Biomarkers and methods for predicting preterm birth

Abstract

To provide biomarker panels, methods and kits for determining the probability for preterm birth in a pregnant female.SOLUTION: The present disclosure is based, in part, on the discovery that certain proteins and peptides in biological samples obtained from a pregnant female are differentially expressed in pregnant females that have an increased risk of developing preterm birth in the future or presently suffering from preterm birth relative to matched controls. The present disclosure is further based, in part, on the unexpected discovery that panels combining one or more of these proteins and peptides can be utilized in methods of determining the probability for preterm birth in a pregnant female with relatively high sensitivity and specificity.SELECTED DRAWING: None

Description

[Technical field]

[0001] This application claims the benefit of U.S. Provisional Patent Application No. 61 / 919,586, filed December 20, 2013, and U.S. Provisional Patent Application No. 61 / 798,504, filed March 15, 2013, the entire contents of which are incorporated herein by reference.

[0002] The present invention relates generally to the field of personalized medicine, and more specifically to compositions and methods for determining the probability for preterm birth in pregnant women. [Background technology]

[0003] background According to the World Health Organization, an estimated 15 million babies are born prematurely (before 37 weeks of gestation) each year. Premature birth rates are increasing in almost every country with reliable data. World Health Organization; March of Dimes; The Partnership for Maternal, Newborn & Child Health; Save the See Children, Born too soon: the global action report on preterm birth, ISBN9789241503433 (2012). An estimated one million infants die each year from complications of preterm birth. Globally, premature birth is the number one cause of death in newborns (infants in the first four weeks of life) and the second leading cause of death in children under five after pneumonia. Many survivors face a life of disabilities, including learning disabilities and vision and hearing problems.

[0004] Across 184 countries with reliable data, preterm birth rates range from 5% to 18% of infants born. Blencowe et al., "National, regional and worldwide estimates of preterm birth." The Lancet, 9;379(9832):2162-72 (2012). Preterm birth is nevertheless a global problem, with over 60% of preterm births occurring in Africa and South Asia. Countries with the highest numbers include Brazil, India, Nigeria, and the United States. Of the 11 countries with preterm birth rates above 15%, all but two are in sub-Saharan Africa. In the poorest countries, on average, 12% of infants are born too early, compared with 9% in high-income countries. Within countries, poor families are at higher risk. Over three-quarters of premature babies could be helped with feasible, cost-effective care, such as prenatal steroid injections given to pregnant women at risk of preterm labor to strengthen the baby's lungs.

[0005] Infants born preterm are at greater risk for death and a variety of health and development problems than infants born at term. Complications include acute respiratory, gastrointestinal, immune, central nervous system, hearing, and visual problems, as well as long-term motor, cognitive, visual, hearing, behavioral, social-emotional, health, and growth problems. The birth of a preterm infant also results in significant emotional and economic costs to families and can have implications for public services, such as health insurance, education, and other social support systems. The greatest risk of death and morbidity is for infants born at the earliest gestational ages. However, infants born closer to term represent the majority of infants born preterm and experience more complications than infants born at term.

[0006] To prevent premature birth in women less than 24 weeks pregnant who have an open cervix on ultrasound, a surgical procedure known as cervical cerclage can be used to close the cervix with strong sutures. For women less than 34 weeks pregnant and in active premature labor, hospitalization and administration of medications to temporarily stop premature labor and / or promote fetal lung development may be required. If a pregnant woman is determined to be at risk for premature birth, health care providers can implement a variety of clinical strategies, which may include preventive medications, such as hydroxyprogesterone caproate (Makena) injections and / or vaginal progesterone gel, cervical pessaries, restrictions on sexual activity and / or other physical activity, and changes in treatment for chronic conditions that increase the risk of premature labor, such as diabetes and hypertension.

[0007] There is a high need to identify women at risk for preterm birth and provide them with appropriate prenatal care. In women identified as high risk, more intensive prenatal surveillance and preventive interventions can be planned. Current strategies for risk assessment are based on obstetric and medical history and clinical examination, but these strategies can only identify a small percentage of women at risk for preterm birth. Reliable early identification of risk for preterm birth would allow for the planning of appropriate monitoring and clinical management to prevent preterm birth. Such monitoring and management may include more frequent prenatal visits, serial cervical length measurements, enhanced education on signs and symptoms of early preterm birth, lifestyle interventions for modifiable risk behaviors, cervical pessaries and progesterone treatments. Finally, reliable prenatal identification of risk for preterm birth is also critical for the cost-effective allocation of monitoring resources. The present invention addresses this need by providing compositions and methods for determining whether a pregnant woman is at risk for preterm birth, and related advantages are also provided. [Prior art documents] [Non-patent literature]

[0008] [Non-Patent Document 1] Blencowe et al., "National, regional and worldwide estimates of preterm birth.", The Lancet, 9;379(9832):2162-72 (2012). Summary of the Invention [Problem to be solved by the invention]

[0009] The present invention provides compositions and methods for predicting the likelihood of preterm birth in a pregnant woman. [Means for solving the problem]

[0010] Abstract In one aspect, the invention provides a panel of isolated biomarkers comprising N of the biomarkers listed in Tables 1-63. In some embodiments, N is a number selected from the group consisting of 2 to 24. In additional embodiments, the biomarker panel comprises at least two of the isolated biomarkers selected from the group consisting of AFTECCVVASQLR, ELLESYIDGR, and ITLPDFTGDLR. In additional embodiments, the biomarker panel comprises at least two of the isolated biomarkers selected from the group consisting of FLNWIK, FGFGGSTDSGPIR, LLELTGPK, VEHSDLSFSK, IEGNLIFDPNYLPK, ALVLELAK, TQILEWAAER, DVLLLVHNLPQNLPGYFWYK, SEPRPGVLLR, ITQDAQLK, ALDLSLK, WWGGQPLWITATK, and LSETNR.

[0011] In a further embodiment, the biomarker panel comprises at least two isolated biomarkers selected from the group consisting of the biomarkers shown in Table 50 and the biomarkers shown in Table 52.

[0012] In a further aspect, the invention provides a panel of isolated biomarkers comprising N of the biomarkers listed in Tables 1 to 63. In some embodiments, N is a number selected from the group consisting of 2 to 24. In additional embodiments, the biomarker panel comprises at least two isolated biomarkers selected from the group consisting of the biomarkers shown in Table 50 and the biomarkers shown in Table 52.

[0013] In some embodiments, the invention provides a biomarker panel comprising at least two isolated biomarkers selected from the group consisting of lipopolysaccharide binding protein (LBP), prothrombin (THRB), complement component C5 (C5 or CO5), plasminogen (PLMN), and complement component C8 gamma chain (C8G or CO8G).

[0014] In some embodiments, the invention provides a biomarker panel comprising at least two of the isolated biomarkers selected from the group consisting of alpha-1B-glycoprotein (A1BG), a disintegrin and metalloproteinase domain-containing protein 12 (ADA12), apolipoprotein B-100 (APOB), beta-2-microglobulin (B2MG), CCAAT / enhancer binding protein alpha / beta (HP8 peptide), corticosteroid binding globulin (CBG), complement component C6, endoglin (EGLN), ectonucleotide pyrophosphatase / phosphodiesterase family member 2 (ENPP2), coagulation factor VII (FA7), hyaluronan binding protein 2 (HABP2), pregnancy specific beta-1-glycoprotein 9 (PSG9), inhibin beta E chain (INHBE).

[0015] In other embodiments, the invention provides a biomarker panel that includes lipopolysaccharide binding protein (LBP), prothrombin (THRB), complement component C5 (C5 or CO5), plasminogen (PLMN), complement component C8 gamma chain (C8G or CO8G), complement component 1, q subcomponent, B chain (C1QB), fibrinogen beta chain (FIBB or FIB), C-reactive protein (CRP), inter-alpha-trypsin inhibitor heavy chain H4 (ITIH4), chorionic somatomammotropic hormone (CSH), and angiotensinogen (ANG or ANGT).

[0016] In other embodiments, the invention provides a biomarker panel comprising alpha-1B-glycoprotein (A1BG), a disintegrin and metalloproteinase domain-containing protein 12 (ADA12), apolipoprotein B-100 (APOB), beta-2-microglobulin (B2MG), CCAAT / enhancer binding protein alpha / beta (HP8 peptide), corticosteroid binding globulin (CBG), complement component C6, endoglin (EGLN), ectonucleotide pyrophosphatase / phosphodiesterase family member 2 (ENPP2), coagulation factor VII (FA7), hyaluronan binding protein 2 (HABP2), pregnancy specific beta-1-glycoprotein 9 (PSG9), inhibin beta E chain (INHBE).

[0017] In additional embodiments, the invention provides a biomarker panel comprising at least two isolated biomarkers selected from the group consisting of the biomarkers shown in Table 51 and the biomarkers shown in Table 53.

[0018] The invention also provides a method of determining the probability of preterm birth in a pregnant woman, comprising detecting a measurable feature for each of N biomarkers selected from the biomarkers listed in Tables 1 to 63 in a biological sample obtained from the pregnant woman, and analyzing the measurable features to determine the probability of preterm birth in the pregnant woman. In some embodiments, the invention provides a method of predicting GAB, comprising detecting a measurable feature for each of N biomarkers selected from the biomarkers listed in Tables 1 to 63 in a biological sample obtained from the pregnant woman, and analyzing the measurable features to predict GAB.

[0019] In some embodiments, the measurable feature comprises a fragment or derivative of each of the N biomarkers selected from the biomarkers listed in Tables 1 to 63. In some embodiments of the disclosed methods, detecting the measurable feature comprises quantifying the amount of each of the N biomarkers selected from the biomarkers listed in Tables 1 to 63, combinations or portions and / or derivatives thereof, in a biological sample obtained from the pregnant woman. In additional embodiments, the disclosed methods of determining a probability for preterm birth in a pregnant woman further comprise detecting a measurable feature for one or more risk indicators associated with preterm birth.

[0020] In some embodiments, the disclosed method of determining the probability for preterm birth in a pregnant woman and related methods disclosed herein comprises detecting a measurable feature for each of N biomarkers, where N is selected from the group consisting of 2 to 24. In further embodiments, the disclosed method of determining the probability for preterm birth in a pregnant woman and related methods disclosed herein comprises detecting a measurable feature for each of at least two isolated biomarkers selected from the group consisting of AFTECCVVASQLR, ELLESYIDGR, and ITLPDFTGDLR. In further embodiments, the disclosed method of determining the probability for preterm birth in a pregnant woman and related methods disclosed herein comprises detecting a measurable feature for each of at least two isolated biomarkers selected from the group consisting of FLNWIK, FGFGGSTDSGPIR, LLELTGPK, VEHSDLSFSK, IEGNLIFDPNYLPK, ALVLELAK, TQILEWAAER, DVLLLVHNLPQNLPGYFWYK, SEPRPGVLLR, ITQDAQLK, ALDLSLK, WWGGQPLWITATK, and LSETNR. In a further embodiment, the disclosed method of determining the probability of preterm birth in a pregnant woman and related methods disclosed herein comprises detecting a measurable feature of each of at least two isolated biomarkers selected from the group consisting of the biomarkers shown in Table 50 and the biomarkers shown in Table 52.

[0021] In other embodiments, the disclosed method of determining the probability for preterm birth in a pregnant woman comprises detecting a measurable characteristic of each of at least two isolated biomarkers selected from the group consisting of lipopolysaccharide binding protein (LBP), prothrombin (THRB), complement component C5 (C5 or CO5), plasminogen (PLMN), and complement component C8 gamma chain (C8G or CO8G).

[0022] In other embodiments, the disclosed method of determining the probability for preterm birth in a pregnant woman comprises detecting a measurable characteristic of each of at least two isolated biomarkers selected from the group consisting of alpha-1B-glycoprotein (A1BG), a disintegrin and metalloproteinase domain-containing protein 12 (ADA12), apolipoprotein B-100 (APOB), beta-2-microglobulin (B2MG), CCAAT / enhancer binding protein alpha / beta (HP8 peptide), corticosteroid binding globulin (CBG), complement component C6, endoglin (EGLN), ectonucleotide pyrophosphatase / phosphodiesterase family member 2 (ENPP2), coagulation factor VII (FA7), hyaluronan binding protein 2 (HABP2), pregnancy specific beta-1-glycoprotein 9 (PSG9), inhibin beta E chain (INHBE).

[0023] In other embodiments, the disclosed method of determining the probability for preterm birth in a pregnant woman comprises detecting a measurable characteristic of each of at least two isolated biomarkers selected from the group consisting of lipopolysaccharide binding protein (LBP), prothrombin (THRB), complement component C5 (C5 or CO5), plasminogen (PLMN), complement component C8 gamma chain (C8G or CO8G), complement component 1, q subcomponent, B chain (C1QB), fibrinogen beta chain (FIBB or FIB), C-reactive protein (CRP), inter-alpha-trypsin inhibitor heavy chain H4 (ITIH4), chorionic somatomammotropic hormone (CSH), and angiotensinogen (ANG or ANGT).

[0024] In a further embodiment, the disclosed method of determining the probability of preterm birth in a pregnant woman comprises detecting a measurable feature of each of at least two isolated biomarkers selected from the group consisting of the biomarkers shown in Table 51 and the biomarkers shown in Table 53.

[0025] In some embodiments of the methods of determining the probability for preterm birth in a pregnant woman, the probability of preterm birth in the pregnant woman is calculated based on the quantified amount of each of N biomarkers selected from the biomarkers listed in Tables 1 to 63. In some embodiments, the disclosed methods for determining the probability of preterm birth include detecting and / or quantifying one or more biomarkers using mass spectrometry, a capture agent, or a combination thereof.

[0026] In some embodiments, the disclosed methods of determining the probability of preterm birth in a pregnant woman include an initial step of providing a biomarker panel that includes N of the biomarkers listed in Tables 1 to 63. In additional embodiments, the disclosed methods of determining the probability of preterm birth in a pregnant woman include an initial step of providing a biological sample from the pregnant woman.

[0027] In some embodiments, the disclosed method of determining the probability of preterm birth in a pregnant woman includes communicating the probability to a health care provider. In additional embodiments, the communication informs subsequent treatment decisions for the pregnant woman. In further embodiments, the one or more treatment decisions are selected from the group consisting of more frequent prenatal visits, serial cervical length measurements, increased education on the signs and symptoms of early preterm birth, lifestyle interventions for modifiable risk behaviors, and progesterone treatment.

[0028] In further embodiments, the disclosed methods of determining the probability of preterm birth in a pregnant woman include analyzing a measurable characteristic of one or more isolated biomarkers using a predictive model. In some embodiments of the disclosed methods, the measurable characteristic of one or more isolated biomarkers is compared to a reference characteristic.

[0029] In additional embodiments, the disclosed method of determining the probability of preterm birth in a pregnant woman includes using one or more analyses selected from a linear discriminant analysis model, a support vector machine classification algorithm, a recursive feature elimination model, a predictive analysis of a microarray model, a logistic regression model, a CART algorithm, a flex tree algorithm, a LART algorithm, a random forest algorithm, a MART algorithm, a machine learning algorithm, a penalized regression method, and combinations thereof. In one embodiment, the disclosed method of determining the probability of preterm birth in a pregnant woman includes logistic regression.

[0030] In some embodiments, the present invention provides a method of determining a probability for preterm birth in a pregnant woman comprising quantifying the amount of each of N biomarkers selected from the biomarkers listed in Tables 1 to 63 in a biological sample obtained from the pregnant woman, multiplying the amounts by a predetermined coefficient, and adding the individual products to obtain a total risk score corresponding to the probability.

[0031] In an additional embodiment, the present invention provides a method of predicting GAB comprising: (a) quantifying the amount of each of N biomarkers selected from the biomarkers listed in Tables 1 to 63 in a biological sample obtained from the pregnant woman; (b) multiplying or thresholding the amounts by a predetermined factor; and (c) determining the predicted GAB birth in the pregnant woman comprising adding the individual products to obtain a total risk score corresponding to the predicted GAB.

[0032] In a further embodiment, the present invention provides a method of predicting time to delivery in a pregnant woman, comprising: (a) obtaining a biological sample from the pregnant woman; (b) quantifying in the biological sample the amount of each of N biomarkers selected from the biomarkers listed in Tables 1 to 63; (c) multiplying or thresholding the amounts by a predetermined coefficient; (d) determining a predicted GAB in the pregnant woman, comprising adding the individual products to obtain a total risk score corresponding to a predicted GAB; and (e) subtracting an estimated gestational age (GA) at the time the biological sample was obtained from the predicted GAB to predict time to delivery in the pregnant woman.

[0033] Other features and advantages of the invention will become apparent from the detailed description, and from the claims. In an embodiment of the present invention, for example, the following items are provided: (Item 1) A panel of isolated biomarkers comprising N of the biomarkers listed in Tables 1 to 63. (Item 2) 2. The panel of item 1, wherein N is a number selected from the group consisting of 2 to 24. (Item 3) The panel of item 2, comprising at least two of the isolated biomarkers selected from the group consisting of AFTECCVVASQLR, ELLESYIDGR, ITLPDFTGDLR, the biomarkers shown in Table 50, and the biomarkers shown in Table 52. (Item 4) The panel of item 2, comprising lipopolysaccharide binding protein (LBP), prothrombin (THRB), complement component C5 (C5 or CO5), plasminogen (PLMN), and complement component C8 gamma chain (C8G or CO8G). (Item 5) The panel of item 2, comprising at least two isolated biomarkers selected from the group consisting of lipopolysaccharide binding protein (LBP), prothrombin (THRB), complement component C5 (C5 or CO5), plasminogen (PLMN), and complement component C8 gamma chain (C8G or CO8G). (Item 6) The panel of item 2, comprising at least two isolated biomarkers selected from the group consisting of lipopolysaccharide binding protein (LBP), prothrombin (THRB), complement component C5 (C5 or CO5), plasminogen (PLMN), complement component C8 gamma chain (C8G or CO8G), complement component 1, q subcomponent, B chain (C1QB), fibrinogen beta chain (FIBB or FIB), C-reactive protein (CRP), inter-alpha-trypsin inhibitor heavy chain H4 (ITIH4), chorionic somatomammotropic hormone (CSH), angiotensinogen (ANG or ANGT), the biomarkers shown in Table 51, and the biomarkers shown in Table 53. (Item 7) 1. A method for determining a probability for preterm birth in a pregnant woman, comprising detecting a measurable feature of each of N biomarkers selected from the biomarkers listed in Tables 1 to 63 in a biological sample obtained from the pregnant woman, and analyzing the measurable features to determine the probability for preterm birth in the pregnant woman. (Item 8) 8. The method of claim 7, wherein the measurable features comprise fragments or derivatives of each of the N biomarkers selected from the biomarkers listed in Tables 1 to 63. (Item 9) 8. The method of claim 7, wherein said detecting a measurable feature comprises quantifying the amount of each of N biomarkers selected from the biomarkers listed in Tables 1 to 63, combinations or portions and / or derivatives thereof in a biological sample obtained from the pregnant female. (Item 10) 10. The method of claim 9, further comprising calculating the probability of preterm birth in the pregnant woman based on the quantified amount of each of N biomarkers selected from the biomarkers listed in Tables 1 to 63. (Item 11) The method of claim 7, further comprising the initial step of providing a biomarker panel comprising N of the biomarkers listed in Tables 1 to 63. (Item 12) 8. The method of claim 7, further comprising the initial step of providing a biological sample from the pregnant female. (Item 13) 8. The method of claim 7, further comprising communicating said probability to a health care provider. (Item 14) 14. The method of claim 13, wherein said communication informs subsequent treatment decisions for said pregnant woman. (Item 15) 8. The method of claim 7, wherein N is a number selected from the group consisting of 2 to 24. (Item 16) 16. The method of claim 15, wherein the N biomarkers comprise at least two isolated biomarkers selected from the group consisting of AFTECCVVASQLR, ELLESYIDGR, ITLPDFTGDLR, the biomarkers shown in Table 50, and the biomarkers shown in Table 52. (Item 17) 8. The method of claim 7, wherein the analysis comprises the use of a predictive model. (Item 18) 18. The method of claim 17, wherein the analyzing comprises comparing the measurable feature to a reference feature. (Item 19) 20. The method of claim 18, wherein the analysis comprises using one or more selected from the group consisting of a linear discriminant analysis model, a support vector machine classification algorithm, a recursive feature elimination model, predictive analysis of microarray models, a logistic regression model, a multiple regression model, a survival model, a CART algorithm, a flextree algorithm, a LART algorithm, a random forest algorithm, a MART algorithm, a machine learning algorithm, a penalized regression method, and combinations thereof. (Item 20) 20. The method of claim 19, wherein the analysis comprises logistic regression. (Item 21) 8. The method of claim 7, wherein the probability is expressed as a risk score. (Item 22) 8. The method of claim 7, wherein the biological sample is selected from the group consisting of whole blood, plasma, and serum. (Item 23) 23. The method of claim 22, wherein the biological sample is serum. (Item 24) 8. The method of claim 7, wherein the quantifying comprises mass spectrometry (MS). (Item 25) 25. The method of claim 24, wherein the MS comprises liquid chromatography mass spectrometry (LC-MS). (Item 26) 25. The method of claim 24, wherein the MS comprises multiple reaction monitoring (MRM) or selected reaction monitoring (SRM). (Item 27) 27. The method of claim 26, wherein the MRM (or SRM) comprises scheduled MRM (SRM). (Item 28) 8. The method of claim 7, wherein the quantifying comprises an assay using a capture agent. (Item 29) 29. The method of claim 28, wherein the capture agent is selected from the group consisting of an antibody, an antibody fragment, a nucleic acid-based protein binding reagent, a small molecule, or a variant thereof. (Item 30) 29. The method of claim 28, wherein the assay is selected from the group consisting of an enzyme immunoassay (EIA), an enzyme linked immunosorbent assay (ELISA), and a radioimmunoassay (RIA). (Item 31) 31. The method of claim 30, wherein said quantifying further comprises mass spectrometry (MS). (Item 32) 32. The method of claim 31, wherein the quantifying comprises co-immunoprecipitation-mass spectrometry (co-IP MS). (Item 33) 8. The method of claim 7, further comprising detecting a measurable characteristic for one or more risk indicators. (Item 34) 8. The method of claim 7, wherein said analyzing said measurable characteristics first includes predicting gestational age at birth (GAB) prior to said determining said probability for preterm birth. (Item 35) 35. The method of claim 34, wherein the prediction of the GAB is used to determine the probability of preterm birth. (Item 36) 34. The method of claim 33, wherein the one or more risk indicators are selected from the group consisting of age, previous pregnancies, previous history of low birth weight or early birth, multiple second trimester spontaneous abortions, previous first trimester induced abortions, familial and intergenerational factors, infertility history, nulliparity, placental abnormalities, cervical and uterine abnormalities, pregnancy bleeding, intrauterine growth restriction, intrauterine diethylstilbestrol exposure, multiple pregnancy, infant sex, short stature, low pre-pregnancy weight / low body mass index, diabetes, hypertension, hypothyroidism, asthma, education level, tobacco use, and urogenital infections. (Item 37) 1. A method for determining a probability for preterm birth in a pregnant woman, comprising: (a) quantifying the amount of each of N biomarkers selected from the biomarkers listed in Tables 1 to 63 in a biological sample obtained from the pregnant woman; (b) multiplying the amounts by a predetermined coefficient; and (c) adding the individual products to obtain a total risk score corresponding to the probability, thereby determining the probability for preterm birth in the pregnant woman. (Item 38) A method for predicting GAB, comprising detecting a measurable feature of each of N biomarkers selected from the biomarkers listed in Tables 1 to 63 in a biological sample obtained from a pregnant woman, and analyzing the measurable features to predict GAB. (Item 39) 39. The method of claim 38, wherein the measurable feature comprises a fragment or derivative of each of the N biomarkers selected from the biomarkers listed in Tables 1 to 63. (Item 40) 39. The method of claim 38, wherein said detecting a measurable feature comprises quantifying the amount of each of N biomarkers selected from the biomarkers listed in Tables 1 to 63, combinations or portions and / or derivatives thereof in a biological sample obtained from the pregnant female. (Item 41) 41. The method of claim 40, further comprising calculating a probability for preterm birth in the pregnant woman based on the quantified amount of each of N biomarkers selected from the biomarkers listed in Tables 1 to 63. (Item 42) The method of claim 38, further comprising the initial step of providing a biomarker panel comprising N of the biomarkers listed in Tables 1 to 63. (Item 43) 40. The method of claim 38, further comprising the initial step of providing a biological sample from the pregnant female. (Item 44) 40. The method of claim 38, further comprising communicating said probability to a health care provider. (Item 45) 45. The method of claim 44, wherein said communication informs subsequent treatment decisions for said pregnant woman. (Item 46) Item 39. The method of item 38, wherein N is a number selected from the group consisting of 2 to 24. (Item 47) 47. The method of claim 46, wherein the N biomarkers comprise at least two of the isolated biomarkers selected from the group consisting of AFTECCVVASQLR, ELLESYIDGR, ITLPDFTGDLR, the biomarkers shown in Table 50, and the biomarkers shown in Table 52. (Item 48) 40. The method of claim 38, wherein the analysis comprises the use of a predictive model. (Item 49) 49. The method of claim 48, wherein the analyzing comprises comparing the measurable feature to a reference feature. (Item 50) 50. The method of claim 49, wherein the analysis comprises using one or more selected from the group consisting of a linear discriminant analysis model, a support vector machine classification algorithm, a recursive feature elimination model, predictive analysis of microarray models, a logistic regression model, a multiple regression model, a survival model, a CART algorithm, a flextree algorithm, a LART algorithm, a random forest algorithm, a MART algorithm, a machine learning algorithm, a penalized regression method, and combinations thereof. (Item 51) 51. The method of claim 50, wherein the analysis comprises a random forest algorithm. (Item 52) 39. The method of claim 38, wherein the probability is expressed as a risk score. (Item 53) 39. The method of claim 38, wherein the biological sample is selected from the group consisting of whole blood, plasma, and serum. (Item 54) 54. The method of claim 53, wherein the biological sample is serum. (Item 55) 39. The method of claim 38, wherein said quantifying comprises mass spectrometry (MS). (Item 56) 56. The method of claim 55, wherein the MS comprises liquid chromatography mass spectrometry (LC-MS). (Item 57) 56. The method of claim 55, wherein the MS comprises multiple reaction monitoring (MRM) or selected reaction monitoring (SRM). (Item 58) 58. The method of claim 57, wherein the MRM (or SRM) comprises a scheduled MRM (SRM). (Item 59) 39. The method of claim 38, wherein the quantifying comprises an assay utilizing a capture agent. (Item 60) 60. The method of claim 59, wherein the capture agent is selected from the group consisting of an antibody, an antibody fragment, a nucleic acid-based protein binding reagent, a small molecule, or a variant thereof. (Item 61) 60. The method of claim 59, wherein the assay is selected from the group consisting of an enzyme immunoassay (EIA), an enzyme linked immunosorbent assay (ELISA), and a radioimmunoassay (RIA). (Item 62) 62. The method of claim 61, wherein the quantifying further comprises mass spectrometry (MS). (Item 63) 63. The method of claim 62, wherein the quantifying comprises co-immunoprecipitation-mass spectrometry (co-IP MS). (Item 64) 40. The method of claim 38, further comprising detecting a measurable characteristic for one or more risk indicators. (Item 65) 65. The method of item 64, wherein the one or more risk indicators are selected from the group consisting of age, previous pregnancy, previous history of low birth weight or early birth, multiple second trimester spontaneous abortions, previous first trimester induced abortions, familial and intergenerational factors, infertility history, nulliparity, placental abnormalities, cervical and uterine abnormalities, pregnancy bleeding, intrauterine growth restriction, intrauterine diethylstilbestrol exposure, multiple pregnancy, infant sex, short stature, low pre-pregnancy weight / low body mass index, diabetes, hypertension, hypothyroidism, asthma, education level, tobacco use, and genitourinary infections. (Item 66) 1. A method for predicting GAB, comprising: (a) quantifying the amount of each of N biomarkers selected from the biomarkers listed in Tables 1 to 63 in a biological sample obtained from a pregnant woman; (b) multiplying and / or thresholding the amounts by a predetermined coefficient; and (c) determining a predicted GAB birth in the pregnant woman, comprising adding the individual products to obtain a total risk score corresponding to the predicted GAB. (Item 67) 1. A method for predicting time to delivery in a pregnant woman, comprising: (a) obtaining a biological sample from the pregnant woman; (b) quantifying in the biological sample the amount of each of N biomarkers selected from the biomarkers listed in Tables 1 to 63; (c) multiplying and / or thresholding the amounts by a predetermined coefficient; (d) determining a predicted GAB in the pregnant woman, comprising adding the individual products to obtain a total risk score corresponding to a predicted GAB; and (e) subtracting the estimated GA at the time the biological sample was obtained from the predicted GAB to predict time to delivery in the pregnant woman. (Item 68) 1. A method of determining a probability for a full-term birth in a pregnant woman, comprising detecting a measurable feature of each of N biomarkers selected from the biomarkers listed in Tables 1 to 63 in a biological sample obtained from the pregnant woman, and analyzing the measurable features to determine the probability for a full-term birth in the pregnant woman. (Item 69) 3. The panel of claim 2, comprising at least two of the isolated biomarkers selected from the group consisting of FLNWIK, FGFGGSTDSGPIR, LLELTGPK, VEHSDLSFSK, IEGNLIFDPNYLPK, ALVLELAK, TQILEWAAER, DVLLLVHNLPQNLPGYFWYK, SEPRPGVLLR, ITQDAQLK, ALDLSLK, WWGGQPLWITATK, and LSETNR. (Item 70) The panel of claim 2, comprising alpha-1B-glycoprotein (A1BG), a disintegrin and metalloproteinase domain-containing protein 12 (ADA12), apolipoprotein B-100 (APOB), beta-2-microglobulin (B2MG), CCAAT / enhancer binding protein alpha / beta (HP8 peptide), corticosteroid binding globulin (CBG), complement component C6, endoglin (EGLN), ectonucleotide pyrophosphatase / phosphodiesterase family member 2 (ENPP2), coagulation factor VII (FA7), hyaluronan binding protein 2 (HABP2), pregnancy specific beta-1-glycoprotein 9 (PSG9), inhibin beta E chain (INHBE). (Item 71) 3. The panel of claim 2, comprising at least two isolated biomarkers selected from the group consisting of alpha-1B-glycoprotein (A1BG), a disintegrin and metalloproteinase domain-containing protein 12 (ADA12), apolipoprotein B-100 (APOB), beta-2-microglobulin (B2MG), CCAAT / enhancer binding protein alpha / beta (HP8 peptide), corticosteroid binding globulin (CBG), complement component C6, endoglin (EGLN), ectonucleotide pyrophosphatase / phosphodiesterase family member 2 (ENPP2), coagulation factor VII (FA7), hyaluronan binding protein 2 (HABP2), pregnancy specific beta-1-glycoprotein 9 (PSG9), and inhibin beta E chain (INHBE). (Item 72) 39. The method of claim 38, wherein the biomarkers comprise at least two of the isolated biomarkers selected from the group consisting of FLNWIK, FGFGGSTDSGPIR, LLELTGPK, VEHSDLSFSK, IEGNLIFDPNYLPK, ALVLELAK, TQILEWAAER, DVLLLVHNLPQNLPGYFWYK, SEPRPGVLLR, ITQDAQLK, ALDLSLK, WWGGQPLWITATK, and LSETNR. (Item 73) 67. The method of claim 66, wherein the biomarkers comprise at least two of the isolated biomarkers selected from the group consisting of FLNWIK, FGFGGSTDSGPIR, LLELTGPK, VEHSDLSFSK, IEGNLIFDPNYLPK, ALVLELAK, TQILEWAAER, DVLLLVHNLPQNLPGYFWYK, SEPRPGVLLR, ITQDAQLK, ALDLSLK, WWGGQPLWITATK, and LSETNR. (Item 74) 68. The method of claim 67, wherein the biomarkers comprise at least two of the isolated biomarkers selected from the group consisting of FLNWIK, FGFGGSTDSGPIR, LLELTGPK, VEHSDLSFSK, IEGNLIFDPNYLPK, ALVLELAK, TQILEWAAER, DVLLLVHNLPQNLPGYFWYK, SEPRPGVLLR, ITQDAQLK, ALDLSLK, WWGGQPLWITATK, and LSETNR. (Item 75) 69. The method of claim 68, wherein the biomarkers comprise at least two of the isolated biomarkers selected from the group consisting of FLNWIK, FGFGGSTDSGPIR, LLELTGPK, VEHSDLSFSK, IEGNLIFDPNYLPK, ALVLELAK, TQILEWAAER, DVLLLVHNLPQNLPGYFWYK, SEPRPGVLLR, ITQDAQLK, ALDLSLK, WWGGQPLWITATK, and LSETNR. (Item 76) 39. The method of claim 38, wherein the biomarkers comprise at least two of the isolated biomarkers selected from the group consisting of alpha-1B-glycoprotein (A1BG), a disintegrin and metalloproteinase domain-containing protein 12 (ADA12), apolipoprotein B-100 (APOB), beta-2-microglobulin (B2MG), CCAAT / enhancer binding protein alpha / beta (HP8 peptide), corticosteroid binding globulin (CBG), complement component C6, endoglin (EGLN), ectonucleotide pyrophosphatase / phosphodiesterase family member 2 (ENPP2), coagulation factor VII (FA7), hyaluronan binding protein 2 (HABP2), pregnancy specific beta-1-glycoprotein 9 (PSG9), and inhibin beta E chain (INHBE). (Item 77) 67. The method of claim 66, wherein the biomarkers comprise at least two of the isolated biomarkers selected from the group consisting of alpha-1B-glycoprotein (A1BG), a disintegrin and metalloproteinase domain-containing protein 12 (ADA12), apolipoprotein B-100 (APOB), beta-2-microglobulin (B2MG), CCAAT / enhancer binding protein alpha / beta (HP8 peptide), corticosteroid binding globulin (CBG), complement component C6, endoglin (EGLN), ectonucleotide pyrophosphatase / phosphodiesterase family member 2 (ENPP2), coagulation factor VII (FA7), hyaluronan binding protein 2 (HABP2), pregnancy specific beta-1-glycoprotein 9 (PSG9), and inhibin beta E chain (INHBE). (Item 78) 68. The method of claim 67, wherein the biomarkers comprise at least two of the isolated biomarkers selected from the group consisting of alpha-1B-glycoprotein (A1BG), a disintegrin and metalloproteinase domain-containing protein 12 (ADA12), apolipoprotein B-100 (APOB), beta-2-microglobulin (B2MG), CCAAT / enhancer binding protein alpha / beta (HP8 peptide), corticosteroid binding globulin (CBG), complement component C6, endoglin (EGLN), ectonucleotide pyrophosphatase / phosphodiesterase family member 2 (ENPP2), coagulation factor VII (FA7), hyaluronan binding protein 2 (HABP2), pregnancy specific beta-1-glycoprotein 9 (PSG9), and inhibin beta E chain (INHBE). (Item 79) 69. The method of claim 68, wherein the biomarkers comprise at least two of the isolated biomarkers selected from the group consisting of alpha-1B-glycoprotein (A1BG), a disintegrin and metalloproteinase domain-containing protein 12 (ADA12), apolipoprotein B-100 (APOB), beta-2-microglobulin (B2MG), CCAAT / enhancer binding protein alpha / beta (HP8 peptide), corticosteroid binding globulin (CBG), complement component C6, endoglin (EGLN), ectonucleotide pyrophosphatase / phosphodiesterase family member 2 (ENPP2), coagulation factor VII (FA7), hyaluronan binding protein 2 (HABP2), pregnancy specific beta-1-glycoprotein 9 (PSG9), and inhibin beta E chain (INHBE). [Brief description of the drawings]

[0034] [Figure 1] Figure 1 is a scatter plot of actual gestational age at birth versus predicted gestational age from the random forest regression model.

[0035] [Diagram 2]Figure 2 shows the distribution of predicted gestational age vs. actual gestational age at birth (GAB) from the random forest regression model, where actual GAB is given in categories: (i) less than 37 weeks, (ii) 37 to 39 weeks, and (iii) 40 weeks or greater (peaks from left to right, respectively). DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0036] Detailed Description The present disclosure is based in part on the discovery that certain proteins and peptides in biological samples obtained from pregnant women are differentially expressed in pregnant women with increased risk of preterm birth compared to controls.The present disclosure is further based in part on the unexpected discovery that a panel combining one or more of these proteins and peptides can be utilized in a method for determining the probability of preterm birth in pregnant women with high sensitivity and specificity.These proteins and peptides disclosed herein serve as biomarkers for classifying test samples, predicting the probability of preterm birth, predicting the probability of full-term birth, predicting gestational age at birth (GAB), predicting time to birth, and / or monitoring the progress of preventive treatment in pregnant women, either individually or in a panel of biomarkers.

[0037] The present disclosure provides a biomarker panel, method and kit for determining the probability of preterm birth in pregnant women.One major advantage of the present disclosure is that the risk of developing preterm birth can be assessed early in pregnancy, and appropriate monitoring and clinical management can be initiated accordingly to prevent preterm birth.The present invention is particularly beneficial for women who lack any risk factors for preterm birth and would otherwise not be identified and treated.

[0038] As an example, the disclosure includes a method for generating results useful in determining a probability for preterm birth in a pregnant woman by obtaining a dataset associated with a sample, where the dataset includes at least quantitative data for biomarkers and panels of biomarkers identified as predictors of preterm birth, and inputting the dataset into an analytical process using the dataset to generate results useful in determining a probability for preterm birth in a pregnant woman. As described further below, this quantitative data can include amino acids, peptides, polypeptides, proteins, nucleotides, nucleic acids, nucleosides, sugars, fatty acids, steroids, metabolites, carbohydrates, lipids, hormones, antibodies, regions of interest that serve as surrogates for biological macromolecules, and combinations thereof.

[0039] In addition to the specific biomarkers identified in this disclosure, for example by accession numbers in public databases, sequences, or references, the present invention also contemplates the use of biomarker variants now known or later discovered that are at least 90% or at least 95% or at least 97% identical to the exemplary sequences having utility for the methods of the present invention. These variants may represent polymorphisms, splice variants, mutations, and the like. In this regard, the present specification discloses a number of art-known proteins in the context of the present invention and provides exemplary accession numbers associated with one or more public databases as well as exemplary references to published journal articles related to these art-known proteins. However, one of skill in the art will readily identify additional accession numbers and journal articles that can provide additional characteristics of the disclosed biomarkers, and will understand that the exemplary references are in no way limiting with respect to the disclosed biomarkers. As described herein, a variety of techniques and reagents find use in the methods of the present invention. Suitable samples in the context of the present invention include, for example, blood, plasma, serum, amniotic fluid, vaginal secretions, saliva, and urine. In some embodiments, the biological sample is selected from the group consisting of whole blood, plasma, and serum.In certain embodiments, the biological sample is serum.As described herein, biomarkers can be detected through various assays and techniques known in the art.As further described herein, such assays include, but are not limited to, mass spectrometry (MS)-based assays, antibody-based assays, and assays that combine aspects of both.

[0040] Protein biomarkers associated with the probability of preterm birth in pregnant women include, but are not limited to, one or more of the isolated biomarkers listed in Tables 1 to 63. In addition to specific biomarkers, the present disclosure further includes biomarker variants that are about 90%, about 95%, or about 97% identical to the exemplary sequences. Variants, as used herein, include polymorphisms, splice variants, mutations, and the like.

[0041] The additional markers may be selected from one or more risk indicators, including, but not limited to, maternal characteristics, medical history, past pregnancy history, and obstetric history. Such additional markers may include, for example, previous low birth weight or early birth, multiple second trimester spontaneous abortions, previous first trimester induced abortions, familial and intergenerational factors, infertility history, nulliparity, placental abnormalities, cervical and uterine abnormalities, short cervical length measurements, pregnancy bleeding, intrauterine growth restriction, intrauterine diethylstilbestrol exposure, multiple pregnancy, infant sex, short stature, low pre-pregnancy weight, low or high body mass index, diabetes, hypertension, genitourinary infections (i.e., urinary tract infections), asthma, anxiety and depression, asthma, hypertension, hypothyroidism. Demographic risk indicators for preterm birth may include, for example, maternal age, race / ethnicity, single marital status, low socioeconomic status, maternal age, employment-related physical activity, occupational and environmental exposures, and stress. Additional risk indicators may include inadequate prenatal care, smoking, marijuana and other illicit drug use, cocaine use, alcohol consumption, caffeine intake, maternal weight gain, dietary intake, sexual activity during the third trimester, and leisure-time physical activity (Preterm (In: Premature Birth: Causes, Consequences, and Prevention, Institute of Medicine (US) Committee on Understanding Premature Birth and Assuring Healthy Outcomes; Behrman RE, Butler AS eds. Washington (DC): National Academies Press (US); 2007). Additional risk signatures useful as markers can be identified using learning algorithms known in the art, such as linear discriminant analysis, support vector machine classification, recursive feature elimination, predictive analysis of microarrays, logistic regression, CART, FlexTree, LART, random forest, MART, and / or survival analysis regression, which are known to those of skill in the art and are further described herein.

[0042] Provided herein is a panel of isolated biomarkers comprising N of the biomarkers selected from the group listed in Tables 1 to 63. In the disclosed panel of biomarkers, N can be a number selected from the group consisting of 2 to 24. In the disclosed methods, the number of biomarkers detected and the levels of which are determined can be 1 or more than 1, for example, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 12, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25 or more. In certain embodiments, the number of biomarkers detected and the levels of which are determined can be 1 or more than 1, for example, 2, 3, 4, 5, 6, 7, 8, 9, 10 or more. The methods of the present disclosure are useful for determining the probability of preterm birth in pregnant women.

[0043] While some of the biomarkers listed in Tables 1 through 63 are only useful for determining the probability for preterm birth in pregnant women, methods are also described herein for grouping multiple subsets of biomarkers, each of which is useful as a panel of three or more biomarkers. In some embodiments, the invention provides a panel comprising N biomarkers, where N is at least three biomarkers. In other embodiments, N is selected to be any number from 3 to 23 biomarkers.

[0044] In yet another embodiment, N is selected to be any number from 2 to 5, 2 to 10, 2 to 15, 2 to 20, or 2 to 23. In another embodiment, N is selected to be any number from 3 to 5, 3 to 10, 3 to 15, 3 to 20, or 3 to 23. In another embodiment, N is selected to be any number from 4 to 5, 4 to 10, 4 to 15, 4 to 20, or 4 to 23. In another embodiment, N is selected to be any number from 5 to 10, 5 to 15, 5 to 20, or 5 to 23. In another embodiment, N is selected to be any number from 6 to 10, 6 to 15, 6 to 20, or 6 to 23. In another embodiment, N is selected to be any number from 7 to 10, 7 to 15, 7 to 20, or 7 to 23. In another embodiment, N is selected to be any number from 8 to 10, 8 to 15, 8 to 20, or 8 to 23. In other embodiments, N is selected to be any number from 9 to 10, 9 to 15, 9 to 20, or 9 to 23. In other embodiments, N is selected to be any number from 10 to 15, 10 to 20, or 10 to 23. It will be appreciated that N can be selected to encompass similar but higher order ranges.

[0045] In certain embodiments, the panel of isolated biomarkers comprises one or more, two or more, three or more, four or more, or five isolated biomarkers comprising an amino acid sequence selected from AFTECCVVASQLR, ELLESYIDGR, ITLPDFTGDLR, TDAPDLPEENQAR, and SFRPFVPR. In some embodiments, the panel of isolated biomarkers comprises one or more, two or more, three or more, four or more, or five isolated biomarkers comprising an amino acid sequence selected from FLNWIK, FGFGGSTDSGPIR, LLELTGPK, VEHSDLSFSK, IEGNLIFDPNNYLPK, ALVLELAK, TQILEWAAER, DVLLLVHNLPQNLPGYFWYK, SEPRPGVLLR, ITQDAQLK, ALDLSLK, WWGGQPLWITATK, and LSETNR.

[0046] In some embodiments, the panel of isolated biomarkers comprises one or more, two or more, or three isolated biomarkers consisting of an amino acid sequence selected from AFTECCVVASQLR, ELLESYIDGR, and ITLPDFTGDLR. In some embodiments, the panel of isolated biomarkers comprises one or more, two or more, or three isolated biomarkers consisting of an amino acid sequence selected from FLNWIK, FGFGGSTDSGPIR, LLELTGPK, VEHSDLSFSK, IEGNLIFDPNNYLPK, ALVLELAK, TQILEWAAER, DVLLLVHNLPQNLPGYFWYK, SEPRPGVLLR, ITQDAQLK, ALDLSLK, WWGGQPLWITATK, and LSETNR.

[0047] In some embodiments, the panel of isolated biomarkers includes one or more, two or more, or three isolated biomarkers consisting of amino acid sequences selected from the biomarkers shown in Table 50 and the biomarkers shown in Table 52.

[0048] In some embodiments, the panel of isolated biomarkers includes lipopolysaccharide-binding protein (LBP), Schumann et al., Science 249(4975), 1429-1431 (1990) (UniProtKB / Swiss-Prot:P18428.3); prothrombin (THRB), Walz et al., Proc. Natl. Acad. Sci. USA 74(5), 1969-1972 (1977) (NCBI Reference Sequence:NP_000497.1); complement component C5 (C5 or CO5) Haviland, J. Immunol. 146(1), 362-368 (1991) (GenBank:AAA51925.1); plasminogen (PLMN) Petersen et al., J. Biol. Chem. 265(11), 6104-6111 (1990) (NCBI Reference Sequences: NP_000292.1 NP_001161810.1); and one or more peptides comprising a fragment from complement component C8 gamma chain (C8G or CO8G), Haefliger et al., Mol. Immunol. 28(1-2), 123-131 (1991) (NCBI Reference Sequence: NP_000597.2).

[0049] In some embodiments, the panel of isolated biomarkers includes a cell adhesion molecule with affinity to complement component 1, q subcomponent B chain (C1QB), Reid, Biochem. J. 179(2):367-371 (1979) (NCBI Reference Fibrinogen beta chain (FIBB or FIB); Watt et al., Biochemistry 18(1), 68-76 (1979) (NCBI Reference Sequences: NP_001171670.1 and NP_005132.2); C-reactive protein (CRP), Oliveira et al., J. Biol. Chem. 254(2), 489-502 (1979) (NCBI Reference Sequence: NP_000558.2); Inter-alpha-trypsin inhibitor heavy chain H4 (ITIH4), Kim et al., Mol. Biosyst. 7(5), 1430-1440 (2011) (NCBI Reference Sequences: NP_001171670.1 and NP_005132.2). Sequences: NP_001159921.1 and NP_002209.2); chorionic somatomammotropic hormone (CSH) Selby et al., J. Biol. Chem. 259(21):13131-13138 (1984) (NCBI Reference Sequence: NP_001308.1); and angiotensinogen (ANG or ANGT) Underwood et al., Metabolism 60(8):1150-7 (2011) (NCBI Reference Sequence: NP_000020.1).

[0050] In additional embodiments, the invention provides a panel of isolated biomarkers comprising N of the biomarkers listed in Tables 1-63. In some embodiments, N is a number selected from the group consisting of 2 to 24. In additional embodiments, the biomarker panel comprises at least two of the isolated biomarkers selected from the group consisting of AFTECCVVASQLR, ELLESYIDGR, and ITLPDFTGDLR. In additional embodiments, the biomarker panel comprises at least two of the isolated biomarkers selected from the group consisting of AFTECCVVASQLR, ELLESYIDGR, ITLPDFTGDLR, TDAPDLPEENQAR, and SFRPFVPR. In additional embodiments, the biomarker panel comprises at least two of the isolated biomarkers selected from the group consisting of FLNWIK, FGFGGSTDSGPIR, LLELTGPK, VEHSDLSFSK, IEGNLIFDPNYLPK, ALVLELAK, TQILEWAAER, DVLLLVHNLPQNLPGYFWYK, SEPRPGVLLR, ITQDAQLK, ALDLSLK, WWGGQPLWITATK, and LSETNR.

[0051] In an additional embodiment, the biomarker panel comprises at least two of the isolated biomarkers selected from the group consisting of the biomarkers shown in Table 50 and the biomarkers shown in Table 52.

[0052] In a further embodiment, the biomarker panel comprises at least two isolated biomarkers selected from the group consisting of lipopolysaccharide binding protein (LBP), prothrombin (THRB), complement component C5 (C5 or CO5), plasminogen (PLMN), and complement component C8 gamma chain (C8G or CO8G). In another embodiment, the invention provides a biomarker panel comprising at least three isolated biomarkers selected from the group consisting of lipopolysaccharide binding protein (LBP), prothrombin (THRB), complement component C5 (C5 or CO5), plasminogen (PLMN), and complement component C8 gamma chain (C8G or CO8G).

[0053] In a further embodiment, the biomarker panel comprises at least two of the isolated biomarkers selected from the group consisting of alpha-1B-glycoprotein (A1BG), a disintegrin and metalloproteinase domain-containing protein 12 (ADA12), apolipoprotein B-100 (APOB), beta-2-microglobulin (B2MG), CCAAT / enhancer binding protein alpha / beta (HP8 peptide), corticosteroid binding globulin (CBG), complement component C6, endoglin (EGLN), ectonucleotide pyrophosphatase / phosphodiesterase family member 2 (ENPP2), coagulation factor VII (FA7), hyaluronan binding protein 2 (HABP2), pregnancy specific beta-1-glycoprotein 9 (PSG9), inhibin beta E chain (INHBE).

[0054] In some embodiments, the present invention provides a biomarker panel that includes lipopolysaccharide binding protein (LBP), prothrombin (THRB), complement component C5 (C5 or CO5), plasminogen (PLMN), complement component C8 gamma chain (C8G or CO8G), complement component 1, q subcomponent, B chain (C1QB), fibrinogen beta chain (FIBB or FIB), C-reactive protein (CRP), inter-alpha-trypsin inhibitor heavy chain H4 (ITIH4), chorionic somatomammotropic hormone (CSH), and angiotensinogen (ANG or ANGT). In some embodiments, the present invention provides a biomarker panel comprising alpha-1B-glycoprotein (A1BG), a disintegrin and metalloproteinase domain-containing protein 12 (ADA12), apolipoprotein B-100 (APOB), beta-2-microglobulin (B2MG), CCAAT / enhancer binding protein alpha / beta (HP8 peptide), corticosteroid binding globulin (CBG), complement component C6, endoglin (EGLN), ectonucleotide pyrophosphatase / phosphodiesterase family member 2 (ENPP2), coagulation factor VII (FA7), hyaluronan binding protein 2 (HABP2), pregnancy specific beta-1-glycoprotein 9 (PSG9), inhibin beta E chain (INHBE).

[0055] In another embodiment, the invention provides a biomarker panel comprising at least two isolated biomarkers selected from the group consisting of lipopolysaccharide binding protein (LBP), prothrombin (THRB), complement component C5 (C5 or CO5), plasminogen (PLMN), complement component C8 gamma chain (C8G or CO8G), complement component 1, q subcomponent, B chain (C1QB), fibrinogen beta chain (FIBB or FIB), C-reactive protein (CRP), inter-alpha-trypsin inhibitor heavy chain H4 (ITIH4), chorionic somatomammotropic hormone (CSH), and angiotensinogen (ANG or ANGT) and the biomarkers shown in Tables 51 and 53.

[0056] In another aspect, the present invention provides a biomarker panel comprising at least two isolated biomarkers selected from the group consisting of alpha-1B-glycoprotein (A1BG), a disintegrin and metalloproteinase domain-containing protein 12 (ADA12), apolipoprotein B-100 (APOB), beta-2-microglobulin (B2MG), CCAAT / enhancer binding protein alpha / beta (HP8 peptide), corticosteroid binding globulin (CBG), complement component C6, endoglin (EGLN), ectonucleotide pyrophosphatase / phosphodiesterase family member 2 (ENPP2), coagulation factor VII (FA7), hyaluronan binding protein 2 (HABP2), pregnancy specific beta-1-glycoprotein 9 (PSG9), inhibin beta E chain (INHBE).

[0057] It must be noted that as used in this specification and the appended claims, the singular forms "a," "an," and "the" include plural referents unless the content clearly dictates otherwise. Thus, for example, reference to a "biomarker" includes a mixture of two or more biomarkers, and the like.

[0058] The term "about," particularly with reference to a given amount, is meant to encompass deviations of plus or minus 5 percent.

[0059] As used in this application, including the appended claims, the singular forms "a," "an," and "the" are inclusive of plural referents and are used interchangeably with "at least one" and "one or more," unless the content clearly dictates otherwise.

[0060] As used herein, the terms "comprises," "comprising," "includes," "including," "contains," "containing" and any variations thereof are intended to cover a non-exclusive inclusion, and a process, method, product-by-process, or composition that comprises, includes, or contains an element or list of elements does not include only those elements, but may include other elements not expressly listed in or inherent to such process, method, product-by-process, or composition.

[0061] As used herein, the term "panel" refers to a composition, such as an array or collection, that includes one or more biomarkers. The term may also refer to a profile or index of the expression pattern of one or more biomarkers described herein. The number of biomarkers useful for a biomarker panel is based on the sensitivity and specificity values ​​for a particular combination of biomarker values.

[0062] As used herein, and unless otherwise specified, the terms "isolated" and "purified" generally refer to a composition that has been removed from its native environment (e.g., the natural environment if it occurs in nature) and thus has been altered by the hand of man from its native state. An isolated protein or nucleic acid is different from the form in which it occurs in nature.

[0063] The term "biomarker" refers to a biological molecule, or a fragment of a biological molecule, the change and / or detection of which may correlate with a particular physical condition or state. The terms "marker" and "biomarker" are used interchangeably throughout this disclosure. For example, the biomarkers of the present invention correlate with an increased likelihood of preterm birth. Such biomarkers include, but are not limited to, biological molecules that include nucleotides, nucleic acids, nucleosides, amino acids, sugars, fatty acids, steroids, metabolites, peptides, polypeptides, proteins, carbohydrates, lipids, hormones, antibodies, regions of interest that serve as surrogates for biological macromolecules, and combinations thereof (e.g., glycoproteins, ribonucleoproteins, lipoproteins). The term also encompasses portions or fragments of biological molecules, e.g., peptide fragments of proteins or polypeptides comprising at least 5 contiguous amino acid residues, at least 6 contiguous amino acid residues, at least 7 contiguous amino acid residues, at least 8 contiguous amino acid residues, at least 9 contiguous amino acid residues, at least 10 contiguous amino acid residues, at least 11 contiguous amino acid residues, at least 12 contiguous amino acid residues, at least 13 contiguous amino acid residues, at least 14 contiguous amino acid residues, at least 15 contiguous amino acid residues, at least 5 contiguous amino acid residues, at least 16 contiguous amino acid residues, at least 17 contiguous amino acid residues, at least 18 contiguous amino acid residues, at least 19 contiguous amino acid residues, at least 20 contiguous amino acid residues, at least 21 contiguous amino acid residues, at least 22 contiguous amino acid residues, at least 23 contiguous amino acid residues, at least 24 contiguous amino acid residues, at least 25 contiguous amino acid residues, or more contiguous amino acid residues.

[0064] The present invention also provides a method of determining a probability for preterm birth in a pregnant woman, comprising detecting a measurable feature of each of N biomarkers selected from the biomarkers listed in Tables 1 to 63 in a biological sample obtained from the pregnant woman, and analyzing the measurable feature to determine a probability for preterm birth in the pregnant woman. As disclosed herein, the measurable feature comprises a fragment or derivative of each of the N biomarkers selected from the biomarkers listed in Tables 1 to 63. In some embodiments of the disclosed methods, detecting the measurable feature comprises quantifying the amount of each of the N biomarkers selected from the biomarkers listed in Tables 1 to 63, combinations or portions and / or derivatives thereof in the biological sample obtained from the pregnant woman.

[0065] The present invention further provides a method of predicting GAB, comprising detecting a measurable feature of each of N biomarkers selected from the biomarkers listed in Tables 1 to 63 in a biological sample obtained from a pregnant woman, and analyzing the measurable features to predict GAB.

[0066] The present invention also provides a method of predicting GAB comprising: (a) quantifying the amount of each of N biomarkers selected from the biomarkers listed in Tables 1 to 63 in a biological sample obtained from a pregnant woman; (b) multiplying or thresholding the amounts by a predetermined factor; and (c) determining a predicted GAB birth in a pregnant woman comprising adding the individual products to obtain a total risk score corresponding to the predicted GAB.

[0067] The present invention further provides a method for predicting time to birth in a pregnant woman, comprising: (a) obtaining a biological sample from the pregnant woman; (b) quantifying the amount of each of N biomarkers selected from the biomarkers listed in Tables 1 to 63 in the biological sample; (c) multiplying or thresholding the amount by a predetermined coefficient; (d) determining a predicted GAB in the pregnant woman, including adding the individual products to obtain a total risk score corresponding to the predicted GAB; and (e) subtracting the estimated gestational age (GA) at the time the biological sample was obtained from the predicted GAB to predict time to birth in the pregnant woman. For the method directed to predicting time to birth, "birth" is understood to mean birth following the spontaneous onset of labor, with or without rupture of membranes.

[0068] Although described and illustrated with reference to the method for determining the probability of premature birth in pregnant women, the present disclosure is equally applicable to the method for predicting GAB, the method for predicting full-term birth, the method for determining the probability of full-term birth in pregnant women, and the method for predicting the time to birth in pregnant women.It will be clear to those skilled in the art that each of the above methods has a specific and substantial utility and benefit in terms of maternal-fetal health considerations.

[0069] In some embodiments, the method of determining the probability for preterm birth in a pregnant woman and related methods disclosed herein comprises detecting a measurable feature for each of N biomarkers, where N is selected from the group consisting of 2 to 24. In further embodiments, the disclosed method of determining the probability for preterm birth in a pregnant woman and related methods disclosed herein comprises detecting a measurable feature for each of at least two isolated biomarkers selected from the group consisting of AFTECCVVASQLR, ELLESYIDGR, and ITLPDFTGDLR. In further embodiments, the disclosed method of determining the probability for preterm birth in a pregnant woman and related methods disclosed herein comprises detecting a measurable feature for each of at least two isolated biomarkers selected from the group consisting of FLNWIK, FGFGGSTDSGPIR, LLELTGPK, VEHSDLSFSK, IEGNLIFDPNYLPK, ALVLELAK, TQILEWAAER, DVLLLVHNLPQNLPGYFWYK, SEPRPGVLLR, ITQDAQLK, ALDLSLK, WWGGQPLWITATK, and LSETNR.

[0070] In additional embodiments, the disclosed method of determining the probability of preterm birth in a pregnant woman and related methods disclosed herein comprises detecting a measurable feature of each of at least two isolated biomarkers selected from the group consisting of the biomarkers shown in Table 50 and the biomarkers shown in Table 52.

[0071] In additional embodiments, the method of determining the probability of preterm birth in a pregnant woman and related methods disclosed herein comprises detecting a measurable characteristic of each of at least two isolated biomarkers selected from the group consisting of lipopolysaccharide binding protein (LBP), prothrombin (THRB), complement component C5 (C5 or CO5), plasminogen (PLMN), and complement component C8 gamma chain (C8G or CO8G).

[0072] In additional embodiments, the method of determining the probability for preterm birth in a pregnant woman and related methods disclosed herein comprises detecting a measurable characteristic of each of at least two isolated biomarkers selected from the group consisting of alpha-1B-glycoprotein (A1BG), a disintegrin and metalloproteinase domain-containing protein 12 (ADA12), apolipoprotein B-100 (APOB), beta-2-microglobulin (B2MG), CCAAT / enhancer binding protein alpha / beta (HP8 peptide), corticosteroid binding globulin (CBG), complement component C6, endoglin (EGLN), ectonucleotide pyrophosphatase / phosphodiesterase family member 2 (ENPP2), coagulation factor VII (FA7), hyaluronan binding protein 2 (HABP2), pregnancy specific beta-1-glycoprotein 9 (PSG9), inhibin beta E chain (INHBE).

[0073] In a further embodiment, the disclosed method of determining the probability of preterm birth in a pregnant woman and related methods disclosed herein comprises detecting a measurable characteristic of each of at least two isolated biomarkers selected from the group consisting of lipopolysaccharide binding protein (LBP), prothrombin (THRB), complement component C5 (C5 or CO5), plasminogen (PLMN), complement component C8 gamma chain (C8G or CO8G), complement component 1, q subcomponent, B chain (C1QB), fibrinogen beta chain (FIBB or FIB), C-reactive protein (CRP), inter-alpha-trypsin inhibitor heavy chain H4 (ITIH4), chorionic somatomammotropic hormone (CSH), and angiotensinogen (ANG or ANGT).

[0074] In a further embodiment, the disclosed method of determining the probability for preterm birth in a pregnant woman and related methods disclosed herein comprises detecting a measurable characteristic of each of at least two isolated biomarkers selected from the group consisting of alpha-1B-glycoprotein (A1BG), a disintegrin and metalloproteinase domain-containing protein 12 (ADA12), apolipoprotein B-100 (APOB), beta-2-microglobulin (B2MG), CCAAT / enhancer binding protein alpha / beta (HP8 peptide), corticosteroid binding globulin (CBG), complement component C6, endoglin (EGLN), ectonucleotide pyrophosphatase / phosphodiesterase family member 2 (ENPP2), coagulation factor VII (FA7), hyaluronan binding protein 2 (HABP2), pregnancy specific beta-1-glycoprotein 9 (PSG9), inhibin beta E chain (INHBE).

[0075] In a further embodiment, the disclosed method of determining the probability for preterm birth in a pregnant woman and related methods disclosed herein comprises detecting a measurable characteristic of each of at least two isolated biomarkers selected from the group consisting of alpha-1B-glycoprotein (A1BG), a disintegrin and metalloproteinase domain-containing protein 12 (ADA12), apolipoprotein B-100 (APOB), beta-2-microglobulin (B2MG), CCAAT / enhancer binding protein alpha / beta (HP8 peptide), corticosteroid binding globulin (CBG), complement component C6, endoglin (EGLN), ectonucleotide pyrophosphatase / phosphodiesterase family member 2 (ENPP2), coagulation factor VII (FA7), hyaluronan binding protein 2 (HABP2), pregnancy specific beta-1-glycoprotein 9 (PSG9), inhibin beta E chain (INHBE).

[0076] In a further embodiment, the disclosed method of determining the probability of preterm birth in a pregnant woman and related methods disclosed herein comprises detecting a measurable feature of each of at least two isolated biomarkers selected from the group consisting of the biomarkers shown in Table 51 and the biomarkers shown in Table 53.

[0077] In additional embodiments, the method of determining the probability for preterm birth in a pregnant woman further comprises detecting a measurable feature for one or more risk indicators associated with preterm birth. In additional embodiments, the risk indicators are selected from the group consisting of previous low birth weight or early birth, multiple second trimester spontaneous abortions, previous first trimester induced abortions, familial and intergenerational factors, infertility history, nulliparity, placental abnormalities, cervical and uterine abnormalities, pregnancy bleeding, intrauterine growth restriction, intrauterine diethylstilbestrol exposure, multiple gestations, infant sex, short stature, low pre-pregnancy weight, low or high body mass index, diabetes, hypertension, and genitourinary infections.

[0078] "Measurable characteristic" is any characteristic, feature, or aspect that can be determined and correlated with the probability of preterm birth in a subject. This term further encompasses any characteristic, feature, or aspect that can be determined and correlated in the context of predicting GAB in pregnant women, predicting full-term birth, or predicting time to birth. For biomarkers, such measurable characteristics can include, for example, the presence, absence, or concentration of the biomarker or a fragment thereof in a biological sample, altered structure, such as the presence or amount of post-translational modification, such as oxidation at one or more positions on the amino acid sequence of the biomarker, or the presence of an altered conformation, such as compared to the conformation of the biomarker in a normal control subject, and / or the presence, amount, or altered structure of the biomarker as part of a profile of more than one biomarker. In addition to biomarkers, measurable characteristics can also include risk indicators, including, for example, maternal characteristics, age, race, ethnicity, medical history, past pregnancy history, obstetric history. For risk indicators, measurable characteristics may include, for example, previous low birth weight or early birth, multiple second trimester spontaneous abortions, previous first trimester induced abortions, familial and intergenerational factors, infertility history, nulliparity, placental abnormalities, cervical and uterine abnormalities, short cervical length measurement, pregnancy bleeding, intrauterine growth restriction, intrauterine diethylstilbestrol exposure, multiple pregnancy, infant sex, short stature, low pre-pregnancy weight / low body mass index, diabetes, hypertension, genitourinary infections, hypothyroidism, asthma, low academic achievement, smoking, drug use, and alcohol consumption.

[0079] In some embodiments of the disclosed methods of determining the probability of preterm birth in a pregnant woman, the probability of preterm birth in a pregnant woman is calculated based on the quantified amount of each of N biomarkers selected from the biomarkers listed in Tables 1 to 63. In some embodiments, the disclosed methods for determining the probability of preterm birth include detecting and / or quantifying one or more biomarkers using mass spectrometry, a capture agent, or a combination thereof.

[0080] In some embodiments, the disclosed methods of determining the probability of preterm birth in a pregnant woman include an initial step of providing a biomarker panel that includes N of the biomarkers listed in Tables 1 to 63. In additional embodiments, the disclosed methods of determining the probability of preterm birth in a pregnant woman include an initial step of providing a biological sample from the pregnant woman.

[0081] In some embodiments, the disclosed method of determining the probability of preterm birth in a pregnant woman includes communicating the probability to a health care provider. The disclosed method of predicting GAB, predicting full-term birth, determining the probability of full-term birth in a pregnant woman, and predicting the time to birth in a pregnant woman also includes communicating the probability to a health care provider. Although described and illustrated above with reference to determining the probability of preterm birth in a pregnant woman, all embodiments described throughout this disclosure are equally applicable to the methods of predicting GAB, predicting full-term birth, determining the probability of full-term birth in a pregnant woman, and predicting the time to birth in a pregnant woman. In particular, the biomarkers and panels listed throughout this application that explicitly refer to the methods for preterm birth can also be used in the methods of predicting GAB, predicting full-term birth, determining the probability of full-term birth in a pregnant woman, and predicting the time to birth in a pregnant woman. It will be apparent to one skilled in the art that each of the above methods has specific and substantial utility and benefit with respect to maternal-fetal health considerations.

[0082] In additional embodiments, the communication informs subsequent treatment decisions for the pregnant woman. In some embodiments, the methods of determining the probability for preterm birth in a pregnant woman include an additional feature that expresses the probability as a risk score.

[0083] As used herein, the term "risk score" refers to a score that can be assigned based on comparing the amount of one or more biomarkers in a biological sample obtained from a pregnant woman with a standard score or reference score that represents the average amount of one or more biomarkers calculated from biological samples obtained from a random pool of pregnant women. Because the level of a biomarker may not be static throughout pregnancy, the standard score or reference score must be obtained for a pregnancy time point that corresponds to that of the pregnant woman when the sample is taken. The standard score or reference score can be pre-determined and incorporated into the predictive model, so that the comparison is indirect, rather than being performed every time a probability is determined for a subject. The risk score can be a standard (e.g., a number) or a threshold (e.g., a line on a graph). The value of the risk score correlates with an upward or downward deviation from the average amount of one or more biomarkers calculated from biological samples obtained from a random pool of pregnant women. In certain embodiments, if the risk score is greater than the standard risk score or reference risk score, the pregnant woman may have an increased likelihood of preterm birth. In some embodiments, the magnitude of a pregnant woman's risk score, or the amount by which it exceeds a baseline risk score, may be indicative of or correlate with the level of risk for that pregnant woman.

[0084] In the context of the present invention, the term "biological sample" encompasses any sample taken from a pregnant woman and includes one or more of the biomarkers listed in Tables 1 to 63. Suitable samples in the context of the present invention include, for example, blood, plasma, serum, amniotic fluid, vaginal secretions, saliva, and urine. In some embodiments, the biological sample is selected from the group consisting of whole blood, plasma, and serum. In certain embodiments, the biological sample is serum. As will be understood by those skilled in the art, the biological sample may include any fraction or component of blood, including, but not limited to, T cells, monocytes, neutrophils, red blood cells, platelets, and microvesicles, such as exosomes and exosome-like vesicles. In certain embodiments, the biological sample is serum.

[0085] Preterm birth refers to birth or delivery at less than 37 completed weeks of gestation. Other commonly used subcategories of preterm birth have been established, delineating moderately preterm birth (birth at 33-36 weeks of gestation), very preterm birth (birth at <33 weeks of gestation), and extremely preterm birth (birth at ≦28 weeks of gestation). With respect to the methods disclosed herein, one of skill in the art will appreciate that the cutoffs delineating preterm and full-term birth, as well as the cutoffs delineating subcategories of preterm birth, can be adjusted in practicing the methods disclosed herein, for example, to maximize certain health benefits. It is further understood that such adjustments are well within the skill of the artisan and are encompassed within the scope of the invention disclosed herein. Gestational age is a surrogate for the degree of fetal development and fetal readiness for birth. Gestational age is typically defined as the length of time from the date of the last normal menstrual period to the date of delivery. However, obstetric measures and ultrasound estimates can also aid in estimating gestational age. Preterm birth is generally classified into two separate subgroups. One is that spontaneous preterm birth occurs following the spontaneous onset of preterm labor or preterm rupture of membranes, regardless of subsequent labor augmentation or Caesarean section.Two is that the indicated preterm birth occurs following induction or Caesarean section for one or more conditions that female caregivers determine to be threatening to the health or life of the mother and / or fetus.In some embodiments, the method disclosed herein is directed to determining the probability of spontaneous preterm birth.In additional embodiments, the method disclosed herein is directed to predicting pregnancy delivery.

[0086] As used herein, the term "estimated gestational age" or "estimated GA" refers to the GA determined based on the date of the last normal menstrual period and additional obstetric measures, ultrasound estimates, or other clinical parameters, including but not limited to those described in the previous section. In contrast, the term "predicted gestational age at birth" or "predicted GAB" refers to the GAB determined based on the inventive method disclosed herein. As used herein, "term birth" refers to birth at or beyond 37 full weeks of gestational age.

[0087] In some embodiments, the pregnant female is between 17-28 weeks gestation at the time the biological sample is collected. In other embodiments, the pregnant female is between 16-29 weeks gestation, between 17-28 weeks gestation, between 18-27 weeks gestation, between 19-26 weeks gestation, between 20-25 weeks gestation, between 21-24 weeks gestation, or between 22-23 weeks gestation at the time the biological sample is collected. In further embodiments, the pregnant female is between about 17-22 weeks gestation, between about 16-22 weeks gestation, between about 22-25 weeks gestation, between about 13-25 weeks gestation, between about 26-28 weeks gestation, or between about 26-29 weeks gestation at the time the biological sample is collected. Thus, the gestational age of the pregnant female at the time the biological sample is collected can be 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, or 30 weeks.

[0088] In some embodiments of the claimed methods, the measurable feature comprises a fragment or derivative of each of the N biomarkers selected from the biomarkers listed in Tables 1 to 63. In additional embodiments of the claimed methods, detecting the measurable feature comprises quantifying the amount of each of the N biomarkers selected from the biomarkers listed in Tables 1 to 63, combinations or portions and / or derivatives thereof in a biological sample obtained from said pregnant female.

[0089] The term "amount" or "level" as used herein refers to the amount of a biomarker that is detectable or measurable in a biological sample and / or control. The amount of a biomarker can be, for example, the amount of a polypeptide, the amount of a nucleic acid, or the amount of a fragment or surrogate. The term can alternatively include combinations thereof. The term "amount" or "level" of a biomarker is a measurable characteristic of that biomarker.

[0090] In some embodiments, calculating the probability for preterm birth in a pregnant woman is based on the quantified amount of each of N biomarkers selected from the biomarkers listed in Tables 1 to 63. Any existing, available, or conventional separation, detection, and quantification method can be used herein to measure the presence or absence (e.g., readout is present vs. absent; or detectable vs. undetectable amount) and / or amount (e.g., readout is absolute amount or relative amount, such as absolute concentration or relative concentration) of a biomarker, peptide, polypeptide, protein, and / or fragment thereof, and optionally one or more other biomarkers or fragments thereof in a sample. In some embodiments, the detection and / or quantification of one or more biomarkers includes an assay that utilizes a capture agent. In further embodiments, the capture agent is an antibody, an antibody fragment, a nucleic acid-based protein binding reagent, a small molecule, or a variant thereof. In additional embodiments, the assay is an enzyme immunoassay (EIA), an enzyme-linked immunosorbent assay (ELISA), and a radioimmunoassay (RIA). In some embodiments, the detection and / or quantification of one or more biomarkers further comprises mass spectrometry (MS). In yet further embodiments, the mass spectrometry is co-immunoprecipitation-mass spectrometry (co-IP MS), in which co-immunoprecipitation, a technique suitable for the isolation of total protein complexes, is followed by mass spectrometry.

[0091] As used herein, the term "mass spectrometer" refers to a device that can volatilize / ionize analytes, form gas phase ions, and determine their absolute or relative molecular mass. Suitable methods of volatilization / ionization are matrix-assisted laser desorption / ionization (MALDI), electrospray, laser / light, thermal, electrical, atomization / nebulization, etc., or combinations thereof. Suitable forms of mass spectrometry include, but are not limited to, ion trap instruments, quadrupole instruments, electrostatic sector instruments, time-of-flight instruments, time-of-flight tandem mass spectrometers (TOF MS / MS), Fourier transform mass spectrometers, Orbitrap, and hybrid instruments composed of various combinations of these types of mass analyzers. These instruments can then be matched with various other instruments that fractionate the sample (e.g., liquid chromatography or solid phase adsorption techniques based on chemical or biological features) and ionize the sample for introduction into the mass spectrometer (including matrix-assisted laser desorption (MALDI), electrospray, or nanospray ionization (ESI), or combinations thereof).

[0092] Generally, any mass spectrometry (MS) technique (e.g., tandem mass spectrometry, MS / MS; or post-source decay, TOF MS) that can provide accurate information on peptide mass, and preferably also on the fragmentation and / or (partial) amino acid sequence of selected peptides, can be used in the methods disclosed herein. Suitable peptide MS and MS / MS techniques and systems are known per se (see, for example, Methods in Molecular Biology, Vol. 146: "Mass Spectrometry of Proteins and Peptides", Chapman, ed., Humana Press 2000; Biemann 1990, Methods Enzymol Vol. 193: pp. 455-79; or Methods in Enzymology, Vol. 402: "Biological Mass Spectrometry", Burlingame, ed., Academic Press 2005) and can be used in carrying out the methods disclosed herein. Thus, in some embodiments, the disclosed methods comprise carrying out quantitative MS to measure one or more biomarkers. Such quantitative methods can be performed in an automated (Villanueva et al., Nature Protocols (2006) vol. 1(2): 880-891) or semi-automated format. In certain embodiments, the MS can be operatively linked to a liquid chromatography device (LC-MS / MS or LC-MS), or a gas chromatography device (GC-MS or GC-MS / MS). Other methods useful in this context include isotope-coded affinity tags (ICAT), tandem mass tags (TMT), or stable isotope labeling with amino acids in cell culture (SILAC), followed by chromatography and MS / MS.

[0093] As used herein, the term "multiple reaction monitoring (MRM)" or "selected reaction monitoring (SRM)" refers to an MS-based quantification method that is particularly useful for quantifying analytes that are in low abundance. In an SRM experiment, a predefined precursor ion and one or more of its fragments are selected by two mass filters of a triple quadrupole instrument and monitored over time for accurate quantification. Multiple SRM precursor and fragment ion pairs can be measured within the same experiment on a chromatographic timescale by rapidly toggling between different precursor / fragment pairs to perform an MRM experiment. A series of transitions (precursor / fragment ion pairs) combined with the retention time of the target analyte (e.g., a peptide or small molecule, such as a chemical, steroid, hormone, etc.) can constitute the final assay. Multiple analytes can be quantified during a single LC-MS experiment. The term "schedule" or "dynamic" in reference to MRM or SRM refers to a variation of the assay where transitions for a particular analyte are acquired only in a time window around the expected retention time, significantly increasing the number of analytes that can be detected and quantified in a single LC-MS experiment and contributing to the selectivity of the test, since the retention time is a feature that depends on the physical properties of the analyte. A single analyte can also be monitored with more than one transition. Finally, a standard that corresponds to the analyte of interest (e.g., the same amino acid sequence) but differs due to the inclusion of a stable isotope can be included in the assay. A stable isotope standard (SIS) can be incorporated into the assay at a precise level and used to quantify the corresponding unknown analyte. An additional level of specificity is contributed by the co-elution of the unknown analyte and its corresponding SIS as well as the characteristics of their transitions (e.g., similarity in the ratio of the levels of two transitions of the unknown and the ratio of two transitions of its corresponding SIS).

[0094] Suitable mass spectrometry assays, instruments, and systems for biomarker peptide analysis include, but are not limited to, matrix-assisted laser desorption / ionization time-of-flight (MALDI-TOF) MS; MALDI-TOF post-source decay (PSD); MALDI-TOF / TOF; surface-enhanced laser desorption / ionization time-of-flight mass spectrometry (SELDI-TOF) MS; electrospray ionization mass spectrometry (ESI-MS); ESI-MS / MS; ESI-MS / (MS) n (n is an integer greater than zero);ESI 3D or linear (2D) ion trap MS;ESI triple quadrupole MS;ESI quadrupole orthogonal TOF (Q-TOF);ESI Fourier transform MS system;Desorption / ionization on silicon (DIOS);Secondary ion mass spectrometry (SIMS);Atmospheric pressure chemical ionization mass spectrometry (APCI-MS);APCI-MS / MS;APCI-(MS) n ;Ion mobility spectroscopy (IMS);Inductively coupled plasma mass spectrometry (ICP-MS);Atmospheric pressure photoionization mass spectrometry (APPI-MS);APPI-MS / MS, and APPI-(MS) nFragmentation of peptide ions in tandem MS (MS / MS) configurations can be accomplished using modalities established in the art, such as collision-induced dissociation (CID). As described herein, detection and quantification of biomarkers by mass spectrometry can include multiple reaction monitoring (MRM), as described, for example, by Kuhn et al., Proteomics 4:1175-86 (2004), among others. Scheduled multiple reaction monitoring (scheduled MRM) mode acquisition during LC-MS / MS analysis enhances the sensitivity and accuracy of peptide quantification. Anderson and Hunter, Molecular and Cellular Proteomics 5(4):573 (2006). As described herein, mass spectrometry-based assays can be advantageously combined with upstream peptide or protein separation or fractionation methods, such as chromatography and other methods described herein below. As further described herein, shotgun quantitative proteomics can be combined with SRM / MRM-based assays for high-throughput identification and validation of prognostic biomarkers for preterm birth.

[0095] One skilled in the art will appreciate that several methods can be used to determine the amount of a biomarker, including mass spectrometry approaches such as MS / MS, LC-MS / MS, multiple reaction monitoring (MRM), or SRM, and product ion monitoring (PIM), as well as antibody-based methods such as immunoassays, such as Western blots, enzyme-linked immunosorbent assays (ELISAs), immunoprecipitation, immunohistochemistry, immunofluorescence, radioimmunoassays, dot blotting, and FACS. Thus, in some embodiments, determining the level of at least one biomarker comprises using an immunoassay and / or mass spectrometry. In additional embodiments, the mass spectrometry method is selected from MS, MS / MS, LC-MS / MS, SRM, PIM, and other such methods known in the art. In other embodiments, the LC-MS / MS further comprises 1D LC-MS / MS, 2D LC-MS / MS, or 3D LC-MS / MS. Immunoassay techniques and protocols are generally known to those of skill in the art (Price and Newman, Principles and Practice of Immunoassay, 2nd ed., Grove's Dictionaries, 1997, and Gosling, Immunoassays: A Practical Approach, Oxford University Press, 2000). A variety of immunoassay techniques can be used, including competitive and noncompetitive immunoassays (Self et al., Curr. Opin. Biotechnol., 7:60-65 (1996)).

[0096] In further embodiments, the immunoassay is selected from Western blot, ELISA, immunoprecipitation, immunohistochemistry, immunofluorescence, radioimmunoassay (RIA), dot blotting, and FACS. In certain embodiments, the immunoassay is ELISA. In yet further embodiments, the ELISA is direct ELISA (enzyme-linked immunosorbent assay), indirect ELISA, sandwich ELISA, competitive ELISA, multiplex ELISA, ELISPOT technology, and other similar techniques known in the art. The principles of these immunoassay methods are known in the art (e.g., John R. Crowther, The ELISA Guidebook, 1st Edition, Humana Press 2000, ISBN 0896037282). Typically, ELISAs are performed with antibodies, but they can be performed with any capture agent that can specifically bind and detect one or more biomarkers of the present invention. Multiplexed ELISAs allow for the simultaneous detection of two or more analytes in a single compartment (e.g., a microplate well), usually at multiple array addresses (Nielsen and Geierstanger 2004, J Immunol Methods 290:107-20 (2004) and Ling et al., 2007, Expert Rev Mol Diagn 7:87-98 (2007)).

[0097] In some embodiments, radioimmunoassays (RIAs) can be used to detect one or more biomarkers in the methods of the invention. RIAs are competition-based assays that are well known in the art and use radioactive labels (e.g., 125 I or 131Radioimmunoassay (RI-labeling) involves mixing a known amount of target analyte with an antibody specific for the analyte, then adding unlabeled analyte from the sample, and measuring the amount of labeled analyte displaced (see, for example, for guidance, An Introduction to Radioimmunoassay and Related Techniques, edited by Chard T, Elsevier Science 1995, ISBN 0444821198).

[0098] Detectable labels can be used in the assays described herein for direct or indirect detection of biomarkers in the methods of the present invention. A variety of detectable labels can be used, and the choice of label depends on the required sensitivity, ease of conjugation with antibody, stability requirements, and available equipment and disposal regulations. Those skilled in the art are familiar with the selection of suitable detectable labels based on the assay detection of biomarkers in the methods of the present invention. Suitable detectable labels include, but are not limited to, fluorescent dyes (e.g., fluorescein, fluorescein isothiocyanate (FITC), Oregon Green™, rhodamine, Texas Red, tetrarhodimine isothiocynate (TRITC), Cy3, Cy5, etc.), fluorescent markers (e.g., green fluorescent protein (GFP), phycoerythrin, etc.), enzymes (e.g., luciferase, horseradish peroxidase, alkaline phosphatase, etc.), nanoparticles, biotin, digoxigenin, metals, etc.

[0099] For mass-sectrometry based analysis, differential tagging with isotopic reagents (e.g., isotope-coded affinity tags (ICAT) or more recent variations using isobaric tagging reagents, iTRAQ (Applied Biosystems, Foster City, Calif.), or tandem mass tags, TMT (Thermo Scientific, Rockford, IL), followed by multidimensional liquid chromatography (LC) and tandem mass spectrometry (MS / MS) analysis) can provide further methodologies in carrying out the methods of the invention.

[0100] Chemiluminescence assays using chemiluminescent antibodies can be used for sensitive non-radioactive detection of protein levels. Antibodies labeled with fluorescent dyes can also be suitable. Examples of fluorescent dyes include, but are not limited to, DAPI, fluorescein, Hoechst 33258, R phycocyanin, B phycoerythrin, R phycoerythrin, rhodamine, Texas Red, and Lissamine. Indirect labels include various enzymes well known in the art, such as horseradish peroxidase (HRP), alkaline phosphatase (AP), beta-galactosidase, urease, etc. Detection systems using suitable substrates for horseradish peroxidase, alkaline phosphatase, and beta-galactosidase are well known in the art.

[0101] Signals from direct or indirect labels can be detected using, for example, spectrophotometers to detect color from chromogenic substrates; radiation counters to detect radiation, for example 125For detection of enzyme-linked antibodies, quantitative analysis can be performed using a spectrophotometer, such as the EMAX Microplate Reader (Molecular Devices; Menlo Park, Calif.), according to the manufacturer's instructions. If desired, the assays used to practice the invention can be automated or performed robotically, and signals from multiple samples can be detected simultaneously.

[0102] In some embodiments, the methods described herein include quantification of biomarkers using mass spectrometry (MS). In further embodiments, the mass spectrometry can be liquid chromatography-mass spectrometry (LC-MS), multiple reaction monitoring (MRM), or selected reaction monitoring (SRM). In additional embodiments, the MRM or SRM can further include scheduled MRM or scheduled SRM.

[0103] As mentioned above, chromatography can also be used in carrying out the method of the present invention. Chromatography encompasses methods for separating chemical substances and generally includes a process in which a mixture of analytes is carried by a moving stream of liquid or gas ("mobile phase") and separated into components as they flow around or over a stationary liquid or solid phase ("stationary phase"), as a result of the differential distribution of analytes between the mobile phase and said stationary phase. The stationary phase can usually be a finely divided solid, a sheet of filter material, or a thin film of liquid on the surface of a solid, etc. Chromatography is well understood by those skilled in the art as a technique applicable for the separation of chemical compounds of biological origin, such as amino acids, proteins, protein fragments, or peptides.

[0104] Chromatography can be columnar (i.e., the stationary phase is deposited or packed in a column), preferably liquid chromatography, even more preferably high performance liquid chromatography (HPLC), or ultra-high performance / high pressure liquid chromatography (UHPLC). Details of chromatography are well known in the art (Bidlingmeyer, Practical HPLC Methodology and Applications, John Wiley & Sons Inc., 1993). Exemplary types of chromatography include, but are not limited to, high performance liquid chromatography (HPLC), UHPLC, normal phase HPLC (NP-HPLC), reverse phase HPLC (RP-HPLC), ion exchange chromatography (IEC) (such as cation or anion exchange chromatography), hydrophilic interaction chromatography (HILIC), hydrophobic interaction chromatography (HIC), size exclusion chromatography (SEC) (including gel filtration or gel permeation chromatography), chromatofocusing, affinity chromatography (such as immunoaffinity, immobilized metal affinity chromatography, etc.), and the like. Chromatography, including single, two or more dimensional chromatography, can be used as a further method of peptide analysis, for example, as a peptide fractionation method in conjunction with downstream mass spectrometry as described elsewhere herein.

[0105] Additional peptide or polypeptide separation, identification, or quantification methods may be used to measure the biomarkers in this disclosure, optionally in conjunction with any of the analytical methods described above. Such methods include, but are not limited to, chemical extraction partitioning, isoelectric focusing (IEF) (including capillary isoelectric focusing (CIEF), capillary isotachophoresis (CITP), capillary electrochromatography (CEC), etc.), one-dimensional polyacrylamide gel electrophoresis (PAGE), two-dimensional polyacrylamide gel electrophoresis (2D-PAGE), capillary gel electrophoresis (CGE), capillary zone electrophoresis (CZE), micellar electrokinetic chromatography (MEKC), free-flow electrophoresis (FFE), etc.

[0106] In the context of the present invention, the term "capture agent" refers to a compound capable of specifically binding to a target, in particular a biomarker. This term includes antibodies, antibody fragments, nucleic acid-based protein binding reagents (e.g., aptamers, Slow Off-rate Modified Aptamers (SOMAmers™)), protein capture agents, natural ligands (i.e., hormones for their receptors or vice versa), small molecules or variants thereof.

[0107] The capture agent can be configured to specifically bind to a target, particularly a biomarker. The capture agent can include, but is not limited to, organic molecules, such as polypeptides, polynucleotides, and other non-polymeric molecules that can be identified by those skilled in the art. In the embodiments disclosed herein, the capture agent includes any agent that can be used to detect, purify, isolate, or enrich a target, particularly a biomarker. Any affinity capture technique known in the art can be used to selectively isolate and enrich / concentrate biomarkers that are components of a complex mixture of biological media for use in the disclosed methods.

[0108] Antibody capture agents that specifically bind to biomarkers can be prepared using any suitable method known in the art. See, for example, Coligan, Current Protocols in Immunology (1991); Harlow & Lane, Antibodies: A Laboratory Manual (1988); Goding, Monoclonal Antibodies: Principles and Practice (2nd ed., 1986). The antibody capture agent can be any immunoglobulin or derivative thereof, whether natural or wholly or partially synthetically produced. All derivatives thereof that maintain specific binding ability are also included in the term. The antibody capture agent has a binding domain that is homologous or largely homologous to an immunoglobulin binding domain, and can be derived from natural sources or can be partially or wholly synthetically produced. The antibody capture agent can be a monoclonal or polyclonal antibody. In some embodiments, the antibody is a single chain antibody. Those skilled in the art will appreciate that the antibody can be provided in any of a variety of forms, including, for example, humanized, partially humanized, chimeric, chimeric humanized, and the like. The antibody capture agent may be an antibody fragment, including, but not limited to, Fab, Fab', F(ab')2, scFv, Fv, dsFv diabody, and Fd fragments. The antibody capture agent may be produced by any means. For example, the antibody capture agent may be enzymatically or chemically produced by fragmentation of an intact antibody, and / or it may be recombinantly produced from a gene encoding a partial antibody sequence. The antibody capture agent may include a single chain antibody fragment. Alternatively, or in addition, the antibody capture agent may include multiple chains linked together, for example, by disulfide bonds; and any functional fragment obtained from such molecules, in which such fragments retain the specific binding characteristics of the parent antibody molecule. Due to their smaller size as a functional component of the whole molecule, antibody fragments may offer advantages over intact antibodies for use in certain immunochemical techniques and experimental applications.

[0109] Suitable capture agents useful for the implementation of the present invention also include aptamers. Aptamers are oligonucleotide sequences that can specifically bind to their targets through a unique three-dimensional (3-D) structure. Aptamers can include any suitable number of nucleotides, and different aptamers can have the same or different number of nucleotides. Aptamers can be DNA or RNA or chemically modified nucleic acids, and can be single-stranded, double-stranded, or can include double-stranded regions and higher dimensional structures. Aptamers can also be photoaptamers, in which a photoreactive or chemically reactive functional group is included in the aptamer so that it can covalently bind to its corresponding target. The use of aptamer capture agents can include the use of two or more aptamers that specifically bind to the same biomarker. Aptamers can include tags. Aptamers can be identified using any known method, including the SELEX (Systematic Evolution of Ligands by Exponential Enrichment) process. Once identified, aptamers can be prepared or synthesized according to any known method, including chemical and enzymatic synthesis, and used in various applications for biomarker detection. Liu et al., Curr Med Chem. 18(27):4117-25 (2011). Capture agents useful in carrying out the methods of the present invention also include SOMAmers (Slow Off-Rate Modified Aptamers), known in the art, and have improved off-rate characteristics. Brody et al., J Mol Biol. 422(5):595-606 (2012). SOMAmers can be generated using any known method, including the SELEX method.

[0110] It will be understood by those skilled in the art that biomarkers can be modified prior to analysis to improve their resolution or to determine their identity. For example, biomarkers can be subjected to proteolytic digestion prior to analysis. Any protease can be used. Proteases that can cleave biomarkers into a distinct number of fragments, such as trypsin, are particularly useful. The fragments resulting from digestion serve as fingerprints for the biomarkers, thereby indirectly allowing their detection. This is particularly useful when there are biomarkers with similar molecular weights that can be confused for the biomarker in question. Proteolytic fragmentation is also useful for biomarkers with high molecular weights, since smaller biomarkers are more easily resolved by mass spectrometry. In another example, biomarkers can be modified to improve detection resolution. For example, neuraminidase can be used to remove terminal sialic acid residues from glycoproteins to improve binding to anionic adsorbents and improve detection resolution. In another example, biomarkers can be modified by attachment of tags of specific molecular weights that specifically bind to molecular biomarkers, further distinguishing them. Optionally, after detecting such modified biomarkers, the identity of the biomarkers can be further determined by matching the physical and chemical characteristics of the modified biomarkers in a protein database (e.g., SwissProt).

[0111] It is further understood in the art that biomarkers in a sample can be captured on a substrate for detection. Traditional substrates include antibody-coated 96-well plates or nitrocellulose membranes that are then probed for the presence of proteins. Alternatively, protein-binding molecules attached to microspheres, microparticles, microbeads, beads, or other particles can be used for capturing and detecting biomarkers. Protein-binding molecules can be antibodies, peptides, peptoids, aptamers, small molecule ligands, or other protein-binding capture agents attached to the surface of a particle. Each protein-binding molecule can contain a unique detectable label that is encoded, allowing it to be distinguished from other detectable labels attached to other protein-binding molecules and allowing the detection of biomarkers in multiplex assays. Examples include, but are not limited to, color-coded microspheres with known fluorescence intensities (see, e.g., microspheres using xMAP technology produced by Luminex (Austin, Tex.)); microspheres containing quantum dot nanocrystals, e.g., with different ratios and combinations of quantum dot colors (see, e.g., Qdot nanocrystals produced by Life Technologies (Carlsbad, Calif.)); glass-coated metal nanoparticles (see, e.g., SERS nanotags produced by Nanoplex Technologies, Inc. (Mountain View, Calif.)); barcode materials (e.g., striped metal rods of submicron size, such as Nanobarcode produced by Nanoplex Technologies, Inc.), coded microparticles with color barcodes (see, e.g., CellCard produced by Vitra Bioscience, vitrabio.com), glass microparticles with digital holographic code images (see, e.g., CyVera microbeads produced by Illumina (San Diego, Calif.)); chemiluminescent dyes, combinations of dye compounds; and detectable beads of different sizes.

[0112] In another embodiment, biochips can be used for the capture and detection of the biomarkers of the present invention. Many protein biochips are known in the art. These include, for example, protein biochips produced by Packard BioScience Company (Meriden Conn.), Zyomyx (Hayward, Calif.), and Phylos (Lexington, Mass.). Generally, protein biochips include a substrate having a surface. A capture reagent or adsorbent is attached to the substrate's surface. Frequently, the surface includes a plurality of addressable locations, each of which has a capture agent bound thereto. The capture agent can be a biological molecule, such as a polypeptide or a nucleic acid, which specifically captures other biomarkers. Alternatively, the capture agent can be a chromatographic material, such as an anion exchange material or a hydrophilic material. Examples of protein biochips are well known in the art.

[0113] Measurement of mRNA in biological samples can be used as a proxy for detecting the level of the corresponding protein biomarker in biological samples. Thus, any of the biomarkers or biomarker panels described herein can also be detected by detecting the appropriate RNA. mRNA levels can be measured by reverse transcription quantitative polymerase chain reaction (RT-PCR, followed by qPCR). RT-PCR is used to generate cDNA from mRNA. cDNA can be used in qPCR assays to produce fluorescence as the DNA amplification process proceeds. By comparison with a standard curve, qPCR can produce absolute measurements, such as the number of mRNA copies per cell. Northern blots, microarrays, invader assays, and RT-PCR combined with capillary electrophoresis have all been used to measure the expression level of mRNA in samples. See Gene Expression Profiling: Methods and Protocols, edited by Richard A. Shimkets, Humana Press, 2004.

[0114] Some embodiments disclosed herein relate to diagnostic and prognostic methods for determining the probability of premature birth in pregnant women.Detection of the expression level of one or more biomarkers and / or determination of the ratio of biomarkers can be used to determine the probability of premature birth in pregnant women.Such detection methods can be used, for example, for early diagnosis of conditions, to determine whether a subject has a predisposition to premature birth, to monitor the progress of premature birth or the progress of treatment protocols, to evaluate the severity of premature birth, to predict the outcome of premature birth and / or the prospect of recovery or full-term birth, or to help determine the appropriate treatment for premature birth.

[0115] The quantification of biomarkers in biological samples can be determined by the methods described above as well as any other methods known in the art, without being limited thereto.The quantitative data thus obtained is then subjected to an analytical classification process.In such a process, raw data is manipulated according to an algorithm that is predefined by a training set of data, for example as described in the examples provided herein.The algorithm can utilize the training set of data provided herein, or can use the guidelines provided herein to generate the algorithm using different sets of data.

[0116] In some embodiments, analyzing the measurable features to determine the probability of preterm birth in pregnant women comprises using a predictive model. In further embodiments, analyzing the measurable features to determine the probability of preterm birth in pregnant women comprises comparing the measurable features to reference features. As those skilled in the art can understand, such comparisons can be direct comparisons with reference features or indirect comparisons where the reference features are incorporated into a predictive model. In further embodiments, analyzing the measurable features to determine the probability of preterm birth in pregnant women comprises one or more of a linear discriminant analysis model, a support vector machine classification algorithm, a recursive feature elimination model, a predictive analysis of microarray models, a logistic regression model, a CART algorithm, a flextree algorithm, a LART algorithm, a random forest algorithm, a MART algorithm, a machine learning algorithm, a penalized regression method, or a combination thereof. In certain embodiments, the analysis comprises logistic regression.

[0117] In the analytical classification process, any one of a variety of statistical analysis methods can be used to manipulate the quantitative data to provide a classification of the sample. Examples of useful methods include linear discriminant analysis, recursive feature elimination, predictive analysis of microarrays, logistic regression, CART algorithms, FlexTree algorithms, LART algorithms, random forest algorithms, MART algorithms, machine learning algorithms, and the like.

[0118] To create a random forest for prediction of GAB, one skilled in the art can consider a set of k subjects (pregnant women) whose gestational age at birth (GAB) is known and N analytes (transitions) have been measured in blood samples taken a few weeks before delivery. The regression tree starts with a root node that includes all subjects. The average GAB for all subjects can be calculated at the root node. The variance of GAB in the root node will be high because there is a mix of women with different GABs. The root node is then partitioned into two branches, each containing women with similar GABs. The average GAB for subjects in each branch is calculated again. The variance of GAB in each branch will be lower than the root node because the subset of women in each branch have relatively more similar GABs than those in the root node. The two branches are created by selecting the analytes and a threshold for the analytes that creates the branches with similar GABs. The analytes and thresholds are chosen among the set of all analytes and thresholds, typically with a random subset of analytes at each node. The procedure continues to recursively generate branches, creating leaves (terminal nodes) where the subjects have very similar GABs. The predicted GAB at each terminal node is the average GAB for the subjects at that terminal node. This procedure creates a single regression tree. A random forest can consist of hundreds or thousands of such trees.

[0119] Classification can be performed according to a predictive model method, which sets a threshold value to determine the probability that a sample belongs to a given class. The probability is preferably at least 50%, or at least 60%, or at least 70%, or at least 80% or higher. Classification can also be performed by determining whether a comparison between the obtained data set and a reference data set results in a statistically significant difference. If so, then the sample from which the data set is obtained is classified as not belonging to the class of the reference data set. Conversely, if such a comparison is not statistically significantly different from the reference data set, then the sample from which the data set is obtained is classified as belonging to the class of the reference data set.

[0120] The predictive ability of a model can be evaluated according to its ability to provide a quality metric (e.g., AUROC (area under the ROC curve) or accuracy at a particular value or range of values). Area under the curve measurements are useful for comparing the accuracy of classifiers across a complete data range. Classifiers with larger AUCs have a greater ability to accurately classify unknowns between two groups of interest. In some embodiments, the desired quality threshold is a predictive model that classifies samples with at least about 0.5, at least about 0.55, at least about 0.6, at least about 0.7, at least about 0.75, at least about 0.8, at least about 0.85, at least about 0.9, at least about 0.95, or higher accuracy. As an alternative measurement, the desired quality threshold may refer to a predictive model that classifies samples with at least about 0.7, at least about 0.75, at least about 0.8, at least about 0.85, at least about 0.9, or higher AUC.

[0121] As known in the art, the relative sensitivity and specificity of a predictive model can be adjusted to favor either the selection metric or the sensitivity metric, where the two metrics have an inverse relationship.The limitations in the models described above can be adjusted to provide a selected level of sensitivity or specificity, depending on the specific requirements of the test being carried out.One or both of sensitivity and specificity can be at least about 0.7, at least about 0.75, at least about 0.8, at least about 0.85, at least about 0.9, or higher.

[0122] The raw data can be analyzed initially by measuring values ​​for each biomarker, usually in triplicate or multiple triplicates. The data can be manipulated, for example, the raw data can be transformed using standard curves and the averages of the triplicate measurements used to calculate the mean and standard deviation for each patient. These values ​​can be transformed (e.g., logarithmic transformation, Box-Cox transformation) before being used in the model (Box and Cox, Royal Stat. Soc., Series B, vol. 26:211-246 (1964)). The data is then input into a predictive model that classifies the sample according to condition. The resulting information can be communicated to the patient or health care provider.

[0123] To generate a prediction model for preterm birth, a robust data set is used in training set, which includes known control samples and samples corresponding to the target preterm birth classification.Sample size can be selected using generally accepted criteria.As discussed above, different statistical methods can be used to obtain highly accurate prediction models.An example of such analysis is provided in Example 2.

[0124] In one embodiment, hierarchical clustering is performed in the derivation of the predictive model, where Pearson correlation is used as the clustering metric. One approach is to consider the preterm birth dataset as a "training sample" in a "supervised learning" problem. CART is a standard in medical applications (Singer, Recursive Partitioning in the Health Sciences, Springer (1999)) and is used to convert any qualitative feature into a quantitative feature; Hotelling T 2 They can be modified by sorting them by the achieved significance level, assessed by the sample reuse method for the statistics; and by appropriate application of the lasso method. Problems in prediction become problems in regression, without loss of vision in prediction, by making practical use of the Gini criterion for classification in assessing the quality of the regression.

[0125] This approach led to what is called FlexTree (Huang, Proc. Nat. Acad. Sci. USA 101:10529-10534 (2004)). FlexTree performs very well in simulations and when applied to multiple forms of data, and is useful for implementing the methods claimed. Software has been developed to automate FlexTree. Alternatively, LARTree or LART can be used (Turnbull (2005) Classification Trees with Subset Analysis Selection by the Lasso, Stanford University). The name reflects the implementation of the Lasso through a binary tree; the lasso described; and what is called LARS by Efron et al. (2004) Annals of Statistics 32:407-451 (2004). See also Huang et al., Proc. Natl. Acad. Sci. USA. 101(29):10529-34 (2004). Other analytical methods that can be used include logistic regression. One method of logistic regression: Ruczinski, Journal of Computational and Graphical Statistics 12:475-512 (2003). Logistic regression is similar to CART in that its classifier can be viewed as a binary tree. It differs in that each node has a Boolean description of the features, which is more general than the simple "and" description produced by CART.

[0126] Another approach is the nearest shrink centroid approach (Tibshirani, Proc. Natl. Acad. Sci. USA 99:6567-72 (2002)). This technique is like k-means, but has the advantage of automatically selecting features by shrinking cluster centers, as in the case of lasso, and focusing attention on the few that are informative. This approach is available as PAM software and is widely used. Two additional sets of algorithms that can be used are Random Forest (Breiman, Machine Learning 45:5-32 (2001)) and MART (Hastie, The Elements of Statistical Learning, Springer (2001)). These two methods are known in the art as "committee methods" that involve predictors that "vote" on the outcome.

[0127] To provide a significant ordering, a false discovery rate (FDR) can be determined. First, a set of null distributions of dissimilarity values ​​is generated. In one embodiment, the observed profile values ​​are permuted to create a set of distributions of correlation coefficients that would be obtained by chance, thereby creating an appropriate set of null distributions of correlation coefficients (Tusher et al., Proc. Natl. Acad. Sci. USA 98:5116-21 (2001)). The set of null distributions is obtained by permuting the values ​​of each profile of all available profiles; calculating pairwise correlation coefficients for all profiles; calculating the probability density function of the correlation coefficients for this permutation; and repeating this procedure N times (where N is a large number, typically 300). The N distribution is used to calculate an appropriate measure (mean, median, etc.) of the count of correlation coefficient values ​​whose values ​​exceed the (similar) value obtained from the distribution of experimentally observed similar values ​​at a given significance level.

[0128] The FDR is the ratio of the expected number of false significant correlations (estimated from correlations in the randomized data set that are greater than this selected Pearson correlation) to the number of correlations in the empirical data that are greater than this selected Pearson correlation (significant correlation). This cutoff correlation value can be applied to the correlation between experimental profiles. Using the above distribution, a level of confidence is chosen for significance. This is used to determine the minimum value of the correlation coefficient that exceeds the result that would have been obtained by chance. This method is used to obtain a threshold for positive correlation, negative correlation, or both. Using this threshold, the user can filter the pairwise correlation coefficient observations and eliminate those that do not exceed the threshold. Additionally, an estimate of the false positive rate can be obtained for a given threshold. For each of the individual "random correlation" distributions, it can be found how many observations are outside the threshold range. This procedure provides a series of counts. The mean and standard deviation of the sequence provide the average number of potential false positives and their standard deviation.

[0129] In an alternative analytical approach, the variables selected in the cross-sectional analysis are used separately as predictors in a time-to-event analysis (survival analysis), where the event is the occurrence of preterm birth, and subjects without an event are considered censored at birth. Assuming a specific pregnancy outcome (preterm birth event or no event), a random length of time each patient is observed, and the selection of proteomic and other features, a parametric approach to analyzing survival may be better than the widely applied semi-parametric Cox model. The Weibull parametric fit of survival allows the hazard ratio to be monotonically increasing, decreasing, or constant, and also has a proportional hazards representation (as in the Cox model) and an accelerated failure time representation. All standard tools available in obtaining approximate maximum likelihood estimators of regression coefficients and corresponding functions are available in this model.

[0130] In addition, Cox model can be used.In particular, because the reduction of the number of covariates to a manageable size with lasso significantly simplifies the analysis, and allows the possibility of non-parametric or semi-parametric approach to predicting the time to preterm birth.These statistical tools are known in the art and are applicable to all types of proteomics data.A set of biomarkers, clinical data, and genetic data that can be easily determined and is highly informative regarding the probability of preterm birth in said pregnant women and the predicted time to preterm birth event is provided.Algorithm also provides information regarding the probability of preterm birth in pregnant women.

[0131] Therefore, those skilled in the art will understand that the probability of premature birth according to the present invention can be determined using either quantitative or categorical variables. For example, when carrying out the method of the present invention, the measurable features of each of the N biomarkers can be subjected to categorical data analysis to determine the probability of premature birth as a binary categorical outcome. Alternatively, the method of the present invention can analyze the measurable features of each of the N biomarkers by first calculating the quantitative variables, in particular the predicted gestational age at birth. The predicted gestational age at birth can then be used as the basis for predicting the risk of premature birth. By first using quantitative variables and then converting them into categorical variables, the method of the present invention takes into account the continuity of measurements detected for the measurable features. For example, rather than making a binary prediction of premature birth versus full-term birth, predicting the gestational age at birth allows treatment to be individualized for pregnant women. For example, an earlier predicted gestational age at birth will result in more intensive prenatal intervention, i.e., monitoring and treatment, than a predicted gestational age approaching full term.

[0132] Among women with a predicted GAB of j days plus or minus k days, p(PTB) can be estimated as the proportion of women in the PAPR clinical trial with a predicted GAB of j days plus or minus k days who actually deliver before 37 weeks gestation (see Example 1). More generally, for women with a predicted GAB of j days plus or minus k days, the probability that the actual gestational age at birth is less than the specified gestational age, p(actual GAB<specified GAB), was estimated as the proportion of women in the PAPR clinical trial with a predicted GAB of j days plus or minus k days who actually deliver before the specified gestational age.

[0133] In developing a predictive model, it may be desirable to select a subset of markers, i.e., at least 3, at least 4, at least 5, at least 6, up to a complete set of markers. Typically, a subset of markers is chosen that provides what is needed for quantitative sample analysis, e.g., availability of reagents, ease of quantification, etc., while maintaining a highly accurate predictive model. Selection of some informative markers for building a classification model requires the definition of a performance metric and a user-defined threshold to produce a model with useful predictive ability based on this metric. For example, the performance metric may be the AUC, the sensitivity and / or specificity of the prediction, and the overall accuracy of the predictive model.

[0134] As will be understood by those skilled in the art, the analytical classification process can use any one of a variety of statistical analysis methods to manipulate quantitative data and provide classification of samples.Examples of useful methods include, but are not limited to, linear discriminant analysis, recursive feature elimination, predictive analysis of microarrays, logistic regression, CART algorithm, FlexTree algorithm, LART algorithm, random forest algorithm, MART algorithm, and machine learning algorithm.

[0135] As described in Example 2, various methods are used in training models. The selection of a subset of markers can be for forward selection or backward selection of a subset of markers. The number of markers that optimizes the performance of the model can be selected without using all markers. One method for defining the optimal number of items is to choose the number of items that produces a model with the desired predictive ability (e.g., AUC>0.75, or equivalent measurement of sensitivity / specificity) using any combination and number of items used for a given algorithm, with a standard error of 1 or less from the maximum value obtained for this metric.

[0136] Table 1. Transitions with p values ​​less than 0.05 in univariate Cox proportional hazards analysis for predicting gestational age at birth [Table 1]

[0137] Table 2. Transitions selected by Cox stepwise AIC analysis [Table 2]

[0138] Table 3. Transitions selected by the Cox lasso model [Table 3]

[0139] Table 4. Area under the ROC curve (AUROC) for individual analytes discriminating preterm from non-preterm subjects. The 77 transitions with the highest AUROC areas are shown. [Table 4-1] [Table 4-2]

[0140] Table 5. AUROC of random forest, boosting, lasso, and logistic regression models for a given number of transitions allowed in the model, estimated by 100 rounds of bootstrap resampling. [Table 5]

[0141] Table 6. Top 15 transitions selected by each multivariate method, ranked by their importance for that method. [Table 6]

[0142] In yet another embodiment, the present invention provides a kit for determining the probability of preterm birth, the kit can be used to detect N of the isolated biomarkers listed in Tables 1 to 63. For example, the kit can be used to detect one or more, two or more, or three of the isolated biomarkers selected from the group consisting of AFTECCVVASQLR, ELLESYIDGR, and ITLPDFTGDLR. For example, the kit can be used to detect one or more, two or more, or three of the isolated biomarkers selected from the group consisting of FLNWIK, FGFGGSTDSGPIR, LLELTGPK, VEHSDLSFSK, IEGNLIFDPNYLPK, ALVLELAK, TQILEWAAER, DVLLLVHNLPQNLPGYFWYK, SEPRPGVLLR, ITQDAQLK, ALDLSLK, WWGGQPLWITATK, and LSETNR.

[0143] In another embodiment, the kit can be used to detect one or more, two or more, three or more, four or more, five or more, six or more, seven or more, or eight isolated biomarkers selected from the group consisting of lipopolysaccharide binding protein (LBP), prothrombin (THRB), complement component C5 (C5 or CO5), plasminogen (PLMN), and complement component C8 gamma chain (C8G or CO8G).

[0144] In another embodiment, the kit can be used to detect one or more, two or more, three or more, four or more, five or more, six or more, seven or more, or eight of the isolated biomarkers selected from the group consisting of alpha-1B-glycoprotein (A1BG), a disintegrin and metalloproteinase domain-containing protein 12 (ADA12), apolipoprotein B-100 (APOB), beta-2-microglobulin (B2MG), CCAAT / enhancer binding protein alpha / beta (HP8 peptide), corticosteroid binding globulin (CBG), complement component C6, endoglin (EGLN), ectonucleotide pyrophosphatase / phosphodiesterase family member 2 (ENPP2), coagulation factor VII (FA7), hyaluronan binding protein 2 (HABP2), pregnancy specific beta-1-glycoprotein 9 (PSG9), inhibin beta E chain (INHBE).

[0145] The kit may include one or more agents for detection of the biomarkers, a container for holding an isolated biological sample from a pregnant woman; and printed instructions for reacting the agents with the biological sample or a portion of the biological sample to detect the presence or amount of the isolated biomarker in the biological sample. The agents may be packaged in separate containers. The kit may further include one or more control reference samples and reagents for performing the immunoassay.

[0146] In one embodiment, the kit comprises agents for measuring the levels of at least N of the isolated biomarkers listed in Tables 1 to 63. The kit can include antibodies that specifically bind to these biomarkers, for example, the kit can include at least one of an antibody that specifically binds to lipopolysaccharide binding protein (LBP), an antibody that specifically binds to prothrombin (THRB), an antibody that specifically binds to complement component C5 (C5 or CO5), an antibody that specifically binds to plasminogen (PLMN), and an antibody that specifically binds to complement component C8 gamma chain (C8G or CO8G).

[0147] In one embodiment, the kit comprises agents for measuring the levels of at least N of the isolated biomarkers listed in Tables 1 to 63. The kit can comprise antibodies that specifically bind to these biomarkers, for example, the kit can comprise at least one of antibodies that specifically bind to alpha-1B-glycoprotein (A1BG), a disintegrin and metalloproteinase domain-containing protein 12 (ADA12), apolipoprotein B-100 (APOB), beta-2-microglobulin (B2MG), CCAAT / enhancer binding protein alpha / beta (HP8 peptide), corticosteroid binding globulin (CBG), complement component C6, endoglin (EGLN), ectonucleotide pyrophosphatase / phosphodiesterase family member 2 (ENPP2), coagulation factor VII (FA7), hyaluronan binding protein 2 (HABP2), pregnancy specific beta-1-glycoprotein 9 (PSG9), inhibin beta E chain (INHBE).

[0148] The kit can include one or more containers for the composition contained in the kit. The composition can be in liquid form or can be lyophilized. Suitable containers for the composition include, for example, bottles, vials, syringes, and test tubes. The container can be made of various materials, including glass or plastic. The kit can also include a package insert that includes instructions on how to determine the probability of premature birth.

[0149] From the foregoing description, it will be apparent that variations and modifications can be made to the invention described herein to adapt it to various usages and conditions, such embodiments also falling within the scope of the following claims.

[0150] The recitation of a list of elements in any definition of a variable herein includes definitions of that variable as any single element or combination (or subcombination) of the listed elements. The recitation of an embodiment herein includes that embodiment as any single embodiment or in combination with any other embodiment or portion thereof.

[0151] All patents and publications mentioned in this specification are herein incorporated by reference to the same extent as if each individual patent and publication was specifically and individually indicated to be incorporated by reference.

[0152] The following examples are offered by way of illustration, and not by way of limitation. EXAMPLES

[0153] Example 1: Development of a sample set for biomarker discovery and validation for preterm birth A standard protocol was developed to govern the conduct of the Proteomic Assessment of Preterm Risk (PAPR) clinical study. The protocol also specified that samples and clinical information could be used to study other pregnancy complications for a portion of the subjects. Specimens were obtained from women at 11 Internal Review Board (IRB)-certified sites across the United States. After providing informed consent, serum and plasma samples were obtained, as well as relevant information regarding patient demographic characteristics, past medical and reproductive history, current pregnancy history, and concomitant medications. Following delivery, data related to maternal and infant status and complications were collected. Serum and plasma samples were processed according to a protocol calling for standard refrigerated centrifugation, dispensing of samples into 0.5 ml 2-D barcoded cryovials, and subsequent freezing at -80°C.

[0154] Following delivery, preterm birth cases were considered individually and their status determined as either spontaneous preterm or medically indicated preterm. Only spontaneous preterm birth cases were used for this analysis. For preterm birth biomarker discovery, 80 samples were analyzed in two gestational age groups: a) a late window comprised of samples from 23-28 weeks gestation (included 13 cases, 13 term controls matched within 1 week of sample collection and 14 random term controls) and b) an early window comprised of samples from 17-22 weeks gestation (included 15 cases, 15 term controls matched within 1 week of sample collection and 10 random term controls).

[0155] The samples were then depleted of abundant proteins using the Human 14 Multiple Affinity Removal System (MARS 14), which removes 14 of the most abundant proteins that are treated as uninformative for the identification of disease-related changes in the serum proteome. For this purpose, equal volumes of each of the clinical or pooled human serum samples (HGS) samples were diluted with column buffer and filtered to remove precipitates. The filtered samples were depleted using a MARS-14 column (4.6 x 100 mm, Cat. No. 5188-6558, Agilent Technologies). The samples were cooled to 4°C in the autosampler, the depletion column was run at room temperature, and the collected fractions were kept at 4°C until further analysis. The unbound fraction was collected for further analysis.

[0156] A second aliquot of each clinical serum sample and of each HGS was diluted in ammonium bicarbonate buffer and depleted of 14 highly and approximately 60 additional moderately abundant proteins using IgY14-SuperMix (Sigma) hand-packed columns consisting of 10 mL of bulk material (50% slurry, Sigma). Shi et al., Methods, 56(2):246-53 (2012). Samples were cooled to 4°C in an autosampler, the depletion column was run at room temperature, and collected fractions were kept at 4°C until further analysis. The unbound fraction was collected for further analysis.

[0157] Depleted serum samples were denatured with trifluorethanol, reduced with dithiotreitol, alkylated using iodoacetamide, and then digested with trypsin at a trypsin:protein ratio of 1:10. Following trypsin digestion, samples were desalted on a C18 column and the eluate was lyophilized. Desalted samples were reconstituted in reconstitution buffer containing five internal standard peptides.

[0158] Depleted and trypsin digested samples were analyzed using scheduled multiple reaction monitoring (sMRM). Peptides were analyzed using a Waters Nano Acquity Separation was performed using UPLC on a 150 mm x 0.32 mm Bio-Basic C18 column (ThermoFisher) at a flow rate of 5 μl / min and eluted using an acetonitrile gradient in an AB SCIEX QTRAP 5500 with a Turbo V source (AB SCIEX, Framingham, MA). The sMRM assay measured 1708 transitions corresponding to 854 peptides and 236 proteins. Chromatographic peaks were integrated using Rosetta Elucidator software (Ceiba Solutions).

[0159] Transitions were excluded from the analysis if their intensity area count was less than 10000 and if they were missing in more than 3 samples per batch. Intensity area counts were log-transformed and mass spectrometry run order trends and depletion batch effects were minimized using regression analysis.

[0160] Example 2: Analysis of Transitions to Identify Preterm Birth Biomarkers I The purpose of these analyses was to examine the data collected in Example 1 to identify transitions and proteins predictive of preterm birth. The specific analyses used were (i) Cox time-to-event analysis and (ii) models with preterm birth as a binary categorical dependent variable. The dependent variable for all Cox analyses was gestational age of the time to event, where the event was preterm birth. For the purposes of the Cox analysis, subjects with preterm birth have an event on their birthday. Subjects with full-term birth were censored on their birthday. Gestational age at the date of sample collection was a covariate in all Cox analyses.

[0161] Assay data were previously adjusted for run order and depletion batch and log-transformed. Values ​​for gestational age at sample collection were adjusted as follows: Transition values ​​were regressed on gestational age at sample collection using only controls (non-preterm subjects). Residuals from the regression were designated as adjusted values. Adjusted values ​​were used in models with preterm birth as a binary categorical dependent variable. Unadjusted values ​​were used in Cox analyses.

[0162] Univariate Cox proportional hazards analysis Univariate Cox proportional hazards analyses were performed to predict gestational age at birth, including gestational age at the date of specimen collection as a covariate. Table 1 shows transitions with p values ​​less than 0.05. Five proteins had multiple transitions among transitions with p values ​​less than 0.05: lipopolysaccharide-binding protein (LBP), prothrombin (THRB), complement component C5 (C5 or CO5), plasminogen (PLMN), and complement component C8 gamma chain (C8G or CO8G).

[0163] Multivariate Cox proportional hazards analysis: stepwise AIC selection Cox proportional hazards analyses were performed to predict gestational age at birth using a stepwise lasso model for variable selection, including gestational age at sample collection date as a covariate. These analyses included a total of n=80 subjects, with number of PTB events=28. The Akaike Information Criterion (AIC) was used as the stopping criterion in the stepwise variable selection analyses. Table 2 shows the transitions selected by the stepwise AIC analysis. The coefficient of determination (R 2 ) is 0.86 (uncorrected for multiple comparisons).

[0164] Multivariate Cox proportional hazards analysis: Lasso selection Lasso variable selection was used as a second method in multivariate Cox proportional hazards analysis to predict gestational age at birth, including gestational age at sample collection date as a covariate. In this analysis, a lambda penalty for lasso estimated by cross-validation is used. The results are shown in Table 3. The lasso variable selection method is significantly more stringent than the stepwise AIC, and only three transitions were selected for the final model, representing three different proteins. These three proteins give the top four transitions from the univariate analysis. Two of the top four univariates are from the same protein, and therefore, not both are selected by the lasso method. The lasso tends to select a relatively small number of variables with low intercorrelation. The coefficient of determination (R 2 ) is 0.21 (uncorrected for multiple comparisons).

[0165] Univariate AUROC analysis of preterm birth as a binary categorical dependent variable Univariate analyses were performed to distinguish preterm subjects from non-preterm subjects (preterm as a binary categorical variable) as estimated by the area under the receiver operating characteristic curve (AUROC). These analyses use transition values ​​adjusted for gestational age at sample collection as described above. Table 4 shows the AUROC curves for the 77 transitions with the highest AUROC area of ​​0.6 or greater.

[0166] Multivariate analysis of preterm birth as a binary categorical dependent variable Multivariate analyses were performed to predict preterm birth as a binary categorical dependent variable using random forest, boosting, lasso, and logistic regression models. Random forest and boosting models grow many classification trees. The trees vote for the assignment of each subject to one possible class. The forest outranks all the trees and chooses the class with the most votes.

[0167] For each of the four methods (random forest, boosting, lasso, and logistic regression), each method was allowed to select and rank its own best 15 transitions. We then built models with 1 to 15 transitions. Each method independently lowered the number of nodes, successively from 15 to 1. The recursive option was used to lower the number of nodes at each step. To determine which nodes to remove, the nodes were ranked at each step based on their importance from the nested cross-validation procedure. The least important nodes were eliminated. The importance measure for lasso and logistic regression is the z-score. For random forest and boosting, the importance of the variables was calculated by permuting the out-of-bag data. For each tree, the classification error rate on the out-of-bag portion of the data was recorded. The error rate was then recalculated after permuting the values ​​of each variable (i.e., transition). If the transitions were indeed important, there would have been a large difference between the two error rates. The difference between the two error rates was then averaged over all trees and normalized by the standard deviation of the difference. The AUC for these models is shown in Table 5, as estimated by 100 rounds of bootstrap resampling. Table 6 shows the top 15 transitions selected by each multivariate method, ranked by importance for that method. These multivariate analyses suggest that models combining three or more transitions give an AUC greater than 0.7 estimated by bootstrap.

[0168] In the multivariate models, the random forest (rf), boosting, and lasso models gave the best AUROC area under the curve. The following transitions were selected by these models as significant in the Cox univariate models and / or as having high univariate ROC: AFTECCVVASQLR_770.87_574.3 ELLESYIDGR_597.8_710.3 ITLPDFTGDLR_624.34_920.4 TDAPDLPEENQAR_728.34_613.3 SFRPFVPR_335.86_635.3

[0169] In summary, univariate and multivariate Cox analyses were performed using transitions to predict gestational age at birth (GAB), including gestational age at the date of specimen collection as a covariate. Five proteins with multiple transitions were identified in the univariate Cox analyses with p-values ​​less than 0.05: lipopolysaccharide-binding protein (LBP), prothrombin (THRB), complement component C5 (C5 or CO5), plasminogen (PLMN), and complement component C8 gamma chain (C8G or CO8G).

[0170] In the multivariate Cox analysis, 24 transitions were selected in the stepwise AIC variable analysis, and 3 transitions were selected in the lasso model, including the top 3 proteins in the univariate analysis. Univariate (AUROC) and multivariate (random forest, boosting, lasso, and logistic regression) analyses were performed to predict preterm birth as a binary categorical variable. In the univariate analysis, 63 analytes with an AUROC of 0.6 or more were identified. The multivariate analysis suggests that a model combining 3 or more transitions, estimated by bootstrap, gives an AUC greater than 0.7.

[0171] Example 3: Study to identify and validate preterm birth biomarkers II Further studies were conducted using essentially the same methods described in the preceding examples, unless otherwise noted below. In this study, two gestational age-matched controls were used for each of the 28 cases and 56 matched controls, all from the early pregnancy window only (17-22 weeks).

[0172] Samples were processed in 4 batches, each consisting of 7 cases, 14 matched controls, and 3 HGS controls. Serum samples were depleted of the 14 most abundant serum samples with MARS14 as described in Example 1. The depleted serum was then reduced with dithiothreitol, alkylated with iodoacetamide, and then digested overnight at 37°C with trypsin at a trypsin to protein ratio of 1:20. Following trypsin digestion, samples were desalted with Empore C18 96-well solid phase extraction plates (3M Company) and lyophilized. The desalted samples were redissolved in reconstitution solution containing five internal standard peptides.

[0173] LC-MS / MS analysis was performed using an Agilent Poroshell 120 EC-C18 column (2.1 × 50 mm, 2.7 μm) and an Agilent 6490 Elution was with an acetonitrile gradient into a Triple Quadrapole mass spectrometer.

[0174] Data analysis involved the use of conditional logistic regression, where each matched triplet (case and two matched controls) was a strata. The p-values ​​reported in the table indicate whether there was a significant difference between the cases and matched controls.

[0175] Table 7. Results of Study II [Table 7]

[0176] Example 4: Study III Shotgun Identification of Preterm Birth Biomarkers Further studies used a hypothesis-independent shotgun approach to identify and quantify additional biomarkers not present in our multiplexed hypothesis-dependent MRM assay. Samples were processed as described in the preceding examples unless noted below.

[0177] Tryptic digests of MARS-depleted patient (preterm cases and term controls) samples were fractionated by two-dimensional liquid chromatography and analyzed by tandem mass spectrometry. Sample aliquots equivalent to 3-4 μl of serum were injected onto a 6 cm × 75 μm self-packed strong cation exchange (Luna SCX, Phenomenex) column. Peptides were eluted from the SCX column with salts (15, 30, 50, 70, and 100% B, where B = 250 mM ammonium acetate, 2% acetonitrile, 0.1% formic acid in water) and bound to 0.5 μl of a C18-packed stem trap (Optimize Technologies, Inc.) consecutively for each salt elution and further fractionated on a 10 cm × 75 μm reversed-phase ProteoPep II PicoFrit column (New Objective). Peptides were eluted from the reversed-phase column using an acetonitrile gradient containing 0.1% formic acid and directly ionized in an LTQ-Orbitrap (ThermoFisher). For each scan, peptide parent ion masses were acquired in the Orbitrap at 60K resolution, and the top seven most abundant ions were fragmented in the LTQ to obtain peptide sequence information.

[0178] The parent and fragment ion data were used to search the human RefSeq database using the Sequest algorithm (Eng et al., J. Am. Soc. Mass Spectrom 1994;5:976-989) and the X!Tandem algorithm (Craig and Beavis, Bioinformatics 2004;20:1466-1467). The Sequest data were searched with a tolerance of 20 ppm for parent ions and 1 AMU for fragment ions. Two losses of trypsin cleavage were allowed and modifications included static carboxyamidomethylcysteine ​​and methionine oxidation. After searching, the data were filtered by charge state versus Xcorr score (charge +1 ≧ 1.5 Xcorr, charge +2 ≧ 2.0, charge +3 ≧ 2.5). Similar search parameters were used for X!tandem, except that the mass tolerance for fragment ions was 0.8 AMU and there was no Xcorr filter. Instead, each X!Tandem peptide spectral assignment was verified using the PeptideProphet algorithm (Keller et al., Anal. Chem 2002;74:5383-5392), and protein assignments were verified using the ProteinProphet algorithm (Nesvizhskii et al., Anal. Chem 2002;74:5383-5392). Data were filtered to include only peptide spectral matches with a PeptideProphet probability of 0.9 or greater. After compiling peptide and protein identifications, the spectral count data for each peptide were imported into DAnTE software (Polpitiya et al., Bioinformatics. 2008;24:1556-1558). Log-transformed data were centered around the mean and missing values ​​were filtered by requiring that a peptide must be identified in at least 4 cases and 4 controls. To determine the significance of the analytes, receiver operating characteristic (ROC) curves for each analyte are constructed and the true positive rate (sensitivity) is plotted as a function of the false positive rate (1 specificity) for different thresholds separating the SPTB and term groups.The area under the ROC curve (AUC) is equal to the probability that a classifier will rank a randomly chosen positive example higher than a randomly chosen negative example. Peptides with AUC greater than or equal to 0.6 uniquely found by Sequest or Xtandem are found in Tables 8 and 9, respectively, and peptides identified by both approaches are found in Table 10.

[0179] Table 8. Significant peptides for Sequest only (AUC>0.6) [Table 8-1] [Table 8-2] [Table 8-3] [Table 8-4] [Table 8-5] [Table 8-6] [Table 8-7] [Table 8-8] [Table 8-9] [Table 8-10] [Table 8-11] [Table 8-12] [Table 8-13] [Table 8-14] [Table 8-15]

[0180] Table 9. Significant peptides (AUC>0.6) for X!Tandem only [Table 9-1] [Table 9-2] [Table 9-3] [Table 9-4] [Table 9-5] [Table 9-6] [Table 9-7] [Table 9-8] [Table 9-9] [Table 9-10] [Table 9-11] [Table 9-12] [Table 9-13]

[0181] Table 10. Significant peptides (AUC>0.6) for both X!Tandem and Sequest [Table 10-1] [Table 10-2] [Table 10-3] [Table 10-4] [Table 10-5] [Table 10-6] [Table 10-7] [Table 10-8] [Table 10-9]

[0182] The differentially expressed proteins identified by the above hypothesis-independent strategy were not already present in our MRM-MS assay and were candidates for incorporation into the MRM-MS assay. Two additional proteins of functional interest (AFP, PGH1) were also selected for MRM development. Candidates were prioritized by AUC and biological function, with new pathways prioritized. The sequence for each protein of interest was imported into Skyline software, which generated a list of tryptic peptides, m / z values ​​for parent and fragment ions, and instrument-specific collision energies (McLean et al. Bioinformatics (2010) 26(7):966-968; McLean et al. Anal. Chem (2010) 82(24):10116-10124).

[0183] The list was refined by eliminating peptides containing cysteine ​​and methionine and using the shotgun data to select a subset of charge states and potential fragment ions for each peptide that had already been observed in the mass spectrometer.

[0184] After prioritizing parent and fragment ions, a list of transitions was exported with a single predicted collision energy. Approximately 100 transitions were added to a single MRM run. For development, MRM data was collected on either a QTRAP5500 (AB Sciex) or 6490QQQ (Agilent). Commercially available human female serum (from pregnant and non-pregnant donors) was depleted, processed into tryptic peptides as described above, and used to "scan" for peptides of interest. In some cases, purified synthetic peptides were used for further optimization. For development, digested serum or purified synthetic peptides were separated on a 2.1 x 50 mm Poroshell 120 EC-C18 column (Agilent) at 40°C with a 15 minute acetonitrile gradient at 100 μl / min.

[0185] The MS / MS data was imported back into Skyline, where all chromatograms for each peptide were overlaid and used to identify consensus peaks corresponding to the peptide of interest as well as transitions with the highest intensity and lowest noise. Table 11 contains a list of the most strongly observed candidate transitions and peptides for transfer to the MRM assay.

[0186] Table 11. Candidate peptides and transitions for transfer to MRM assay [Table 11-1] [Table 11-2] [Table 11-3] [Table 11-4] [Table 11-5] [Table 11-6] [Table 11-7] [Table 11-8] [Table 11-9] [Table 11-10]

[0187] The top 2-10 transitions per peptide and up to 7 peptides per protein were then selected for collision energy (CE) optimization on an Agilent 6490. Using Skyline or MassHunter Qual software, the optimized CE value for each transition was determined based on peak area or signal to noise. Two transitions with the largest peak area per peptide and at least two peptides per protein were chosen for the final MRM method. Replacement of the transition with the lower peak area was performed if the transition with the larger peak area had a high background level or a low m / z value with more potential for interference.

[0188] Finally, the retention times of selected peptides were mapped using the same column and gradient as our established sMRM assay. The newly discovered analytes were then added to the sMRM method and used in further hypothesis-driven discovery studies, described in Example 5 below.

[0189] The above method was typical for most proteins. However, in some cases, differentially expressed peptides identified in the shotgun method did not uniquely identify the protein, for example, in protein families with high sequence identity. In these cases, MRM methods were developed for each family member. Also, note that for any given protein, peptides found to be significant plus fragment ions not observed in the Orbitrap could have been included in the MRM optimization and added to the final sMRM method if they produced the best signal intensity.

[0190] Example 5: Study IV to identify and validate preterm birth biomarkers Further hypothesis-driven discovery studies were performed using the scheduled MRM assay used in Example 3, but this time augmented with the newly discovered analytes from Example 4. Less robust transitions (from the original 1708 described in Example 1) were removed to improve analytical performance and make space for the newly discovered analytes. Samples included approximately 30 cases and 60 matched controls from each of three gestational ages (early, 17-22 weeks; mid, 23-25 ​​weeks and late, 26-28 weeks). Log-transformed peak areas for each transition were corrected for run order and batch effects by regression. The ability of each analyte to separate cases and controls was determined by calculating univariate AUC values ​​from the ROC curves. Ranked univariate AUC values ​​(0.6 or greater) are reported for the individual gestational age window sample sets (Tables 12, 13, 15) and the mid and late window combination (Table 14). Multivariate classifiers were constructed using different subsets of analytes (described below) by lasso and random forest methods. Significant transitions in lasso corresponded to transitions with non-zero coefficients, and random forest analyte rankings were determined by Gini importance values ​​(average decrease in model accuracy when that variable was removed). We report all analytes with non-zero lasso coefficients (Tables 16-32) and the top 30 analytes from each random forest analysis (Tables 33-49). Models were constructed considering the top 32 or 100 univariate analytes, the top 50 proteins or a single best univariate analyte for all analytes. Finally, 1000 rounds of bootstrap resampling were performed and non-zero lasso coefficients or random forest Gini importance values ​​were summed for each analyte among panels with an AUC of 0.85 or above.

[0191] Table 12. Individual statistics for early windows [Table 12-1] [Table 12-2] [Table 12-3] [Table 12-4] [Table 12-5] [Table 12-6] [Table 12-7] [Table 12-8]

[0192] Table 13. Individual statistics for medium-term windows [Table 13-1] [Table 13-2] [Table 13-3]

[0193] Table 14. Individual statistics for middle and late stages [Table 14-1] [Table 14-2]

[0194] Table 15. Late window individual statistics [Table 15-1] [Table 15-2] [Table 15-3] [Table 15-4]

[0195] Table 16. Lasso Early 32 [Table 16]

[0196] Table 17. Lasso Early 100 [Table 17]

[0197] Table 18. Lasso Protein Early Windows [Table 18]

[0198] Table 19. Lasso Full Early Window [Table 19]

[0199] Table 20. Lasso SummedCoef Early Window [Table 20-1] [Table 20-2] [Table 20-3] [Table 20-4] [Table 20-5] [Table 20-6]

[0200] Table 21. Lasso 32 mid-term window [Table 21]

[0201] Table 22. Lasso 100 mid-term windows [Table 22]

[0202] Table 23. Lasso Protein Metaphase Windows [Table 23]

[0203] Table 24. Lasso Full Midterm Windows [Table 24]

[0204] Table 25. Lasso 32 mid-late window [Table 25]

[0205] Table 26. Lasso 100 mid-late window [Table 26]

[0206] Table 27. Lasso protein mid-late window [Table 27]

[0207] Table 28. Lasso full mid-late window [Table 28]

[0208] Table 29. Lasso 32 Late Window [Table 29]

[0209] Table 30: Lasso 100 Late Window [Table 30]

[0210] Table 31: Lasso Protein Late Window [Table 31]

[0211] Table 32: Lasso Full Late Window [Table 32]

[0212] Table 33: Random Forest 32 Early Windows [Table 33]

[0213] Table 34. Random Forest 100 Early Windows [Table 34]

[0214] Table 35. Random Forest Protein Early Windows [Table 35]

[0215] Table 36. Random Forest All Early Windows [Table 36]

[0216] Table 37. Random Forest SummedGini Early Windows [Table 37-1] [Table 37-2] [Table 37-3] [Table 37-4] [Table 37-5] [Table 37-6]

[0217] Table 38. Random Forest 32 Medium Term Windows [Table 38]

[0218] Table 39. Random Forest 100 Mid-term Windows [Table 39]

[0219] Table 40. Random Forest Protein Mid-term Windows [Table 40]

[0220] Table 41. Random Forest All Mid-Term Windows [Table 41]

[0221] Table 42. Random Forest 32 mid-late windows [Table 42]

[0222] Table 43. Random Forest 100 mid-late windows [Table 43]

[0223] Table 44. Random Forest Protein Mid-Late Windows [Table 44]

[0224] Table 45. Random Forest All Mid-Late Windows [Table 45]

[0225] Table 46. Random Forest 32 Late Window [Table 46]

[0226] Table 47. Random Forest 100 Late Windows [Table 47]

[0227] Table 48. Random Forest Protein Late Windows [Table 48]

[0228] Table 49. Random Forest All Late Windows [Table 49]

[0229] Table 50. Selected transitions for early windows [Table 50]

[0230] Table 51. Selected proteins for early window [Table 51]

[0231] Table 52. Selected transitions for mid-late windows [Table 52]

[0232] Table 53. Selected proteins for the mid-late window [Table 53]

[0233] Example 6: Study V to further refine preterm birth biomarkers An additional hypothesis-dependent discovery study was performed with a further improved, scheduled MRM assay. Less robust transitions were again removed, improving analytical performance and making space for the inclusion of stable isotope-labeled standards (SIS) corresponding to the 79 analytes of interest identified in the previous study. SIS peptides have identical amino acid sequences, chromatographic and MS fragmentation behaviors as their endogenous peptide counterparts, but at different masses. They can therefore be used to reduce LC-MS analytical variability and confirm analyte identity. Samples included approximately 60 spontaneous PTB cases (delivery at less than 37 weeks, 0 days) and 180 term controls (delivery greater than or equal to 37 weeks, 0 days). Each case was assigned a "matched" control within 1 day of blood collection and two "random" controls matched to the same 3-week blood collection window (17-19, 20-22, or 23-25 ​​weeks of gestation). For analytical purposes, these three blood collection windows were combined. In this study, samples were processed essentially as described previously, except that tryptic digests were reconstituted in a solution containing SIS standards. Raw analyte peak areas were Box-Cox transformed, corrected for run order and batch effects by regression, and used for univariate and multivariate statistical analysis. Univariate analysis included determination of p-values ​​for adjusted peak areas for all analytes from t-tests considering cases vs. controls defined as either birth at >37 weeks (Table 54) or birth at >40 weeks (Table 55). Univariate analysis also included determination of p-values ​​for linear models assessing the dependence of adjusted peak areas for each analyte on time to delivery (gestational age at birth minus gestational age at blood draw) (Table 56) and gestational age at birth (Table 57). Additional raw peak area ratios were calculated for endogenous analytes and their corresponding SIS counterparts, Box-Cox transformed, and then used for univariate and multivariate statistical analysis. The above univariate analysis was repeated for the analyte / SIS peak area ratio values ​​and summarized in Tables 58-61, respectively.

[0234] Multivariate random forest regression models were constructed using analyte values ​​and clinical variables (e.g., maternal age (MAGE), body mass index (BMI)) to predict gestational age at birth (GAB). The accuracy of the random forest was evaluated in terms of the correlation of predicted and actual GAB, as well as in terms of the mean absolute deviation (MAD) of predicted GAB from actual GAB. Accuracy was further evaluated by determining the area under the receiver operating characteristic curve (AUC) when using predicted GAB as a quantitative variable for classifying subjects as full-term or preterm. Random forest importance values ​​were fitted to an empirical cumulative distribution function and probabilities (P) were calculated. We report analytes by their importance ranking (P>0.7) in the random forest model using adjusted analyte peak area values ​​(Table 62) and analyte / SIS peak area ratio values ​​(Table 63).

[0235] The probability of preterm birth, p(PTB), can be estimated using the predicted gestational age at birth (GAB) as follows: Estimates are based on women enrolled in the Sera PAPR clinical trial, which provided the subjects used to develop the PTB prediction method.

[0236] Among women with predicted GAB of j days plus or minus k days, p(PTB) was estimated as the proportion of women in the PAPR clinical trial with predicted GAB of j days plus or minus k days who actually delivered before 37 weeks' gestation.

[0237] More generally, for women with a predicted GAB of j days plus or minus k days, the probability p (actual GAB < specified GAB) that the actual gestational age at birth would be less than the specified gestational age was estimated as the proportion of women in the PAPR clinical trials with a predicted GAB of j days plus or minus k days who delivered before the actual specified gestational age. Figure 1 shows a scatter plot of actual gestational age at birth versus predicted gestational age from the random forest regression model. Figure 2 shows the distribution of predicted gestational age from the random forest regression model versus actual gestational age at birth (GAB), where the actual GAB was given in the categories (i) less than 37 weeks, (ii) 37-39 weeks, and (iii) 40 weeks or greater.

[0238] Table 54. Univariate p-values ​​for adjusted peak areas (<37 weeks vs. >37 weeks) [Table 54-1] [Table 54-2]

[0239] Table 55. Univariate p-values ​​for adjusted peak areas (<37 weeks vs. >40 weeks) [Table 55-1] [Table 55-2]

[0240] Table 56. Univariate p-values ​​for adjusted peak areas in linear models of time to birth [Table 56]

[0241] Table 57. Univariate p-values ​​for adjusted peak areas in linear models of gestational age at birth [Table 57-1] [Table 57-2] [Table 57-3]

[0242] Table 58. Univariate p-values ​​for peak area ratios (<37 weeks vs. >37 weeks) [Table 58]

[0243] Table 59. Univariate p-values ​​for peak area ratios (<37 weeks vs. >40 weeks) [Table 59]

[0244] Table 60. Univariate p-values ​​for peak area ratios in linear models of time to birth [Table 60-1] [Table 60-2]

[0245] Table 61. Univariate p-values ​​for peak area ratios in linear models of gestational age at birth [Table 61]

[0246] Table 62. Random Forest Importance Values ​​with Adjusted Peak Areas [Table 62-1] [Table 62-2] [Table 62-3] [Table 62-4]

[0247] Table 63. Random Forest Importance Values ​​Using Peak Area Ratios [Table 63-1] [Table 63-2]

[0248] From the foregoing, it will be apparent that variations and modifications can be made to the invention described herein to adapt it to various usages and conditions, such embodiments also falling within the scope of the following claims.

[0249] The recitation of a list of elements in any definition of a variable herein includes definitions of that variable as any single element or combination (or subcombination) of the listed elements. The recitation of an embodiment herein includes that embodiment as any single embodiment or in combination with any other embodiment or portion thereof.

[0250] All patents and publications mentioned in this specification are herein incorporated by reference to the same extent as if each individual patent and publication was specifically and individually indicated to be incorporated by reference.

Claims

[Claim 1] Compositions and methods for predicting the probability of premature birth in pregnant women.

Images