选平均表达量差异倍数大于等于2,校正后p值小于0.2),最终通过筛选的基因成为候选的基因标志物。随后,可采用两种模型根据特征重要性进行筛选。两种模型共同使用有利于保证特征的稳定性。优选地,可用广义线性模型和随机森林根据特征重要性进行筛选,例如,每次筛选可从中筛选出30个最重要的分子,筛选过程进行20次,挑选出现频率较高的基因标志物作为最佳基因标志物。
在优选的实施方式中,在训练集中,基于最终筛选出来的最佳基因标记物,采用四种机器学习方法(广义线性模型,梯度提升机、随机森林和支持向量机)进行胎膜早破早产的风险预测。优选每一种算法都采用7折交叉验证的方式挑选出最优参数进行预测模型构建。形成的模型可在验证集中验证效果。
优选地,可通过验证集的效果验证,挑选出效果最优的模型并计算特征重要性。
优选的,mRNA基因和lncRNA基因可共同作为基因标志物进行效果验证,从而构建风险预测模型。
在优选的实施例中,本发明方法构建的预测模型可以在孕中期且最多可以提前23周,以及只需要采取孕妇外周血就可以用无创的方法对胎膜早破早产进行风险预测,预测的灵敏性可达75%,特异性可达83%,接收器工作特性曲线下面积(AUC)在训练集0.94,验证集0.82,均高于现有技术水平。
在第六种典型的实施方式中,提供了一种用于预测孕妇不明原因自发早产风险的方法,该方法包括:
步骤S1:获取来源于所述孕妇的生物样品中基因标志物的表达谱,基因标志物包括以下一种或多种基因:AKAP2、CCNB1IP1、CEACAM19、EMP3、FAR1、FOXN3、GSAP、GTF3C2、HPS3、MTURN、NR1D2、PIK3CG、TMUB1、UPF1、WDR34、ZFR、AC005332.6、AC016727.1、AC018716.2、AC021087.2、AC022613.3、AC084759.3、AC092338.2、AC093525.9、AC099689.1、AC105020.6、AL109936.2、AL138921.1、AL606760.3、AP000688.4、FP671120.4、LINC00221、LINC00511、LINC00689、LINC02076、TTLL10-AS1;
步骤S2:基于基因标志物的表达谱,鉴别孕妇的不明原因自发早产风险。
本申请首次发现孕妇生物样品中的基因标志物与孕妇不明原因自发早产疾病有着显著的相关性,因而可以作为预测孕妇不明原因自发早产的标志物。这些基因标志物包括16个mRNA基因和20个lncRNA基因,其中mRNA基因标志物包括AKAP2、CCNB1IP1、CEACAM19、EMP3、FAR1、FOXN3、GSAP、GTF3C2、HPS3、MTURN、NR1D2、PIK3CG、TMUB1、UPF1、WDR34、ZFR;lncRNA基因标志物包括AC005332.6、AC016727.1、AC018716.2、AC021087.2、AC022613.3、AC084759.3、AC092338.2、AC093525.9、AC099689.1、AC105020.6、AL109936.2、AL138921.1、AL606760.3、AP000688.4、FP671120.4、LINC00221、LINC00511、LINC00689、LINC02076、TTLL10-AS1。
在本发明的方法中,基因标志物优选包括FP671120.4、TTLL10-AS1、AL109936.2、LINC02076、AC021087.2、AL606760.3、AC018716.2、LINC00221、LINC00511、AC099689.1、AC005332.6、AL138921.1、AC093525.9、LINC00689、AP000688.4、AC022613.3、AC105020.6、AC084759.3、AC016727.1、AC092338.2中的一种或多种。
上面列出的各基因可单独或组合使用。例如,可以采用以下全部基因的组合作为基因标志物:FP671120.4、TTLL10-AS1、AL109936.2、LINC02076、AC021087.2、AL606760.3、 AC018716.2、LINC00221、LINC00511、AC099689.1、AC005332.6、AL138921.1、AC093525.9、LINC00689、AP000688.4、AC022613.3、AC105020.6、AC084759.3、AC016727.1、AC092338.2,从而实现不明原因自发早产的风险预测。
在上述步骤S2中,鉴别孕妇的不明原因自发早产风险可以通过利用孕妇不明原因自发早产风险预测模型来实施,通过利用来源于已发生不明原因自发早产的孕妇的生物样品中上述基因标志物的表达谱训练计算机来产生孕妇不明原因自发早产风险预测模型。
训练计算机可通过机器学习方法来实施。机器学习方法选自回归法、分类法或其组合。“机器学习”一般表示在未明确编程的情况下,给予计算机学习能力的算法,包括从数据学习并对数据做出预测的算法。本发明所使用的机器学习方法可以包括随机森林、最小绝对收缩和选择算子逻辑回归、正则化逻辑回归、XGBoost、决策树学习、人工神经网络、深度神经网络、支持向量机、基于规则的机器学习、广义线性模型、梯度提升机等。优选机器学习方法包括以下一种或多种:广义线性模型、梯度提升机、随机森林、支持向量机。
在预测模型中,可通过模型自动计算得出的风险分数,来评价和预测不明原因自发早产风险高低。例如,若风险分数大于0.5,认为不明原因自发早产高风险,若风险分数小于0.5,则认为不明原因自发早产低风险。
来源于孕妇的生物样品可以为以下一种或多种:血浆、血清、全血、尿液、羊水。优选采用来源于孕妇的血浆、血清或全血,用于本发明的检测和鉴别步骤。该生物样品最优选为血浆,例如,可以从孕妇获取外周血并实施血浆分离,从而获得待使用的血浆生物样品。除了血浆、血清或全血,还可以使用其他体液样品,如尿液、羊水等。生物样品的获取可以采用本领域常规的方法实施。
在本发明中,生物样品的采集可以在孕妇第11至25孕周时进行。通过采用上述特定的基因标志物作为预测因子,本发明的应用群体不必区分孕妇是否早产高危,可以适用于一般孕妇群体。利用上述基因标志物,本发明在孕中期可以实现不明原因自发早产的预测。本发明最高可以提早23周实现早产预测。因此,本发明的方法适用人群更广,更具有临床应用性。
在上述方法的步骤S1中,通过对生物样品中的胞外游离RNA(cfRNA)进行定量分析,从而获取所述基因标志物的表达谱;优选地,采用高通量测序法或RT-PCR法对生物样品中的胞外游离RNA进行定量分析;更优选地,采用下一代测序法对生物样品中的胞外游离RNA进行定量分析。
具体来说,生物样品中的胞外游离RNA可采用本领域常用的方法或试剂盒或两者组合提取获得。例如,可以使用TRIzol LS标准的RNA提取步骤,从血浆生物样品中提取胞外游离RNA。
在一种具体的实施方式中,对胞外游离RNA进行定量分析,优选包括利用全转录组测序,使用下一代测序法对孕妇生物样品(优选血浆样品)中的胞外游离RNA进行测序。该方法能 同时对血浆游离mRNA和游离lncRNA进行测序。也可以采用RT-PCR的方法进行分析。还可以采用本领域已知的其他方法如qPCR法对胞外游离RNA的表达谱进行定量分析。
优选地,对胞外游离RNA进行定量分析,还包括将原始的胞外游离RNA测序数据进行质控的步骤,优选包括剪切接头,去除低质量读长,去除<17bp长度的读长,去除rRNA序列和value RNA及Y RNA序列,将剩余读长先比对到人源转录组(顺序为miRNA、tRNA和piRNA,mRNA和lncRNA,最后为其他RNA)。在优选的实施方式中,RNA比对用bowtie软件,定量用RSEM进行。
通过将本发明的基因标志物作为预测孕妇不明原因自发早产风险的标志物,根据现有试剂盒的制备原则,可以制备出针对本发明所述基因标志物的预测试剂盒。还可以针对这些基因标志物,制备出用于预测孕妇不明原因自发早产风险的检测探针、芯片等。
本发明通过采用特定的基因标志物作为检测靶标,基于基因标志物的表达谱与孕妇不明原因自发早产疾病的关联性,实现了对孕妇不明原因自发早产的高特异性和高灵敏性的风险预测。
在第七种典型的实施方式中,本发明提供了一种用于预测孕妇不明原因自发早产风险的试剂盒,该试剂盒包括基因标志物的检测试剂,基因标志物包括以下一种或多种基因:AKAP2、CCNB1IP1、CEACAM19、EMP3、FAR1、FOXN3、GSAP、GTF3C2、HPS3、MTURN、NR1D2、PIK3CG、TMUB1、UPF1、WDR34、ZFR、AC005332.6、AC016727.1、AC018716.2、AC021087.2、AC022613.3、AC084759.3、AC092338.2、AC093525.9、AC099689.1、AC105020.6、AL109936.2、AL138921.1、AL606760.3、AP000688.4、FP671120.4、LINC00221、LINC00511、LINC00689、LINC02076、TTLL10-AS1。采用试剂盒进行预测,使得预测更加方便、简单、快速。优选上述基因标志物包括以下一种或多种基因:FP671120.4、TTLL10-AS1、AL109936.2、LINC02076、AC021087.2、AL606760.3、AC018716.2、LINC00221、LINC00511、AC099689.1、AC005332.6、AL138921.1、AC093525.9、LINC00689、AP000688.4、AC022613.3、AC105020.6、AC084759.3、AC016727.1、AC092338.2。
在试剂盒中,基因标志物的检测试剂可包括用于检测基因标志物的探针和/或引物,具体为一种或多种特异性结合(杂交)至基因标志物的探针和/或一种或多种特异性扩增基因标志物的引物。
由于RNA测序通常包括产生用于测序的cDNA分子的反转录步骤,因而在采用RNA测序时,本发明的试剂盒还可以包含将生物样品中的RNA转化为cDNA片段文库的试剂。
在第八种典型的实施方式中,提供了基因标志物的检测试剂在制备预测孕妇不明原因自发早产风险的试剂盒中的应用,基因标志物包括以下一种或多种基因:AKAP2、CCNB1IP1、CEACAM19、EMP3、FAR1、FOXN3、GSAP、GTF3C2、HPS3、MTURN、NR1D2、PIK3CG、TMUB1、UPF1、WDR34、ZFR、AC005332.6、AC016727.1、AC018716.2、AC021087.2、AC022613.3、AC084759.3、AC092338.2、AC093525.9、AC099689.1、AC105020.6、AL109936.2、AL138921.1、AL606760.3、AP000688.4、FP671120.4、LINC00221、LINC00511、LINC00689、 LINC02076、TTLL10-AS1。优选地,基因标志物包括以下一种或多种基因:FP671120.4、TTLL10-AS1、AL109936.2、LINC02076、AC021087.2、AL606760.3、AC018716.2、LINC00221、LINC00511、AC099689.1、AC005332.6、AL138921.1、AC093525.9、LINC00689、AP000688.4、AC022613.3、AC105020.6、AC084759.3、AC016727.1、AC092338.2。基因标志物的检测试剂包括用于检测基因标志物的探针和/或引物,具体地是一种或多种特异性结合(杂交)至基因标志物的探针和/或一种或多种特异性扩增基因标志物的引物。
在第九种典型的实施方式中,本发明提供了一种用于预测孕妇不明原因自发早产风险的装置,该装置内置有孕妇不明原因自发早产风险预测模型,该预测模型是通过利用来源于已发生不明原因自发早产的孕妇的生物样品中基因标志物的表达谱训练计算机而产生,所述基因标志物包括以下一种或多种基因:AKAP2、CCNB1IP1、CEACAM19、EMP3、FAR1、FOXN3、GSAP、GTF3C2、HPS3、MTURN、NR1D2、PIK3CG、TMUB1、UPF1、WDR34、ZFR、AC005332.6、AC016727.1、AC018716.2、AC021087.2、AC022613.3、AC084759.3、AC092338.2、AC093525.9、AC099689.1、AC105020.6、AL109936.2、AL138921.1、AL606760.3、AP000688.4、FP671120.4、LINC00221、LINC00511、LINC00689、LINC02076、TTLL10-AS1。优选地,基因标志物包括以下一种或多种基因:FP671120.4、TTLL10-AS1、AL109936.2、LINC02076、AC021087.2、AL606760.3、AC018716.2、LINC00221、LINC00511、AC099689.1、AC005332.6、AL138921.1、AC093525.9、LINC00689、AP000688.4、AC022613.3、AC105020.6、AC084759.3、AC016727.1、AC092338.2。在一种优选的实施方式中,该预测模型为广义线性模型、梯度提升机、随机森林或支持向量机模型。
在第十种典型的实施方式中,提供了一种孕妇不明原因自发早产风险预测模型的构建方法,该构建方法包括:检测来源于不明原因自发早产的孕妇群体和足月分娩的孕妇群体的生物样品中的基因标志物的差异表达;将部分不明原因自发早产的孕妇群体和部分足月分娩的孕妇群体作为训练集,利用训练集筛选出最佳基因标志物;在训练集中,利用最佳基因标志物训练计算机,从而得到孕妇不明原因自发早产风险预测模型;将剩余部分的不明原因自发早产的孕妇群体和剩余部分的足月分娩的孕妇群体作为验证集,利用验证集验证孕妇不明原因自发早产风险预测模型;其中,最佳基因标志物包括以下一种或多种基因:AKAP2、CCNB1IP1、CEACAM19、EMP3、FAR1、FOXN3、GSAP、GTF3C2、HPS3、MTURN、NR1D2、PIK3CG、TMUB1、UPF1、WDR34、ZFR、AC005332.6、AC016727.1、AC018716.2、AC021087.2、AC022613.3、AC084759.3、AC092338.2、AC093525.9、AC099689.1、AC105020.6、AL109936.2、AL138921.1、AL606760.3、AP000688.4、FP671120.4、LINC00221、LINC00511、LINC00689、LINC02076、TTLL10-AS1。优选上述最佳基因标志物包括以下一种或多种基因:FP671120.4、TTLL10-AS1、AL109936.2、LINC02076、AC021087.2、AL606760.3、AC018716.2、LINC00221、LINC00511、AC099689.1、AC005332.6、AL138921.1、AC093525.9、LINC00689、AP000688.4、AC022613.3、AC105020.6、AC084759.3、AC016727.1、AC092338.2。
本发明的模型构建方法所采用的生物样品优选为以下一种或多种:血浆、血清、全血、尿液、羊水;特别优选血浆、血清、全血;最优选血浆。并且,生物样品可在孕妇第11至25孕周时采集获得。
本发明训练计算机时可采用机器学习方法,优选机器学习方法包括以下一种或多种:广义线性模型、梯度提升机、随机森林和支持向量机。
在本发明的模型构建方法中,训练集和验证集可以根据需要按照一定比例进行拆分,优选地,将所有不明原因自发早产的孕妇按照7:3的人数比例随机拆分为训练集和验证集,将所有足月分娩的孕妇按照7:3的人数比例随机拆分为训练集和验证集。最佳基因标志物的筛选在训练集完成,验证集则用于检验最佳基因标志物及模型的预测效果。
在优选的实施方式中,通过比较不明原因自发早产孕妇群体和足月分娩孕妇群体的基因表达谱差异来初步筛选候选的基因标志物,基因标志物可包括mRNA基因和lncRNA基因。该步骤例如可使用DESeq2包(R软件包)实施。对于每一个基因,在两群体中的平均表达量的差异和稳定性会在该步骤中考虑(优选平均表达量差异倍数大于等于2,校正后p值小于0.2),最终通过筛选的基因成为候选的基因标志物。随后,可用广义线性模型和随机森林根据特征重要性进行筛选,例如,每次筛选可从中筛选出30个最重要的分子,筛选过程进行20次,挑选出现频率较高的基因标志物作为最佳基因标志物。
在优选的实施方式中,在训练集中,基于最终筛选出来的最佳基因标记物,采用四种机器学习方法(广义线性模型,梯度提升机、随机森林和支持向量机)进行不明原因自发早产的风险预测。优选每一种算法都采用7折交叉验证的方式挑选出最优参数进行预测模型构建。形成的模型可在验证集中验证效果。
优选地,可通过验证集的效果验证,挑选出效果最优的模型并计算特征重要性。
优选的,mRNA基因和lncRNA基因可共同作为基因标志物进行效果验证,从而构建风险预测模型。
在优选的实施例中,本发明方法构建的预测模型可以在孕中期且最多可以提前23周,以及只需要采取孕妇外周血就可以用无创的方法对不明原因自发早产进行风险预测,预测的灵敏性可达74%,特异性可达90%,接收器工作特性曲线下面积(AUC)在训练集0.96,验证集0.91,均高于现有技术水平。
需要说明的是,对于前述的各方法实施例,为了简单描述,故将其都表述为一系列的动作组合,但是本领域技术人员应该知悉,本发明并不受所描述的动作顺序的限制,因为依据本发明,某些步骤可以采用其他顺序或者同时进行。其次,本领域技术人员也应该知悉,说明书中所描述的实施例均属于优选实施例,所涉及的动作并不一定是本发明所必须的。
通过以上的实施方式的描述可知,本领域的技术人员可以清楚地了解到本申请可借助软件加必需的检测仪器等硬件设备的方式来实现。基于这样的理解,本申请的技术方案中数据处理的部分可以以软件产品的形式体现出来,该计算机软件产品可以存储在存储介质中,如ROM/RAM、磁碟、光盘等,包括若干指令用以使得一台计算机设备(可以是个人计算机,服务器,或者网络设备等)执行本申请各个实施例或者实施例的某些部分的方法。
本申请可用于众多通用或专用的计算系统环境或配置中。例如:个人计算机、服务器计算机、手持设备或便携式设备、平板型设备、多处理器系统、基于微处理器的系统、置顶盒、可编程的消费电子设备、网络PC、小型计算机、大型计算机、包括以上任何系统或设备的分布式计算环境等等。
显然,本领域的技术人员应该明白,上述的本申请的部分模块或步骤可以在通用的计算装置来实现,它们可以集中在单个的计算装置上,或者分布在多个计算装置所组成的网络上,可选地,它们可以用计算装置可执行的程序代码来实现,从而,可以将它们存储在存储装置中由计算装置来执行,或者将它们分别制作成各个集成电路模块,或者将它们中的多个模块或步骤制作成单个集成电路模块来实现。这样,本申请不限制于任何特定的硬件和软件结合。
在一种优选的实施例中,提供了一种存储介质,该存储介质包括存储的程序,其中,在程序运行时控制存储介质所在设备执行上述用于预测孕妇胎膜早破早产风险的方法或执行上述孕妇胎膜早破早产风险预测模型的构建方法。
在一种优选的实施例中,提供了一种处理器,处理器用于运行程序,其中,程序运行时执行上述用于预测孕妇胎膜早破早产风险的方法或执行上述孕妇胎膜早破早产风险预测模型的构建方法。
在一种优选的实施例中,提供了一种存储介质,该存储介质包括存储的程序,其中,在程序运行时控制存储介质所在设备执行上述用于预测孕妇不明原因自发早产风险的方法或执行上述孕妇不明原因自发早产风险预测模型的构建方法。
在一种优选的实施例中,提供了一种处理器,处理器用于运行程序,其中,程序运行时执行上述用于预测孕妇不明原因自发早产风险的方法或执行上述孕妇不明原因自发早产风险预测模型的构建方法。
此外,本发明的基因标志物可能对预测孕妇分娩孕周有效果。
下面将结合具体的实施例来进一步说明本申请的有益效果。
实施例1
(1)孕妇血浆样品的获取
277例的单胎孕妇外周血从医院获取,血液收集孕周为11至25,如图1示出。血液来源于早产和足月的孕妇,其中胎膜早破早产104例,不明原因的自发早产74例,足月孕妇99例。早产孕妇从采血到分娩相差孕周为6~23周,如图2示出。所有血液样品立即存储在4℃下,并在8小时内实行血浆分离。血浆分离采用2步离心法,在4℃以1,600g转速离心10分钟,再以12,000g转速离心10分钟。血浆分离之后立即存储在-80℃等待下一步的处理。
(2)胞外游离RNA(cfRNA)的提取
在血浆中加入Trizol LS并立即震荡混匀,后续的cfRNA提取步骤使用TRIzol LS标准的RNA提取方法进行。
(3)cfRNA的测序
cfRNA的测序利用全转录组测序,使用下一代测序法对早产(分别为胎膜早破早产和不明原因自发早产)和足月孕妇的血浆样品进行测序。该方法能同时对血浆游离mRNA和游离lncRNA进行测序。
(4)cfRNA的表达谱定量
将原始的cfRNA测序数据进行质控,包括剪切接头,去除低质量读长,去除<17bp长度的的读长,去除rRNA序列和value RNA及Y RNA序列。将剩余读长比对到人转录组(顺序为miRNA、tRNA和piRNA,mRNA和lncRNA,最后为其他RNA),接着剩余读长比对到人基因组。长RNA(包括mRNA和lncRNA)的表达量矫正为TPM,公式如下:
TPM=(Ni/Li)*1000000/(sum(N1/L1+N2/L2+N3/L3+…+Nn/Ln))
Ni为比对到第i个基因的读长数;Li为第i个基因的长度;sum(N1/L1+N2/L2+...+Nn/Ln)为所有(n个)基因按长度进行标准化之后数值的和。
TotalMappingReads为所有比对上的读长数总和。
(5)最佳基因标志物的筛选
将胎膜早破早产孕妇群体、不明原因自发早产孕妇群体和足月分娩孕妇群体分别按照7:3的比例随机拆分成训练集和验证集,训练集包含72个胎膜早破早产的样本,51个不明原因自发早产的样本和69个足月样本,验证集包含32个胎膜早破早产的样本,23个不明原因自发早产的样本和30个足月样本。基因标志物的筛选在训练集完成,验证集用于检验基因标志物及模型的预测效果。孕妇群体的相关数据请参见表1。
表1:实施例1中早产孕妇群体和足月分娩孕妇群体的相关数据
通过比较胎膜早破早产、不明原因自发早产和足月分娩各孕妇组的表达谱差异来初步筛选候选的基因标志物,该步骤使用DESeq2包(R软件包)实现。对于每一个基因,两组中平均表达量的差异和稳定性在该步骤中加以考虑(平均表达量差异倍数大于等于2,校正后p值小于0.2),最终通过筛选的基因成为候选的基因标志物。用广义线性模型和随机森林根据特征重要性进行筛选,每次筛选都从中筛选出30个最重要的分子。这个过程进行20次,并挑选出现频率较高的基因标志物作为最佳基因标志物。基因标志物的筛选流程图见图3。
对mRNA和lncRNA分子进行特征挑选后,产生21个mRNA基因标志物(CCNB1IP1、COL9A2、DNAJC13、FAM45A、FBXO38、FZD3、HCK、KIAA1257、KIF2A、LARP1B、LRRC56、PLD5、PROS1、SEPT10、SLC41A3、SPIN1、TMUB1、TUBB4A、UPF1、WDR34、ZBTB10)和18个lncRNA基因标志物(AC004803.1、AC009779.2、AC011461.1、AC015878.2、AC016727.1、AC022568.1、AC084759.3、AC092338.2、AC093249.2、AC103876.1、AC105020.6、AC108099.1、AL031733.2、AL451074.2、AP000688.4、NORAD、PINK1-AS、REV3L-IT1),作为胎膜早破早产的最佳基因标志物;产生16个mRNA基因标志物(AKAP2、CCNB1IP1、CEACAM19、EMP3、FAR1、FOXN3、GSAP、GTF3C2、HPS3、MTURN、NR1D2、PIK3CG、TMUB1、UPF1、WDR34、ZFR)和20个lncRNA基因标志物(AC005332.6、AC016727.1、AC018716.2、AC021087.2、AC022613.3、AC084759.3、AC092338.2、AC093525.9、AC099689.1、AC105020.6、AL109936.2、AL138921.1、AL606760.3、AP000688.4、FP671120.4、LINC00221、LINC00511、LINC00689、LINC02076、TTLL10-AS1),作为不明原因自发早产的最佳基因标志物。
本实施例筛选得到的胎膜早破早产的最佳基因标志物和不明原因自发早产的最佳基因标志物分别示出在以下表2和表3中。
表2:实施例1筛选得到的胎膜早破早产的最佳基因标志物的基因和转录本信息
基因名称 | 基因登记号 | 最长转录本登记号 | 最长转录本长度(bp) |
SEPT10 | ENSG00000186522.14_4 | ENST00000356688.8_3 | 3091 |
AC004803.1 | ENSG00000250132.6_6 | ENST00000503602.6_1 | 3207 |
AC009779.2 | ENSG00000258056.2_5 | ENST00000552576.2_1 | 1443 |
AC011461.1 | ENSG00000267197.1_6 | ENST00000592671.1_1 | 357 |
AC015878.2 | ENSG00000265948.1_6 | ENST00000584331.1_1 | 361 |
AC016727.1 | ENSG00000270820.5_5 | ENST00000605437.1_1 | 569 |
AC022568.1 | ENSG00000253474.1_5 | ENST00000518222.1_1 | 833 |
AC084759.3 | ENSG00000280362.1_5 | ENST00000624468.1_1 | 3944 |
AC092338.2 | ENSG00000260790.1_5 | ENST00000568827.1_1 | 1415 |
AC093249.2 | ENSG00000260167.1_5 | ENST00000563540.1_1 | 482 |
AC103876.1 | ENSG00000259986.1_5 | ENST00000568634.1_1 | 418 |
AC105020.6 | ENSG00000275454.1_5 | ENST00000621523.1_1 | 1217 |
AC108099.1 | ENSG00000260786.1_5 | ENST00000564402.1_1 | 3258 |
AL031733.2 | ENSG00000241666.2_5 | ENST00000451992.2_1 | 2523 |
AL451074.2 | ENSG00000224358.1_5 | ENST00000423121.1_1 | 2150 |
AP000688.4 | ENSG00000236677.1_5 | ENST00000436303.1_1 | 446 |
CCNB1IP1 | ENSG00000100814.17_2 | ENST00000437553.6_1 | 1686 |
COL9A2 | ENSG00000049089.14_5 | ENST00000372748.7_2 | 2862 |
DNAJC13 | ENSG00000138246.16_4 | ENST00000260818.10_1 | 7730 |
FAM45A | ENSG00000119979.17_4 | ENST00000648560.1_1 | 2829 |
FBXO38 | ENSG00000145868.16_3 | ENST00000340253.9_2 | 4424 |
FZD3 | ENSG00000104290.10_4 | ENST00000537916.2_1 | 13740 |
HCK | ENSG00000101336.13_4 | ENST00000538448.5_2 | 2215 |
KIAA1257 | ENSG00000114656.11_4 | ENST00000265068.9_3 | 6034 |
KIF2A | ENSG00000068796.16_4 | ENST00000381103.6_1 | 3360 |
LARP1B | ENSG00000138709.18_4 | ENST00000326639.10_3 | 4891 |
LRRC56 | ENSG00000161328.10_2 | ENST00000270115.7_1 | 2769 |
NORAD | ENSG00000260032.1_3 | ENST00000565493.1_1 | 5339 |
PINK1-AS | ENSG00000117242.7_6 | ENST00000451424.1_1 | 4443 |
PLD5 | ENSG00000180287.16_3 | ENST00000536534.6_1 | 8721 |
PROS1 | ENSG00000184500.15_3 | ENST00000394236.8_2 | 3583 |
REV3L-IT1 | ENSG00000229276.1_5 | ENST00000411895.1_1 | 382 |
SLC41A3 | ENSG00000114544.16_4 | ENST00000383598.6_4 | 2330 |
SPIN1 | ENSG00000106723.16_2 | ENST00000375859.3_1 | 4484 |
TMUB1 | ENSG00000164897.12_2 | ENST00000392818.7_1 | 1563 |
TUBB4A | ENSG00000104833.11_5 | ENST00000264071.6_1 | 2552 |
UPF1 | ENSG00000005007.12_3 | ENST00000262803.9_2 | 5348 |
WDR34 | ENSG00000119333.11_2 | ENST00000372715.6_1 | 1755 |
ZBTB10 | ENSG00000205189.11_2 | ENST00000430430.5_1 | 10132 |
表3:实施例1筛选得到的不明原因自发早产的最佳基因标志物的基因和转录本信息
基因名称 | 基因登记号 | 最长转录本登记号 | 最长转录本长度(bp) |
AC005332.6 | ENSG00000278730.1_5 | ENST00000620266.1_1 | 2921 |
AC016727.1 | ENSG00000270820.5_5 | ENST00000605437.1_1 | 569 |
AC018716.2 | ENSG00000255267.3_5 | ENST00000528390.1_1 | 2597 |
AC021087.2 | ENSG00000260774.1_6 | ENST00000565521.1_1 | 2881 |
AC022613.3 | ENSG00000259644.1_6 | ENST00000560740.1_1 | 1346 |
AC084759.3 | ENSG00000280362.1_5 | ENST00000624468.1_1 | 3944 |
AC092338.2 | ENSG00000260790.1_5 | ENST00000568827.1_1 | 1415 |
AC093525.9 | ENSG00000279568.1_5 | ENST00000624961.1_1 | 1746 |
AC099689.1 | ENSG00000279416.1_6 | ENST00000624383.1_1 | 1302 |
AC105020.6 | ENSG00000275454.1_5 | ENST00000621523.1_1 | 1217 |
AKAP2 | ENSG00000241978.9_4 | ENST00000374525.5_2 | 6866 |
AL109936.2 | ENSG00000271420.1_5 | ENST00000605350.1_1 | 650 |
AL138921.1 | ENSG00000227492.1_5 | ENST00000444359.1_1 | 846 |
AL606760.3 | ENSG00000259818.1_5 | ENST00000569869.1_1 | 3591 |
AP000688.4 | ENSG00000236677.1_5 | ENST00000436303.1_1 | 446 |
CCNB1IP1 | ENSG00000100814.17_2 | ENST00000437553.6_1 | 1686 |
CEACAM19 | ENSG00000186567.12_3 | ENST00000358777.8_1 | 2249 |
EMP3 | ENSG00000142227.10_3 | ENST00000270221.10_1 | 876 |
FAR1 | ENSG00000197601.12_3 | ENST00000532502.1_1 | 5820 |
FOXN3 | ENSG00000053254.15_4 | ENST00000345097.8_2 | 7832 |
FP671120.4 | ENSG00000281383.1_5 | ENST00000629969.1_1 | 917 |
GSAP | ENSG00000186088.15_3 | ENST00000257626.11_3 | 3251 |
GTF3C2 | ENSG00000115207.13_3 | ENST00000359541.6_1 | 3992 |
HPS3 | ENSG00000163755.8_3 | ENST00000296051.6_2 | 4665 |
LINC00221 | ENSG00000270816.5_4 | ENST00000603633.2_1 | 1652 |
LINC00511 | ENSG00000227036.7_3 | ENST00000650033.1_1 | 3766 |
LINC00689 | ENSG00000231419.6_2 | ENST00000413238.1_1 | 4684 |
LINC02076 | ENSG00000220161.4_4 | ENST00000577684.1_1 | 1871 |
MTURN | ENSG00000180354.15_4 | ENST00000324453.12_3 | 5937 |
NR1D2 | ENSG00000174738.12_2 | ENST00000312521.8_1 | 5258 |
PIK3CG | ENSG00000105851.10_3 | ENST00000359195.3_1 | 5377 |
TMUB1 | ENSG00000164897.12_2 | ENST00000392818.7_1 | 1563 |
TTLL10-AS1 | ENSG00000205231.1_5 | ENST00000379317.1_1 | 3532 |
UPF1 | ENSG00000005007.12_3 | ENST00000262803.9_2 | 5348 |
WDR34 | ENSG00000119333.11_2 | ENST00000372715.6_1 | 1755 |
ZFR | ENSG00000056097.15_2 | ENST00000265069.12_1 | 4738 |
上述基因标志物的具体序列信息可在Genbank中根据序列编号获取。
(6)基于最佳基因标志物的模型构建及验证
在训练集中,基于最终筛选出来的最佳基因标志物(包括mRNA和lncRNA),采用4种机器学习算法(广义线性模型,梯度提升机、随机森林和支持向量机)进行胎膜早破和不明原因自发早产的风险预测。每一种算法都采用7折交叉验证的方式挑选出最优参数进行预测模型构建。得到的模型在验证集中验证效果,并从中挑选最好的模型作为最优模型(胎膜早破采用随机森林模型作为最优模型;不明原因自发早产采用支持向量机作为最优模型)并计 算特征重要性。mRNA基因标志物和lncRNA基因标志物共同应用验证效果,一起构建模型。模型构建流程图可参见图4。
(7)基因标志物对早产风险的预测效果评估
(7.1)基因标志物对胎膜早破早产风险的预测效果
在筛选获得的针对胎膜早破早产的最佳基因标志物(21个mRNA基因标志物和18个lncRNA基因标志物)中,使用其中20个基因标志物(包括6个mRNA分子和14个lncRNA分子,在图5-A中示出,基因标志物的重要性进行了0到100的归一化)的组合,进行预测效果评估,结果可参见图5-B以及表4(其中,PPROM_Group3代表20个基因标志物的组合)。可以看到,该基因标志物的组合在验证集达到了很好的预测效果,敏感性75%,特异性83%,AUC(Area under the receiver operating characteristic curve,接收者操作特征曲线面积)0.818。同时,单独采用AC084759.3、AC092338.2、AP000688.4以及采用另外两种组合(分别是三个基因标志物的组合PPROM_Group1和六个基因标志物的组合PPROM_Group2)进行预测效果评估,发现这些基因标志物单独或组合使用对胎膜早破早产均有预测效果,结果可参见图5-B以及表4。
(7.2)基因标志物对不明原因自发早产风险的预测效果
在筛选获得的针对不明原因自发早产的最佳基因标志物(16个mRNA基因标志物和20个lncRNA基因标志物)中,使用其中20个基因标志物(包括20个lncRNA分子,在图6-A中示出,基因标志物的重要性进行了0到100的归一化)的组合,进行预测效果评估,结果可参见图6-B以及表4(其中,PTL_Group3代表20个基因标志物的组合)。可以看到,该基因标志物的组合在验证集达到了很好的预测效果,敏感性74%,特异性90%,AUC 0.91。同时,单独采用AC092338.2、AP000688.4、AC016727.1、AC084759.3以及采用另外两种组合(分别是四个基因标志物的组合PTL_Group1和六个基因标志物的组合PTL_Group2)进行预测效果评估,发现这些基因标志物单独或组合使用对不明原因自发早产均有预测效果,结果可参见图5-B以及表4。
从以上的结果可以看出,本发明上述的实施例实现了如下技术效果:利用血浆中本发明多个基因标志物的组合,结合机器学习模型,可最高提早23周预测胎膜早破早产和不明原因自发早产。本发明只需要采取孕妇外周血就可以用无创的方法对早产进行风险预测。本发明的基因标志物可以单独使用或组合使用。在单独使用的情况下,本发明的胎膜早破早产基因标志物的预测灵敏性和特异性分别可至少达到44%和57%,不明原因早产基因标志物的预测灵敏性和特异性分别可至少达到30%和70%,高于现有技术单独采用基因标志物的早产预测效果。本发明的基因标志物在随机组合的情况下,针对胎膜早破早产,可以实现63%以上的预测灵敏性和83%以上的预测特异性,针对不明原因早产,可以实现74%以上的预测灵敏性和80%以上的预测特异性,均高于现有技术水平。在20个基因组合的情况下,胎膜早破早产预测的灵敏性可达75%,特异性可达83%,接收器工作特性曲线下面积(AUC)在训练集达到0.94,验证集达到0.82,均高于现有技术水平;不明原因自发早产预测的灵敏性可达74%,特异性可达90%,接收器工作特性曲线下面积在训练集达到0.96,验证集达到0.91,远远高于现有技术水平。本发明的方法可适用于无症状的一般孕妇群体,不区分是否高危,在孕中期就可以预测,最高可提早23周预测早产,相比于现有技术提前了15周。本发明的方法适用人群更广,更具有临床应用性。经过数据验证,本发明的预测模型的准确性比较高,适合用于早期预测孕妇的早产风险,从而实现尽早干预。
以上所述仅为本发明的优选实施例而已,并不用于限制本发明,对于本领域的技术人员来说,本发明可以有各种更改和变化。凡在本发明的精神和原则之内,所作的任何修改、等同替换、改进等,均应包含在本发明的保护范围之内。
Claims (32)
- 一种用于预测孕妇胎膜早破早产风险的方法,其特征在于,所述方法包括:步骤S1:获取来源于所述孕妇的生物样品中基因标志物的表达谱;步骤S2:基于所述基因标志物的表达谱,鉴别所述孕妇的胎膜早破早产风险。
- 根据权利要求1所述的方法,其特征在于,在步骤S1中,所述基因标志物包括以下一种或多种基因:CCNB1IP1、COL9A2、DNAJC13、FAM45A、FBXO38、FZD3、HCK、KIAA1257、KIF2A、LARP1B、LRRC56、PLD5、PROS1、SEPT10、SLC41A3、SPIN1、TMUB1、TUBB4A、UPF1、WDR34、ZBTB10、AC004803.1、AC009779.2、AC011461.1、AC015878.2、AC016727.1、AC022568.1、AC084759.3、AC092338.2、AC093249.2、AC103876.1、AC105020.6、AC108099.1、AL031733.2、AL451074.2、AP000688.4、NORAD、PINK1-AS、REV3L-IT1。
- 根据权利要求1所述的方法,其特征在于,在步骤S2中,鉴别所述孕妇的胎膜早破早产风险是通过利用孕妇胎膜早破早产风险预测模型来实施的,所述孕妇胎膜早破早产风险预测模型是通过利用来源于已发生胎膜早破早产的孕妇的生物样品中所述基因标志物的表达谱训练计算机而产生。
- 根据权利要求3所述的方法,其特征在于,所述训练计算机是通过机器学习方法来实施,优选所述机器学习方法包括以下一种或多种:广义线性模型、梯度提升机、随机森林、支持向量机。
- 根据权利要求1至4中任一项所述的方法,其特征在于,所述生物样品为以下一种或多种:血浆、血清、全血、尿液、羊水;优选所述生物样品在所述孕妇第11至25孕周时采集获得。
- 根据权利要求1至4中任一项所述的方法,其特征在于,在步骤S1中,通过对所述生物样品中的胞外游离RNA进行定量分析,从而获取所述基因标志物的表达谱;优选地,采用高通量测序法或RT-PCR法对所述生物样品中的胞外游离RNA进行定量分析;更优选地,采用高通量测序法对所述生物样品中的胞外游离RNA进行定量分析。
- 一种用于预测孕妇胎膜早破早产风险的试剂盒,其特征在于,所述试剂盒包括基因标志物的检测试剂,所述基因标志物包括以下一种或多种基因:CCNB1IP1、COL9A2、DNAJC13、FAM45A、FBXO38、FZD3、HCK、KIAA1257、KIF2A、LARP1B、LRRC56、PLD5、PROS1、SEPT10、SLC41A3、SPIN1、TMUB1、TUBB4A、UPF1、WDR34、ZBTB10、AC004803.1、AC009779.2、AC011461.1、AC015878.2、AC016727.1、AC022568.1、AC084759.3、AC092338.2、AC093249.2、AC103876.1、AC105020.6、AC108099.1、AL031733.2、AL451074.2、AP000688.4、NORAD、PINK1-AS、REV3L-IT1。
- 根据权利要求7所述的试剂盒,其特征在于,所述基因标志物的检测试剂包括用于检测所述基因标志物的探针和/或引物;优选为将所述基因标志物的RNA制备成高通量测序文库的相关试剂。
- 基因标志物的检测试剂在制备预测孕妇胎膜早破早产风险的试剂盒中的应用,其特征在于,所述基因标志物包括以下一种或多种基因:CCNB1IP1、COL9A2、DNAJC13、FAM45A、FBXO38、FZD3、HCK、KIAA1257、KIF2A、LARP1B、LRRC56、PLD5、PROS1、SEPT10、SLC41A3、SPIN1、TMUB1、TUBB4A、UPF1、WDR34、ZBTB10、AC004803.1、AC009779.2、AC011461.1、AC015878.2、AC016727.1、AC022568.1、AC084759.3、AC092338.2、AC093249.2、AC103876.1、AC105020.6、AC108099.1、AL031733.2、AL451074.2、AP000688.4、NORAD、PINK1-AS、REV3L-IT1。
- 根据权利要求9所述的应用,其特征在于,所述基因标志物的检测试剂包括用于检测所述基因标志物的探针和/或引物;优选为将所述基因标志物的RNA制备成高通量测序文库的相关试剂。
- 一种用于预测孕妇胎膜早破早产风险的装置,其特征在于,所述装置内置有孕妇胎膜早破早产风险预测模型,所述预测模型是通过利用来源于已发生胎膜早破早产的孕妇的生物样品中基因标志物的表达谱训练计算机而产生,所述基因标志物包括以下一种或多种基因:CCNB1IP1、COL9A2、DNAJC13、FAM45A、FBXO38、FZD3、HCK、KIAA1257、KIF2A、LARP1B、LRRC56、PLD5、PROS1、SEPT10、SLC41A3、SPIN1、TMUB1、TUBB4A、UPF1、WDR34、ZBTB10、AC004803.1、AC009779.2、AC011461.1、AC015878.2、AC016727.1、AC022568.1、AC084759.3、AC092338.2、AC093249.2、AC103876.1、AC105020.6、AC108099.1、AL031733.2、AL451074.2、AP000688.4、NORAD、PINK1-AS、REV3L-IT1。
- 一种孕妇胎膜早破早产风险预测模型的构建方法,其特征在于,所述构建方法包括:检测来源于胎膜早破早产的孕妇群体和足月分娩的孕妇群体的生物样品中的基因标志物的差异表达;将部分所述胎膜早破早产的孕妇群体和部分所述足月分娩的孕妇群体作为训练集,利用所述训练集筛选出最佳基因标志物;在所述训练集中,利用所述最佳基因标志物训练计算机,从而得到孕妇胎膜早破早产风险预测模型;将剩余部分的所述胎膜早破早产的孕妇群体和剩余部分的所述足月分娩的孕妇群体作为验证集,利用所述验证集验证所述孕妇胎膜早破早产风险预测模型;其中,所述最佳基因标志物包括以下一种或多种基因:CCNB1IP1、COL9A2、DNAJC13、FAM45A、FBXO38、FZD3、HCK、KIAA1257、KIF2A、LARP1B、LRRC56、PLD5、PROS1、SEPT10、SLC41A3、SPIN1、TMUB1、TUBB4A、UPF1、WDR34、ZBTB10、 AC004803.1、AC009779.2、AC011461.1、AC015878.2、AC016727.1、AC022568.1、AC084759.3、AC092338.2、AC093249.2、AC103876.1、AC105020.6、AC108099.1、AL031733.2、AL451074.2、AP000688.4、NORAD、PINK1-AS、REV3L-IT1。
- 根据权利要求12所述的构建方法,其特征在于,所述生物样品为以下一种或多种:血浆、血清、全血、尿液、羊水;优选所述生物样品在孕妇第11至25孕周时采集获得。
- 根据权利要求12或13的构建方法,其特征在于,所述训练计算机是通过机器学习方法来实施,优选所述机器学习方法包括以下一种或多种:广义线性模型、梯度提升机、随机森林、支持向量机。
- 一种计算机可读存储介质,其特征在于,所述存储介质包括存储的程序,其中,在所述程序运行时控制所述存储介质所在设备执行权利要求1至6中任一项所述的用于预测孕妇胎膜早破早产风险的方法或权利要求12至14中任一项所述的孕妇胎膜早破早产风险预测模型的构建方法。
- 一种处理器,其特征在于,所述处理器用于运行程序,其中,所述程序运行时执行权利要求1至6中任一项所述的用于预测孕妇胎膜早破早产风险的方法或权利要求12至14中任一项所述的孕妇胎膜早破早产风险预测模型的构建方法。
- 一种用于预测孕妇不明原因自发早产风险的方法,其特征在于,所述方法包括:步骤S1:获取来源于所述孕妇的生物样品中基因标志物的表达谱;步骤S2:基于所述基因标志物的表达谱,鉴别所述孕妇的不明原因自发早产风险。
- 根据权利要求17所述的方法,其特征在于,在步骤S1中,所述基因标志物包括以下一种或多种基因:AKAP2、CCNB1IP1、CEACAM19、EMP3、FAR1、FOXN3、GSAP、GTF3C2、HPS3、MTURN、NR1D2、PIK3CG、TMUB1、UPF1、WDR34、ZFR、AC005332.6、AC016727.1、AC018716.2、AC021087.2、AC022613.3、AC084759.3、AC092338.2、AC093525.9、AC099689.1、AC105020.6、AL109936.2、AL138921.1、AL606760.3、AP000688.4、FP671120.4、LINC00221、LINC00511、LINC00689、LINC02076、TTLL10-AS1。
- 根据权利要求17所述的方法,其特征在于,在步骤S2中,鉴别所述孕妇的不明原因自发早产风险是通过利用孕妇不明原因自发早产风险预测模型来实施的,所述孕妇不明原因自发早产风险预测模型是通过利用来源于已发生不明原因自发早产的孕妇的生物样品中所述基因标志物的表达谱训练计算机而产生。
- 根据权利要求19所述的方法,其特征在于,所述训练计算机是通过机器学习方法来实施,优选所述机器学习方法包括以下一种或多种:广义线性模型、梯度提升机、随机森林、支持向量机。
- 根据权利要求17至20中任一项所述的方法,其特征在于,所述生物样品为以下一种或多种:血浆、血清、全血、尿液、羊水;优选所述生物样品在所述孕妇第11至25孕周时采集获得。
- 根据权利要求17至20中任一项所述的方法,其特征在于,在步骤S1中,通过对所述生物样品中的胞外游离RNA进行定量分析,从而获取所述基因标志物的表达谱;优选地,采用高通量测序法或RT-PCR法对所述生物样品中的胞外游离RNA进行定量分析;更优选地,采用高通量测序法对所述生物样品中的胞外游离RNA进行定量分析。
- 一种用于预测孕妇不明原因自发早产风险的试剂盒,其特征在于,所述试剂盒包括基因标志物的检测试剂,所述基因标志物包括以下一种或多种基因:AKAP2、CCNB1IP1、CEACAM19、EMP3、FAR1、FOXN3、GSAP、GTF3C2、HPS3、MTURN、NR1D2、PIK3CG、TMUB1、UPF1、WDR34、ZFR、AC005332.6、AC016727.1、AC018716.2、AC021087.2、AC022613.3、AC084759.3、AC092338.2、AC093525.9、AC099689.1、AC105020.6、AL109936.2、AL138921.1、AL606760.3、AP000688.4、FP671120.4、LINC00221、LINC00511、LINC00689、LINC02076、TTLL10-AS1。
- 根据权利要求23所述的试剂盒,其特征在于,所述基因标志物的检测试剂包括用于检测所述基因标志物的探针和/或引物;优选为将所述基因标志物的RNA制备成高通量测序文库的相关试剂。
- 基因标志物的检测试剂在制备预测孕妇不明原因自发早产风险的试剂盒中的应用,其特征在于,所述基因标志物包括以下一种或多种基因:AKAP2、CCNB1IP1、CEACAM19、EMP3、FAR1、FOXN3、GSAP、GTF3C2、HPS3、MTURN、NR1D2、PIK3CG、TMUB1、UPF1、WDR34、ZFR、AC005332.6、AC016727.1、AC018716.2、AC021087.2、AC022613.3、AC084759.3、AC092338.2、AC093525.9、AC099689.1、AC105020.6、AL109936.2、AL138921.1、AL606760.3、AP000688.4、FP671120.4、LINC00221、LINC00511、LINC00689、LINC02076、TTLL10-AS1。
- 根据权利要求25所述的应用,其特征在于,所述基因标志物的检测试剂包括用于检测所述基因标志物的探针和/或引物;优选为将所述基因标志物的RNA制备成高通量测序文库的相关试剂。
- 一种用于预测孕妇不明原因自发早产风险的装置,其特征在于,所述装置内置有孕妇不明原因自发早产风险预测模型,所述预测模型是通过利用来源于已发生不明原因自发早产的孕妇的生物样品中基因标志物的表达谱训练计算机而产生,所述基因标志物包括以下一种或多种基因:AKAP2、CCNB1IP1、CEACAM19、EMP3、FAR1、FOXN3、GSAP、GTF3C2、HPS3、MTURN、NR1D2、PIK3CG、TMUB1、UPF1、WDR34、ZFR、AC005332.6、AC016727.1、AC018716.2、AC021087.2、AC022613.3、AC084759.3、AC092338.2、 AC093525.9、AC099689.1、AC105020.6、AL109936.2、AL138921.1、AL606760.3、AP000688.4、FP671120.4、LINC00221、LINC00511、LINC00689、LINC02076、TTLL10-AS1。
- 一种孕妇不明原因自发早产风险预测模型的构建方法,其特征在于,所述构建方法包括:检测来源于不明原因自发早产的孕妇群体和足月的孕妇群体的生物样品中的基因标志物的差异表达;将部分所述不明原因自发早产的孕妇群体和部分所述足月的孕妇群体作为训练集,利用所述训练集筛选出最佳基因标志物;在所述训练集中,利用所述最佳基因标志物训练计算机,从而得到孕妇不明原因自发早产风险预测模型;将剩余部分的所述不明原因自发早产的孕妇群体和剩余部分的所述足月的孕妇群体作为验证集,利用所述验证集验证所述孕妇不明原因自发早产风险预测模型;其中,所述最佳基因标志物包括以下一种或多种基因:AKAP2、CCNB1IP1、CEACAM19、EMP3、FAR1、FOXN3、GSAP、GTF3C2、HPS3、MTURN、NR1D2、PIK3CG、TMUB1、UPF1、WDR34、ZFR、AC005332.6、AC016727.1、AC018716.2、AC021087.2、AC022613.3、AC084759.3、AC092338.2、AC093525.9、AC099689.1、AC105020.6、AL109936.2、AL138921.1、AL606760.3、AP000688.4、FP671120.4、LINC00221、LINC00511、LINC00689、LINC02076、TTLL10-AS1。
- 根据权利要求28所述的构建方法,其特征在于,所述生物样品为以下一种或多种:血浆、血清、全血、尿液、羊水;优选所述生物样品在孕妇第11至25孕周时采集获得。
- 根据权利要求28或29的构建方法,其特征在于,所述训练计算机是通过机器学习方法来实施,优选所述机器学习方法包括以下一种或多种:广义线性模型、梯度提升机、随机森林、支持向量机。
- 一种计算机可读存储介质,其特征在于,所述存储介质包括存储的程序,其中,在所述程序运行时控制所述存储介质所在设备执行权利要求17至22中任一项所述的用于预测孕妇不明原因自发早产风险的方法或权利要求28至30中任一项所述的孕妇不明原因自发早产风险预测模型的构建方法。
- 一种处理器,其特征在于,所述处理器用于运行程序,其中,所述程序运行时执行权利要求17至22中任一项所述的用于预测孕妇不明原因自发早产风险的方法或权利要求28至30中任一项所述的孕妇不明原因自发早产风险预测模型的构建方法。
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/CN2021/136566 WO2023102786A1 (zh) | 2021-12-08 | 2021-12-08 | 基因标志物在预测孕妇早产风险中的应用 |
CN202180102281.XA CN118056016A (zh) | 2021-12-08 | 2021-12-08 | 基因标志物在预测孕妇早产风险中的应用 |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/CN2021/136566 WO2023102786A1 (zh) | 2021-12-08 | 2021-12-08 | 基因标志物在预测孕妇早产风险中的应用 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2023102786A1 true WO2023102786A1 (zh) | 2023-06-15 |
Family
ID=86729272
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2021/136566 WO2023102786A1 (zh) | 2021-12-08 | 2021-12-08 | 基因标志物在预测孕妇早产风险中的应用 |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN118056016A (zh) |
WO (1) | WO2023102786A1 (zh) |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101309929A (zh) * | 2005-09-15 | 2008-11-19 | 创源生物科技(武汉)有限公司 | 胎膜早破的一个标志物 |
CN109142565A (zh) * | 2018-07-27 | 2019-01-04 | 重庆早柒天生物科技股份有限公司 | 基于iTRAQ技术的胎膜早破孕妇阴道分泌物差异蛋白的筛选方法 |
CN110191963A (zh) * | 2016-08-05 | 2019-08-30 | 赛拉预测公司 | 用于预测由于早产胎膜早破相对于特发性自然分娩所造成的早产的生物标志物 |
CN113692624A (zh) * | 2019-02-14 | 2021-11-23 | 米尔维公司 | 用于确定受试者的妊娠相关状态的方法和系统 |
-
2021
- 2021-12-08 WO PCT/CN2021/136566 patent/WO2023102786A1/zh unknown
- 2021-12-08 CN CN202180102281.XA patent/CN118056016A/zh active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101309929A (zh) * | 2005-09-15 | 2008-11-19 | 创源生物科技(武汉)有限公司 | 胎膜早破的一个标志物 |
CN110191963A (zh) * | 2016-08-05 | 2019-08-30 | 赛拉预测公司 | 用于预测由于早产胎膜早破相对于特发性自然分娩所造成的早产的生物标志物 |
CN109142565A (zh) * | 2018-07-27 | 2019-01-04 | 重庆早柒天生物科技股份有限公司 | 基于iTRAQ技术的胎膜早破孕妇阴道分泌物差异蛋白的筛选方法 |
CN113692624A (zh) * | 2019-02-14 | 2021-11-23 | 米尔维公司 | 用于确定受试者的妊娠相关状态的方法和系统 |
Non-Patent Citations (1)
Title |
---|
VILLE YVES; ROZENBERG PATRICK: "Predictors of preterm birth", BAILLIERE'S BEST PRACTICE AND RESEARCH. CLINICAL OBSTETRICS ANDGYNAECOLOGY, BAILLIERE TINDALL, LONDON, GB, vol. 52, 7 July 2018 (2018-07-07), GB , pages 23 - 32, XP085545852, ISSN: 1521-6934, DOI: 10.1016/j.bpobgyn.2018.05.002 * |
Also Published As
Publication number | Publication date |
---|---|
CN118056016A (zh) | 2024-05-17 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Liao et al. | Noninvasive prenatal diagnosis of common aneuploidies by semiconductor sequencing | |
WO2020168118A1 (en) | Methods and systems for determining a pregnancy-related state of a subject | |
EP4073805B1 (en) | Systems and methods for predicting homologous recombination deficiency status of a specimen | |
AU2020244763A1 (en) | Systems and methods for deriving and optimizing classifiers from multiple datasets | |
US20200219587A1 (en) | Systems and methods for using fragment lengths as a predictor of cancer | |
US20190228131A1 (en) | Novel method capable of differentiating fetal sex and fetal sex chromosome abnormality on various platforms | |
Reggiardo et al. | LncRNA biomarkers of inflammation and cancer | |
CN111833963A (zh) | 一种cfDNA分类方法、装置和用途 | |
Camunas-Soler et al. | Predictive RNA profiles for early and very early spontaneous preterm birth | |
Zhou et al. | Maternal blood EBF1-based microRNA transcripts as biomarkers for detecting risk of spontaneous preterm birth: A nested case-control study | |
Budis et al. | Combining count-and length-based z-scores leads to improved predictions in non-invasive prenatal testing | |
CN112382341A (zh) | 一种用于鉴定食管鳞癌预后相关的生物标志物的方法 | |
Wong et al. | Regional and bilateral MRI and gene signatures in facioscapulohumeral dystrophy: implications for clinical trial design and mechanisms of disease progression | |
WO2023102786A1 (zh) | 基因标志物在预测孕妇早产风险中的应用 | |
US20220042109A1 (en) | Methods of assessing breast cancer using circulating hormone receptor transcripts | |
US11535896B2 (en) | Method for analysing cell-free nucleic acids | |
WO2023246808A1 (zh) | 利用癌症中剪接异常的短外显子辅助癌症诊断和预后 | |
WO2023102840A1 (zh) | 基因标志物在预测孕妇子痫前期风险中的应用 | |
US20240150837A1 (en) | Methods and systems for methylation profiling of pregnancy-related states | |
CN116287175A (zh) | 标志物在制备预测妊娠期肝内胆汁淤积症的相关产品中的应用 | |
CN116656802A (zh) | 一种用于预测早期慢阻肺的生物标志物及其应用和筛选方法 | |
WO2024155681A1 (en) | Methods and systems for detecting and assessing liver conditions | |
Wong et al. | Validation of the association between MRI and gene signatures in facioscapulohumeral dystrophy muscle: implications for clinical trial design | |
Diaz-Gimeno | Asynchronous and pathological windows of implantation: two causes of recurrent implantation failure | |
WO2021252574A1 (en) | Nucleic acid biomarkers for placental dysfunction |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 21966710 Country of ref document: EP Kind code of ref document: A1 |