Biomarkers and methods for predicting premature birth

Abstract

To provide a biomarker panel, method, and kit for determining the probability of premature birth in pregnant women. [Solution] This disclosure is partly based on the finding that certain proteins and peptides in biological samples obtained from pregnant women are expressed differently in pregnant women who have an increased risk of developing preterm birth in the future or who are currently suffering from preterm birth, compared to corresponding controls. This disclosure is further partly based on the unexpected finding that a panel of one or more of these proteins and peptides can be used in a method for determining the probability of preterm birth in pregnant women with relatively high sensitivity and specificity.

Description

[Technical Field]

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

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

[0003] background According to the World Health Organization, an estimated 15 million infants are born prematurely each year (before 37 weeks of gestation). Premature birth rates are increasing in almost all countries 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, ISBN 9789241503433 (2012). An estimated one million infants die each year from complications of preterm birth. Globally, preterm birth is the leading cause of neonatal death (infants within the first four weeks of life) and the second leading cause of death in children under five years of age, following pneumonia. Many survivors face a life of disability, including learning disabilities and visual 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):2162-2172 (2012)). Although over 60% of preterm births occur in Africa and South Asia, preterm birth remains a global problem. Countries with the highest rates include Brazil, India, Nigeria, and the United States. Of the 11 countries with preterm birth rates exceeding 15%, all but two are in sub-Saharan Africa. In the poorest countries, on average, 12% of infants are born too prematurely, compared to 9% in high-income countries. Within a country, the risk is higher in impoverished families. More than three-quarters of premature babies can be helped with feasible, cost-effective care, such as prenatal steroid injections given to pregnant women at risk of premature labor to strengthen the infant's lungs.

[0005] Premature babies are at greater risk than full-term babies of death and various health and developmental problems. Complications include acute respiratory, digestive, immune, central nervous system, hearing, and vision problems, as well as long-term motor, cognitive, visual, auditory, behavioral, socio-emotional, health, and growth problems. The birth of a premature baby also incurs significant emotional and financial costs for the family 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 babies born in the earliest stages of pregnancy. However, babies born closer to full term represent the largest number of premature babies and experience more complications than full-term babies.

[0006] In women less than 24 weeks pregnant who show cervical opening on ultrasound, a surgical procedure known as cervical cerclage, which involves suturing the cervix closed with strong sutures, may be used to prevent preterm birth. Women less than 34 weeks pregnant and those experiencing active premature labor may require hospitalization and administration of medication to temporarily halt premature labor and / or promote fetal lung development. If a pregnant woman is determined to be at risk of preterm birth, healthcare providers may implement a variety of clinical strategies, which may include prophylactic oral medications such as hydroxyprogesterone caproate (Makena) injections and / or vaginal progesterone gel, cervical pessaries, restrictions on sexual 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 of preterm birth and provide them with appropriate prenatal care. For women identified as high-risk, more intensive prenatal surveillance and preventative interventions can be planned. Current strategies for risk assessment are based on obstetric and medical history and clinical examinations, but these strategies can only identify a small percentage of women at risk of preterm birth. Reliable early identification of risk for preterm birth would enable the planning of appropriate monitoring and clinical management to prevent preterm birth. Such monitoring and management may include more frequent prenatal visits, continuous measurement of cervical length, enhanced education on signs and symptoms of preterm birth, lifestyle interventions for modifiable risk behaviors, and cervical pessaries and progesterone treatment. Finally, reliable prenatal identification of risk for preterm birth is also crucial for the cost-effective allocation of monitoring resources. This invention addresses this need by providing a composition and method for determining whether a pregnant woman is at risk of premature birth. 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(no.9832):2162-2172 (2012). [Overview of the Initiative] [Problems that the invention aims to solve]

[0009] This invention provides a composition and method for predicting the probability of premature birth in pregnant women. [Means for solving the problem]

[0010] Abstract In one embodiment, the present invention provides a panel of isolated biomarkers comprising N 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 AFTECCVVASQLR, ELLESYIDGR, and ITLPDFTGDLR. In additional embodiments, the biomarker panel comprises at least two isolated biomarkers selected from the group consisting of FLNWIK, FGFGGSTDSGPIR, LLELTGPK, VEHSDLSFSK, IEGNLIFDPNNYLPK, ALVLELAK, TQILEWAAER, DVLLLVHNLPQNLPGYFWYK, SEPRPGVLLR, ITQDAQLK, ALDLSLK, WWGGQPLWITATK, and LSETNR.

[0011] In a further embodiment, the biomarker panel includes 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 embodiment, the present invention provides a panel of isolated biomarkers comprising N biomarkers from among those 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 present 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 present invention provides a biomarker panel comprising at least two isolated biomarkers selected from the group consisting of alpha-1B glycoprotein (A1BG), 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).

[0015] In other embodiments, the present invention provides a biomarker panel comprising 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 somatomammotropin hormone (CSH), and angiotensinogen (ANG or ANGT).

[0016] In other embodiments, the present invention provides a biomarker panel comprising alpha-1B glycoprotein (A1BG), 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).

[0017] In an additional embodiment, the present 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] Furthermore, the present invention also provides a method for determining the probability of premature birth in a pregnant woman, comprising detecting measurable features 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 determine the probability of premature birth in the pregnant woman. In some embodiments, the present invention provides a method for predicting GAB, comprising detecting measurable features 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.

[0019] In some embodiments, the measurable features include fragments or derivatives of each of N biomarkers selected from the biomarkers listed in Tables 1 to 63. In some embodiments of the disclosed method, detecting measurable features involves quantifying the amount of each of N biomarkers, combinations or parts and / or derivatives thereof, selected from the biomarkers listed in Tables 1 to 63, in a biological sample obtained from a pregnant woman. In additional embodiments, the disclosed method for determining the probability of preterm birth in a pregnant woman further encompasses detecting measurable features for one or more risk indicators associated with preterm birth.

[0020] In some embodiments, the disclosed methods for determining the probability of preterm birth in a pregnant female and the related methods disclosed herein include detecting the measurable characteristics of each of N biomarkers, where N is selected from the group consisting of 2 to 24. In further embodiments, the disclosed methods for determining the probability of preterm birth in a pregnant female and the related methods disclosed herein include detecting the measurable characteristics of each of at least two isolated biomarkers selected from the group consisting of AFTECCVVASQLR, ELLESYIDGR, and ITLPDFTGDLR. In further embodiments, the disclosed methods for determining the probability of preterm birth in a pregnant female and the related methods disclosed herein include detecting the measurable characteristics of each of at least two isolated biomarkers selected from the group consisting of FLNWIK, FGFGGSTDSGPIR, LLELTGPK, VEHSDLSFSK, IEGNLIFDPNNYLPK, ALVLELAK, TQILEWAAER, DVLLLVHNLPQNLPGYFWYK, SEPRPGVLLR, ITQDAQLK, ALDLSLK, WWGGQPLWITATK, and LSETNR. In further embodiments, the disclosed methods for determining the probability of preterm birth in a pregnant female and the related methods disclosed herein include detecting the measurable characteristics 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 for determining the probability of preterm birth in a pregnant female includes detecting the measurable characteristics 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 for determining the probability of preterm birth in a pregnant woman comprises detecting the measurable characteristics of each of at least two isolated biomarkers selected from the group consisting of alpha-1B-glycoprotein (A1BG), 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).

[0023] In other embodiments, the disclosed method for determining the probability of preterm birth in a pregnant woman comprises detecting the measurable characteristics 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 somatomammotropin hormone (CSH), and angiotensinogen (ANG or ANGT).

[0024] In a further embodiment, the disclosed method for determining the probability of preterm birth in a pregnant woman comprises detecting the measurable characteristics 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 methods for determining the probability of preterm birth in pregnant women, 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 method for determining the probability of preterm birth involves detecting and / or quantifying one or more biomarkers using mass sprectrometry, a capture agent, or a combination thereof.

[0026] In some embodiments, the disclosed method for determining the probability of preterm birth in a pregnant woman comprises a first step of preparing a biomarker panel containing N biomarkers listed in Tables 1 to 63. In additional embodiments, the disclosed method for determining the probability of preterm birth in a pregnant woman comprises a first step of preparing a biological sample from the pregnant woman.

[0027] In some embodiments, the disclosure method for determining the probability of preterm birth in a pregnant woman includes informing the healthcare provider of the probability. In additional embodiments, the communication informs the pregnant woman of the decision regarding subsequent treatment. In further embodiments, the decision regarding one or more treatments is selected from a group consisting of more frequent prenatal visits, continuous measurement of cervical length, enhanced education on signs and symptoms of preterm birth, lifestyle interventions for modifiable risk behaviors, and progesterone treatment.

[0028] In further embodiments, a disclosed method for determining the probability of preterm birth in a pregnant woman includes analyzing measurable features of one or more isolated biomarkers using a predictive model. In some embodiments of the disclosed method, the measurable features of one or more isolated biomarkers are compared to reference features.

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

[0030] In some embodiments, the present invention provides a method for determining the probability of premature 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 for predicting GAB, comprising determining the predicted GAB birth in the pregnant woman, which includes: (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 or thresholding them; and (c) 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 for predicting the time to delivery 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) determining the predicted GAB in the pregnant woman, including multiplying the amounts by a predetermined coefficient or thresholding them; (d) adding the individual products to obtain a total risk score corresponding to the predicted GAB; and (e) predicting the time to delivery in the pregnant woman by subtracting the estimated gestational age (GA) at the time the biological sample was obtained from the predicted GAB.

[0033] Other features and advantages of the present invention will become apparent from the detailed description and the claims. In embodiments of the present invention, for example, the following items are provided. (Item 1) A panel of isolated biomarkers containing N biomarkers from those listed in Tables 1 through 63. (Item 2) The panel described in item 1, where N is a number selected from the group consisting of 2 to 24. (Item 3) The panel according to item 2, comprising 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 4) The panel described in item 2, containing 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 described in 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 described in item 2 includes 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 somatomammotropin hormone (CSH), angiotensinogen (ANG or ANGT), the biomarkers shown in Table 51, and the biomarkers shown in Table 53. (Item 7) A method for determining the probability of premature birth in a pregnant woman, comprising detecting measurable characteristics 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 characteristics to determine the probability of premature birth in the pregnant woman. (Item 8) The method according to item 7, 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 9) The method according to item 7, wherein detecting measurable features includes quantifying the amount of each of N biomarkers, combinations thereof, or parts and / or derivatives thereof, selected from the biomarkers listed in Tables 1 to 63, in a biological sample obtained from the pregnant woman. (Item 10) The method according to item 9, further comprising calculating the probability of preterm birth in the pregnant woman based on the quantified amount of each of the N biomarkers selected from the biomarkers listed in Tables 1 to 63. (Item 11) The method of item 7, further comprising the first step of preparing a biomarker panel containing N biomarkers from among those listed in Tables 1 to 63. (Item 12) The method of item 7, further comprising the first step of preparing a biological sample from the aforementioned pregnant woman. (Item 13) The method of item 7, further comprising informing healthcare providers of the aforementioned probability. (Item 14) The method described in item 13, which, based on the aforementioned contact, notifies the pregnant woman of the decision regarding subsequent treatment. (Item 15) The method according to item 7, wherein N is a number selected from the group consisting of 2 to 24. (Item 16) The method according to item 15, wherein the N biomarkers include 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) The method described in item 7, which includes the use of a predictive model. (Item 18) The method of item 17, wherein the analysis includes comparing the measurable features with reference features. (Item 19) The method described in item 18, wherein the analysis involves using one or more selected from the group consisting of linear discriminant analysis models, support vector machine classification algorithms, recursive feature exclusion models, predictive analysis of microarray models, logistic regression models, multiple regression models, survival models, CART algorithms, flextree algorithms, LART algorithms, random forest algorithms, MART algorithms, machine learning algorithms, penalized regression methods, and combinations thereof. (Item 20) The method described in item 19, which includes logistic regression. (Item 21) The method described in item 7, wherein the aforementioned probability is expressed as a risk score. (Item 22) The method according to item 7, wherein the biological sample is selected from the group consisting of whole blood, plasma, and serum. (Item 23) The method according to item 22, wherein the biological sample is serum. (Item 24) The method described in item 7, wherein the quantification includes mass spectrometry (MS). (Item 25) The method according to item 24, wherein the MS includes liquid chromatography-mass spectrometry (LC-MS). (Item 26) The method according to item 24, wherein the MS includes multiple reaction monitoring (MRM) or selective reaction monitoring (SRM). (Item 27) The method according to item 26, wherein the aforementioned MRM (or SRM) includes a scheduled MRM (SRM). (Item 28) The method described in item 7, wherein the quantification includes an assay using a capture agent. (Item 29) The method according to item 28, wherein the scavenger is selected from the group consisting of antibodies, antibody fragments, nucleic acid-based protein-binding reagents, small molecules, or variants thereof. (Item 30) The method according to item 28, wherein the assay is selected from the group consisting of enzyme immunoassay (EIA), enzyme-linked immunosorbent assay (ELISA), and radioimmunoassay (RIA). (Item 31) The method described in item 30, wherein the quantification further includes mass spectrometry (MS). (Item 32) The method described in item 31, wherein the quantification includes co-immunoprecipitation-mass spectrometry (co-IP MS). (Item 33) The method according to item 7, further comprising detecting measurable features for one or more risk indicators. (Item 34) The method of item 7, wherein the analysis of the measurable features first includes predicting the general age at birth (GAB) before determining the probability of preterm birth. (Item 35) The method of item 34, wherein the probability of preterm birth is determined using the prediction of the GAB. (Item 36) The method according to item 33, wherein one or more of the aforementioned risk indicators are selected from the group consisting of age, previous pregnancy, previous history of low birth weight or premature birth, multiple spontaneous second-term abortions, previous induced first-term abortions, familial and intergenerational factors, history of infertility, nulliparity, placental abnormalities, cervical and uterine abnormalities, pregnancy bleeding, intrauterine growth restriction, intrauterine diethylstilbestrol exposure, multiple pregnancy, sex of the infant, short stature, low pre-pregnancy weight / low body mass index, diabetes mellitus, hypertension, hypothyroidism, asthma, education level, tobacco use, and genitourinary tract infections. (Item 37) A method for determining the probability of premature 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. (Item 38) A method for predicting GAB, comprising detecting measurable features 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) The method of item 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) The method according to item 38, wherein detecting measurable features involves quantifying the amount of each of N biomarkers, combinations thereof, or parts and / or derivatives thereof, selected from the biomarkers listed in Tables 1 to 63, in a biological sample obtained from the pregnant woman. (Item 41) The method of item 40, 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 42) The method of item 38, further comprising the initial step of preparing a biomarker panel containing N biomarkers from Tables 1 to 63. (Item 43) The method of item 38, further comprising the initial step of preparing a biological sample from the aforementioned pregnant woman. (Item 44) The method of item 38, further comprising informing healthcare providers of the aforementioned probability. (Item 45) The method described in item 44, which, in response to the aforementioned contact, notifies the pregnant woman of the decision regarding subsequent treatment. (Item 46) The method described in item 38, wherein N is a number selected from the group consisting of 2 to 24. (Item 47) The method according to item 46, wherein the N biomarkers include 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 48) The aforementioned analysis is the method described in item 38, which includes the use of a predictive model. (Item 49) The method of item 48, wherein the analysis includes comparing the measurable features with reference features. (Item 50) The method described in item 49, wherein the analysis involves using one or more selected from the group consisting of linear discriminant analysis models, support vector machine classification algorithms, recursive feature exclusion models, predictive analysis of microarray models, logistic regression models, multiple regression models, survival models, CART algorithms, flextree algorithms, LART algorithms, random forest algorithms, MART algorithms, machine learning algorithms, penalized regression methods, and combinations thereof. (Item 51) The method described in item 50, wherein the analysis includes a random forest algorithm. (Item 52) The method described in item 38, wherein the aforementioned probability is expressed as a risk score. (Item 53) The method according to item 38, wherein the biological sample is selected from the group consisting of whole blood, plasma, and serum. (Item 54) The method according to item 53, wherein the biological sample is serum. (Item 55) The method described in item 38, wherein the quantification includes mass spectrometry (MS). (Item 56) The method according to item 55, wherein the MS includes liquid chromatography-mass spectrometry (LC-MS). (Item 57) The method according to item 55, wherein the MS includes multiple reaction monitoring (MRM) or selective reaction monitoring (SRM). (Item 58) The method according to item 57, wherein the aforementioned MRM (or SRM) includes a scheduled MRM (SRM). (Item 59) The method described above, which includes an assay using a capture agent, as described in item 38. (Item 60) The method according to item 59, wherein the scavenger is selected from the group consisting of antibodies, antibody fragments, nucleic acid-based protein-binding reagents, small molecules, or variants thereof. (Item 61) The method according to item 59, wherein the assay is selected from the group consisting of enzyme immunoassay (EIA), enzyme-linked immunosorbent assay (ELISA), and radioimmunoassay (RIA). (Item 62) The method according to item 61, wherein the quantification further includes mass spectrometry (MS). (Item 63) The method described in item 62, wherein the quantification includes co-immunoprecipitation-mass spectrometry (co-IP MS). (Item 64) The method of item 38, further comprising detecting measurable features for one or more risk indicators. (Item 65) The method according to item 64, wherein one or more of the aforementioned risk indicators are selected from the group consisting of age, previous pregnancy, previous history of low birth weight or premature birth, multiple spontaneous second-term abortions, previous induced first-term abortions, familial and intergenerational factors, history of infertility, nulliparity, placental abnormalities, cervical and uterine abnormalities, pregnancy bleeding, intrauterine growth restriction, intrauterine diethylstilbestrol exposure, multiple pregnancy, sex of the infant, short stature, low pre-pregnancy weight / low body mass index, diabetes mellitus, hypertension, hypothyroidism, asthma, education level, tobacco use, and genitourinary tract infections. (Item 66) A method for predicting GAB, comprising determining a predicted GAB birth in a pregnant woman, which includes: (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 / or thresholding them; and (c) adding the individual products to obtain a total risk score corresponding to the predicted GAB. (Item 67) A method for predicting the time to delivery 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) determining the predicted GAB in the pregnant woman, including multiplying the amounts by a predetermined coefficient and / or thresholding them; (d) adding the individual products to obtain a total risk score corresponding to the predicted GAB; and (e) predicting the time to delivery in the pregnant woman by subtracting the estimated GA at the time the biological sample was obtained from the predicted GAB. (Item 68) A method for determining the probability of full-term delivery in a pregnant woman, comprising detecting measurable characteristics 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 characteristics to determine the probability of full-term delivery in the pregnant woman. (Item 69) The panel according to item 2, comprising at least two of the isolated biomarkers selected from the group consisting of FLNWIK, FGFGGSTDSGPIR, LLELTGPK, VEHSDLSFSK, IEGNLIFDPNNYLPK, ALVLELAK, TQILEWAAER, DVLLLVHNLPQNLPGYFWYK, SEPRPGVLLR, ITQDAQLK, ALDLSLK, WWGGQPLWITATK, and LSETNR. (Item 70) The panel described in item 2 includes alpha-1B glycoprotein (A1BG), 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 71) The panel described in item 2 includes at least two isolated biomarkers selected from the group consisting of alpha-1B-glycoprotein (A1BG), 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) The method according to item 38, wherein the biomarker comprises at least two isolated biomarkers selected from the group consisting of FLNWIK, FGFGGSTDSGPIR, LLELTGPK, VEHSDLSFSK, IEGNLIFDPNNYLPK, ALVLELAK, TQILEWAAER, DVLLLVHNLPQNLPGYFWYK, SEPRPGVLLR, ITQDAQLK, ALDLSLK, WWGGQPLWITATK, and LSETNR. (Item 73) The method according to item 66, wherein the biomarker comprises at least two isolated biomarkers selected from the group consisting of FLNWIK, FGFGGSTDSGPIR, LLELTGPK, VEHSDLSFSK, IEGNLIFDPNNYLPK, ALVLELAK, TQILEWAAER, DVLLLVHNLPQNLPGYFWYK, SEPRPGVLLR, ITQDAQLK, ALDLSLK, WWGGQPLWITATK, and LSETNR. (Item 74) The method according to item 67, wherein the biomarker comprises at least two isolated biomarkers selected from the group consisting of FLNWIK, FGFGGSTDSGPIR, LLELTGPK, VEHSDLSFSK, IEGNLIFDPNNYLPK, ALVLELAK, TQILEWAAER, DVLLLVHNLPQNLPGYFWYK, SEPRPGVLLR, ITQDAQLK, ALDLSLK, WWGGQPLWITATK, and LSETNR. (Item 75) The method according to item 68, wherein the biomarker comprises at least two isolated biomarkers selected from the group consisting of FLNWIK, FGFGGSTDSGPIR, LLELTGPK, VEHSDLSFSK, IEGNLIFDPNNYLPK, ALVLELAK, TQILEWAAER, DVLLLVHNLPQNLPGYFWYK, SEPRPGVLLR, ITQDAQLK, ALDLSLK, WWGGQPLWITATK, and LSETNR. (Item 76) The method according to item 38, wherein the biomarker comprises at least two isolated biomarkers selected from the group consisting of alpha-1B-glycoprotein (A1BG), 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) The method according to item 66, wherein the biomarker comprises at least two isolated biomarkers selected from the group consisting of alpha-1B-glycoprotein (A1BG), 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) The method according to item 67, wherein the biomarker comprises at least two isolated biomarkers selected from the group consisting of alpha-1B-glycoprotein (A1BG), 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) The method according to item 68, wherein the biomarker comprises at least two isolated biomarkers selected from the group consisting of alpha-1B-glycoprotein (A1BG), 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 explanation of the drawing]

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

[0035] [Figure 2]Figure 2 shows the distribution of predicted gestational age versus actual gestational age at birth (GAB) from a random forest regression model, where the actual GAB is categorized into (i) less than 37 weeks, (ii) 37 to 39 weeks, and (iii) 40 weeks or more (peaks are from left to right, respectively). [Modes for carrying out the invention]

[0036] Detailed explanation This disclosure is partly based on the finding that certain proteins and peptides in biological samples obtained from pregnant women are expressed differently in pregnant women with an increased risk of preterm birth compared to controls. Furthermore, this disclosure is partly based on the unexpected finding that a panel of one or more of these proteins and peptides can be used 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 general age at birth (GAB), predicting time to delivery, and / or monitoring the progress of prophylactic treatment in pregnant women, either individually or within a panel of biomarkers.

[0037] This disclosure provides a biomarker panel, method, and kit for determining the probability of preterm birth in pregnant women. One key advantage of this disclosure is that the risk of developing preterm birth can be assessed in early pregnancy, allowing for timely initiation of appropriate monitoring and clinical management to prevent premature birth. This 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, this disclosure includes a method for obtaining a dataset associated with a sample to generate useful results in determining the probability of preterm birth in pregnant women (wherein the dataset includes at least quantitative data on biomarkers and panels of biomarkers identified as predictors of preterm birth), and for inputting the dataset into an analytical process using the dataset to generate useful results in determining the probability of preterm birth in pregnant women. As further described below, this quantitative data may include amino acids, peptides, polypeptides, proteins, nucleotides, nucleic acids, nucleosides, sugars, fatty acids, steroids, metabolites, carbohydrates, lipids, hormones, antibodies, regions of interest that act as surrogates for biological macromolecules, and combinations thereof.

[0039] For example, in addition to the specific biomarkers identified herein by accession numbers in public databases, sequences, or references, the Invention also considers the use of biomarker variants that are at least 90%, at least 95%, or at least 97% identical to exemplary sequences that are currently known or to be discovered later and have utility for the methods of the Invention. These variants may exhibit polymorphisms, splice variants, mutations, etc. In this regard, in the context of the Invention, this Specification discloses several proteins known in the Art and provides exemplary accession numbers associated with one or more public databases and exemplary references to published journal articles relating to these proteins known in the Art. However, those skilled in the art will understand that additional accession numbers and journal articles that may provide additional features of the disclosed biomarkers can be easily identified, and that exemplary references are by no means limiting with respect to the disclosed biomarkers. As described herein, various techniques and reagents can be found to be used in the methods of the Invention. Suitable samples in the context of the 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 by 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 combining 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, this disclosure further includes biomarker variants that are approximately 90%, approximately 95%, or approximately 97% identical to the exemplary sequences. Variants include polymorphisms, splice variants, mutations, etc., as used herein.

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

[0042] This specification provides a panel of isolated biomarkers containing N biomarkers selected from the groups listed in Tables 1 to 63. In the disclosed panel of biomarkers, N may be a number selected from the groups from 2 to 24. In the disclosed method, the number of biomarkers detected and whose levels are determined may be 1 or greater 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 a particular embodiment, the number of biomarkers detected and whose levels are determined may be 1 or greater than 1, for example, 2, 3, 4, 5, 6, 7, 8, 9, 10 or more. The method of this disclosure is useful for determining the probability of preterm birth in pregnant women.

[0043] While some of the biomarkers listed in Tables 1 to 63 are useful only for determining the probability of preterm birth in pregnant women, methods are also described herein for grouping multiple subsets of biomarkers, each useful as a panel of three or more biomarkers. In some embodiments, the present 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 yet 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 yet 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 yet another embodiment, N is selected to be any number from 5 to 10, 5 to 15, 5 to 20, or 5 to 23. In yet another embodiment, N is selected to be any number from 6 to 10, 6 to 15, 6 to 20, or 6 to 23. In yet another embodiment, N is selected to be any number from 7 to 10, 7 to 15, 7 to 20, or 7 to 23. In yet 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 understood that N may be selected to encompass a similar but higher-order range.

[0045] In a particular embodiment, the panel of isolated biomarkers includes one or more, two or more, three or more, four or more, or five isolated biomarkers, each containing an amino acid sequence selected from AFTECCVVASQLR, ELLESYIDGR, ITLPDFTGDLR, TDAPDLPEENQAR, and SFRPFVPR. In some embodiments, the panel of isolated biomarkers includes one or more, two or more, three or more, four or more, or five isolated biomarkers, each containing 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 includes one or more, two or more, or three isolated biomarkers consisting of amino acid sequences selected from AFTECCVVASQLR, ELLESYIDGR, and ITLPDFTGDLR. 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 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 Vol. 249 (No. 4975), pp. 1429-1431 (1990) (UniProtKB / Swiss-Prot:P18428.3); prothrombin (THRB), Walz et al., Proc. Natl. Acad. Sci. USA Vol. 74 (No. 5), pp. 1969-1972 (1977) (NCBI Reference Sequence:NP_000497.1); complement component C5 (C5 or CO5), Haviland, J. Immunol. Vol. 146 (No. 1), pp. 362-368 (1991) (GenBank:AAA51925.1); plasminogen (PLMN), Petersen et al., J. Biol. It contains one or more peptides including fragments from Chem. Vol. 265 (No. 11), pp. 6104-6111 (1990) (NCBI Reference Sequences: NP_000292.1 NP_001161810.1); and complement component C8 gamma chain (C8G or CO8G), Haefliger et al., Mol. Immunol. Vol. 28 (No. 1-2), pp. 123-131 (1991) (NCBI Reference Sequence: NP_000597.2).

[0049] In some embodiments, a panel of isolated biomarkers is described as a cell adhesion molecule exhibiting phasicity with complement component 1, q subcomponent B chain (C1QB), Reid, Biochem. J. Vol. 179 (No. 2), pp. 367-371 (1979) (NCBI Reference) Sequence:NP_000482.3); Fibrinogen beta chain (FIBB or FIB); Watt et al., Biochemistry Vol. 18 (No. 1), pp. 68-76 (1979) (NCBI Reference Sequences:NP_001171670.1 and NP_005132.2); C-reactive protein (CRP), Oliveira et al., J. Biol. Chem. Vol. 254 (No. 2), pp. 489-502 (1979) (NCBI Reference Sequence:NP_000558.2); Inter-alpha-trypsin inhibitor heavy chain H4 (ITIH4), Kim et al., Mol. Biosyst. Vol. 7 (No. 5), pp. 1430-1440 (2011) (NCBI Reference The sequence contains one or more peptides, including fragments from: chorionic somatomammotropin hormone (CSH) Selby et al., J. Biol. Chem. Vol. 259 (No. 21), pp. 13131-13138 (1984) (NCBI Reference Sequence: NP_001308.1); and angiotensinogen (ANG or ANGT) Underwood et al., Metabolism Vol. 60 (No. 8): pp. 1150-1177 (2011) (NCBI Reference Sequence: NP_000020.1).

[0050] In additional embodiments, the present invention provides a panel of isolated biomarkers comprising N 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 AFTECCVVASQLR, ELLESYIDGR, and ITLPDFTGDLR. In additional embodiments, the biomarker panel comprises at least two isolated biomarkers selected from the group consisting of AFTECCVVASQLR, ELLESYIDGR, ITLPDFTGDLR, TDAPDLPEENQAR, and SFRPFVPR. In an additional embodiment, the biomarker panel includes at least two isolated biomarkers selected from the group consisting of FLNWIK, FGFGGSTDSGPIR, LLELTGPK, VEHSDLSFSK, IEGNLIFDPNNYLPK, ALVLELAK, TQILEWAAER, DVLLLVHNLPQNLPGYFWYK, SEPRPGVLLR, ITQDAQLK, ALDLSLK, WWGGQPLWITATK, and LSETNR.

[0051] In an additional embodiment, the biomarker panel includes at least two 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 present 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 includes at least two isolated biomarkers selected from the group consisting of alpha-1B-glycoprotein (A1BG), 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).

[0054] In some embodiments, the present invention provides a biomarker panel comprising 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 somatomammotropin hormone (CSH), and angiotensinogen (ANG or ANGT). In some embodiments, the present invention provides a biomarker panel comprising alpha-1B glycoprotein (A1BG), 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).

[0055] In another embodiment, the present 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 somatomammotropin hormone (CSH), and angiotensinogen (ANG or ANGT), as well as the biomarkers shown in Tables 51 and 53.

[0056] In another embodiment, the present invention provides a biomarker panel comprising at least two isolated biomarkers selected from the group consisting of alpha-1B-glycoprotein (A1BG), 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).

[0057] When used herein and in the appended claims, the singular forms "a," "an," and "the" should be noted to include multiple references unless otherwise explicitly indicated. Thus, for example, a reference to "biomarker" may include a mixture of two or more biomarkers.

[0058] The term "approximately," especially when referring to a given quantity, means to include a deviation of plus or minus 5 percent.

[0059] When used in this application, including the attached claims, the singular forms "a," "an," and "the" are used interchangeably with "at least one" and "one or more" to include multiple references unless the content otherwise clearly indicates.

[0060] As used herein, the terms “comprises,” “comprising,” “includes,” “including,” “contains,” “containing,” and any variations thereof are intended to describe non-exclusive inclusion, and a process, method, product-by-process, or composition that comprises, includes, or contains elements or a list of elements does not necessarily consist solely of those elements, but may include other elements not explicitly listed in or specific to such process, method, product-by-process, or composition.

[0061] As used herein, the term “panel” refers to a composition containing one or more biomarkers, such as an array or collection. The term may also refer to a profile or index of the expression patterns 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 extracted from its natural environment (e.g., the natural environment, if it exists naturally) and is therefore altered by human hands from its natural state. Isolated proteins or nucleic acids differ from the forms in which they exist in nature.

[0063] The term “biomarker” refers to a biological molecule or fragment of a biological molecule whose changes and / or detection 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 premature birth. Such biomarkers include, but are not limited to, biological molecules including nucleotides, nucleic acids, nucleosides, amino acids, sugars, fatty acids, steroids, metabolites, peptides, polypeptides, proteins, carbohydrates, lipids, hormones, antibodies, regions of interest that act as surrogates for biological macromolecules, and combinations thereof (e.g., glycoproteins, ribonucleoproteins, lipoproteins). This term also encompasses a portion or fragment of a biological molecule, for example, a peptide fragment of a protein or polypeptide containing at least 5 consecutive amino acid residues, at least 6 consecutive amino acid residues, at least 7 consecutive amino acid residues, at least 8 consecutive amino acid residues, at least 9 consecutive amino acid residues, at least 10 consecutive amino acid residues, at least 11 consecutive amino acid residues, at least 12 consecutive amino acid residues, at least 13 consecutive amino acid residues, at least 14 consecutive amino acid residues, at least 15 consecutive amino acid residues, at least 5 consecutive amino acid residues, at least 16 consecutive amino acid residues, at least 17 consecutive amino acid residues, at least 18 consecutive amino acid residues, at least 19 consecutive amino acid residues, at least 20 consecutive amino acid residues, at least 21 consecutive amino acid residues, at least 22 consecutive amino acid residues, at least 23 consecutive amino acid residues, at least 24 consecutive amino acid residues, at least 25 consecutive amino acid residues, or more consecutive amino acid residues.

[0064] The present invention also provides a method for determining the probability of preterm birth in a pregnant woman, comprising detecting measurable features 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 of preterm birth in the pregnant woman. As disclosed herein, the measurable features include fragments or derivatives of each of the N biomarkers selected from the biomarkers listed in Tables 1 to 63. In some embodiments of the disclosed method, detecting measurable features involves quantifying the amount of each of the N biomarkers selected from the biomarkers listed in Tables 1 to 63, their combinations or parts and / or derivatives in the biological sample obtained from the pregnant woman.

[0065] The present invention further provides a method for predicting GAB, which includes detecting measurable features 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 predicting GAB by analyzing the measurable features.

[0066] The present invention also provides a method for predicting GAB, comprising determining a predicted GAB birth in a pregnant woman, which includes (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 the amounts by a predetermined coefficient or thresholding them; and (c) 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 the time to delivery 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) determining a predicted GAB in the pregnant woman, including multiplying or thresholding the amounts by a predetermined coefficient; (d) adding the individual products to obtain a total risk score corresponding to the predicted GAB; and (e) predicting the time to delivery in the pregnant woman by subtracting the estimated gestational age (GA) at the time the biological sample was obtained from the predicted GAB. In the context of methods directed towards predicting the time to delivery, “delivery, birth” is understood to mean delivery following the spontaneous onset of labor, with or without rupture of membranes.

[0068] While the disclosure has been described and illustrated with reference to methods for determining the probability of preterm birth in pregnant women, it is equally applicable to methods for predicting GAB, predicting full-term birth, determining the probability of full-term birth in pregnant women, and predicting the time to delivery in pregnant women. It will be apparent to those skilled in the art that each of the above methods has specific substantial utility and benefits with respect to maternal-fetal health considerations.

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

[0070] In additional embodiments, a disclosed method for determining the probability of preterm birth in a pregnant woman and related methods disclosed herein include detecting measurable features 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, methods for determining the probability of preterm birth in pregnant women and related methods disclosed herein include detecting measurable features 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, methods for determining the probability of preterm birth in pregnant women and related methods disclosed herein include detecting measurable features of each of at least two isolated biomarkers selected from the group consisting of alpha-1B-glycoprotein (A1BG), 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).

[0073] In further embodiments, a disclosed method for determining the probability of preterm birth in a pregnant woman and related methods disclosed herein include detecting measurable features 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 somatomammotropin hormone (CSH), and angiotensinogen (ANG or ANGT).

[0074] In further embodiments, a disclosed method for determining the probability of preterm birth in a pregnant woman and related methods disclosed herein include detecting measurable features of each of at least two isolated biomarkers selected from the group consisting of alpha-1B glycoprotein (A1BG), 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).

[0075] In further embodiments, a disclosed method for determining the probability of preterm birth in a pregnant woman and related methods disclosed herein include detecting measurable features of each of at least two isolated biomarkers selected from the group consisting of alpha-1B glycoprotein (A1BG), 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).

[0076] In a further embodiment, a disclosed method for determining the probability of preterm birth in a pregnant woman and related methods disclosed herein include detecting measurable features 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 an additional embodiment, a method for determining the probability of preterm birth in a pregnant woman further comprises detecting measurable features for one or more risk indicators associated with preterm birth. In an additional embodiment, risk indicators are selected from the group consisting of a previous low birth weight or premature birth, multiple second-term spontaneous abortions, a previous first-term induced abortion, familial and intergenerational factors, a history of infertility, nulliparity, placental abnormalities, cervical and uterine abnormalities, pregnancy bleeding, intrauterine growth restriction, intrauterine diethylstilbestrol exposure, multiple pregnancies, infant sex, short stature, low pre-pregnancy weight, low or high body mass index, diabetes mellitus, hypertension, and genitourinary tract infections.

[0078] A “measurable feature” is any feature, characteristic, or aspect that can be determined and correlated with the probability of preterm birth in a subject. The term further includes any feature, characteristic, or aspect that can be determined and correlated with the prediction of GAB, the prediction of full-term birth, or the prediction of time to birth in pregnant women. With respect to biomarkers, such measurable features may include, for example, the presence or absence of the biomarker or fragments of it in a biological sample, the concentration, altered structure, for example, the presence or amount of post-translational modifications, for example, oxidation at one or more positions on the amino acid sequence of the biomarker, or, for example, the presence of an altered structure compared to the three-dimensional structure 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 features may further include risk indicators, for example, maternal characteristics, age, race, ethnicity, medical history, past pregnancy history, and obstetric history. Regarding risk indicators, measurable features may include, for example, a history of low birth weight or premature birth, multiple second-term spontaneous abortions, a history of first-term induced abortion, familial and intergenerational factors, a history of infertility, nulliparity, placental abnormalities, cervical and uterine abnormalities, short cervical length measurements, pregnancy bleeding, intrauterine growth restriction, intrauterine diethylstilbestrol exposure, multiple pregnancies, infant sex, short stature, low pre-pregnancy weight / low body mass index, diabetes, hypertension, genitourinary tract infections, hypothyroidism, asthma, poor academic performance, smoking, drug use, and alcohol consumption.

[0079] In some embodiments of the disclosed method for 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 method for determining the probability of preterm birth includes detecting and / or quantifying one or more biomarkers using mass spectrometry, a capture agent, or a combination thereof.

[0080] In some embodiments, the disclosed method for determining the probability of preterm birth in a pregnant woman comprises a first step of preparing a biomarker panel containing N biomarkers listed in Tables 1 to 63. In additional embodiments, the disclosed method for determining the probability of preterm birth in a pregnant woman comprises a first step of preparing a biological sample from the pregnant woman.

[0081] In some embodiments, the disclosed method for determining the probability of preterm birth in a pregnant woman includes communicating the probability to a healthcare provider. Similarly, the disclosed methods for predicting GAB, predicting full-term birth, determining the probability of full-term birth in a pregnant woman, and predicting time to delivery in a pregnant woman also include communicating the probability to a healthcare provider. As 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 methods for predicting GAB, methods for predicting full-term birth, methods for determining the probability of full-term birth in a pregnant woman, and methods for predicting time to delivery in a pregnant woman. Specifically, the biomarkers and panels cited throughout this application, with explicit reference to methods for preterm birth, can also be used in methods for predicting GAB, methods for predicting full-term birth, methods for determining the probability of full-term birth in a pregnant woman, and methods for predicting time to delivery in a pregnant woman. It will be apparent to those skilled in the art that each of the above methods has specific substantial utility and benefit with respect to maternal-fetal health considerations.

[0082] In additional embodiments, the pregnant woman is notified by communication of the decision regarding subsequent treatment. In some embodiments, the method for determining the probability of premature birth in a pregnant woman includes an additional feature of expressing 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 baseline score representing the average amount of one or more biomarkers calculated from biological samples obtained from a random pool of pregnant women. Because biomarker levels may not be static throughout pregnancy, the standard score or baseline score must be obtained for a point in pregnancy corresponding to that of the pregnant woman at the time the sample was taken. The standard score or baseline score can be predetermined and incorporated into a predictive model so that the comparison is indirect rather than actually performed each 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 a particular embodiment, if the risk score is greater than the standard risk score or baseline risk score, the pregnant woman may have an increased risk of preterm birth. In some embodiments, the magnitude of a pregnant woman's risk score, or the amount exceeding a baseline risk score, may indicate 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, non-limited to T cells, monocytes, neutrophils, erythrocytes, 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 weeks of gestation. Other commonly used subcategories of preterm birth have been established, dividing them into moderate preterm (birth at 33–36 weeks of gestation), very preterm (birth at <33 weeks of gestation), and extremely preterm (birth at ≤28 weeks of gestation). With respect to the methods disclosed herein, those skilled in the art will understand that when implementing the methods disclosed herein, cutoffs for dividing preterm birth from full-term birth and cutoffs for dividing the subcategories of preterm birth can be adjusted, for example, to maximize certain health benefits. It will be further understood that such adjustments are well within the scope of the skills of those skilled in the art and are included within the scope of the invention disclosed herein. Gestational age is a surname for the degree of fetal development and the preparation of the fetus for birth. Gestational age is typically defined as the length of time from the last normal menstrual period to the day of delivery. However, obstetric measures and ultrasound estimates can also be helpful in estimating gestational age. Preterm births are generally classified into two separate subgroups. One is spontaneous preterm birth, which occurs following the spontaneous onset of early labor or premature rupture of membranes, regardless of subsequent labor augmentation or cesarean section. The second is indicated preterm birth, which occurs following induction or cesarean section for one or more conditions that the woman's caregiver determines threaten the health or life of the mother and / or fetus. In some embodiments, the methods disclosed herein are directed towards determining the probability of spontaneous preterm birth. In additional embodiments, the methods disclosed herein are directed towards predicting pregnancy and childbirth.

[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 preceding paragraph. In contrast, the term “expected gestational age at birth” or “expected GAB” refers to the GAB determined based on the methods of the present invention disclosed herein. As used herein, “full-term birth” refers to a birth that is equal to or exceeds 37 weeks of gestation.

[0087] In some embodiments, the pregnant woman is between 17 and 28 weeks of gestation at the time the biological sample is collected. In other embodiments, the pregnant woman is between 16 and 29 weeks of gestation, between 17 and 28 weeks of gestation, between 18 and 27 weeks of gestation, between 19 and 26 weeks of gestation, between 20 and 25 weeks of gestation, between 21 and 24 weeks of gestation, or between 22 and 23 weeks of gestation at the time the biological sample is collected. In further embodiments, the pregnant woman is between approximately 17 and 22 weeks of gestation, between approximately 16 and 22 weeks of gestation, between approximately 22 and 25 weeks of gestation, between approximately 13 and 25 weeks of gestation, between approximately 26 and 28 weeks of gestation, or between approximately 26 and 29 weeks of gestation at the time the biological sample is collected. Thus, the gestation period of the pregnant woman at the time the biological sample is collected may 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 method, the measurable feature comprises fragments or derivatives of each of N biomarkers selected from the biomarkers listed in Tables 1 to 63. In additional embodiments of the claimed method, detecting the measurable feature comprises quantifying the amount of each of the N biomarkers, combinations or parts and / or derivatives thereof, selected from the biomarkers listed in Tables 1 to 63, in a biological sample obtained from the pregnant woman.

[0089] The terms “quantity” or “level,” as used herein, refer to the amount of a biomarker that is detectable or measurable in a biological sample and / or control. The quantity of a biomarker may be, for example, the amount of a polypeptide, the amount of a nucleic acid, or the amount of a fragment or surrogate. The term may also include combinations thereof. The term “quantity” or “level” of a biomarker is the measurable characteristic of that biomarker.

[0090] In some embodiments, 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. Any existing, available, or conventional separation, detection, and quantification methods can be used herein to measure the presence or absence (e.g., readout is present vs. absent; or detectable amount vs. undetectable amount) and / or amount (e.g., readout is absolute or relative, e.g., absolute or relative concentration) of biomarkers, peptides, polypeptides, proteins, and / or their fragments in a sample, as well as, optionally, one or more other biomarkers or their fragments. In some embodiments, the detection and / or quantification of one or more biomarkers involves assays utilizing a capture agent. In further embodiments, the capture agent is an antibody, antibody fragment, nucleic acid-based protein-binding reagent, small molecule, or variant thereof. In additional embodiments, the assay is an enzyme-linked immunosorbent assay (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 further embodiments, the mass spectrometry is co-immunoprecipitation-mass spectrometry (co-IP MS), where mass spectrometry is followed by co-immunoprecipitation, which is a suitable technique for the isolation of the whole protein complex.

[0091] As used herein, the term “mass spectrometer” refers to a device capable of volatilizing / ionizing an analyte, forming gaseous ions, and determining their absolute or relative molecular weight. Suitable methods of volatilization / ionization include matrix-assisted laser desorption / ionization (MALDI), electrospray, laser / optical, thermal, electrical, atomization / spray, or a combination 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 consisting of various combinations of these types of mass spectrometers. These instruments can then be adapted with various other instruments to fractionate the sample (e.g., liquid chromatography or solid-phase adsorption techniques based on chemical or biological characteristics) and ionize the sample for introduction into the mass spectrometer (including matrix-assisted laser desorption (MALDI), electrospray, or nanospray ionization (ESI), or a combination thereof).

[0092] In general, any mass spectrometry (MS) technique (e.g., tandem mass spectrometry, MS / MS; or post-source degradation, TOF MS) that can provide accurate information regarding the mass of a peptide, preferably, and also regarding the fragmentation and / or (partial) amino acid sequence of a selected peptide, can be used in the methods disclosed herein. Suitable MS and MS / MS techniques and systems for peptides are well known in themselves (see, for example, Methods in Molecular Biology, Vol. 146: "Mass Spectrometry of Proteins and Peptides," edited by Chapman, Humana Press 2000; Biemann 1990 Methods Enzymol, Vol. 193: pp. 455-459; or Methods in Enzymology, Vol. 402: "Biological Mass Spectrometry," edited by Burlingame, Academic Press 2005) and can be used when performing the methods disclosed herein. Thus, in some embodiments, the methods disclosed include performing quantitative MS to measure one or more biomarkers. Such quantitative methods can be carried out in an automated (Villanueva et al., Nature Protocols (2006) Vol. 1 (No. 2): pp. 880-891) or semi-automated format. In certain embodiments, MS can be operably coupled 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 tagging (ICAT), tandem mass tagging (TMT), or stable isotope labeling with amino acids in cell culture (SILAC), followed by chromatography and MS / MS.

[0093] As used herein, the terms “Multiple Reaction Monitoring (MRM)” or “Selective Reaction Monitoring (SRM)” refer to MS-based quantification methods particularly useful for quantifying analytes in small quantities. In an SRM experiment, one or more predefined precursor ions and their fragments are selected by two mass filters in 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., peptides or small molecules, such as chemicals, steroids, or hormones), can constitute a final assay. Numerous analytes can be quantified during a single LC-MS experiment. In reference to MRM or SRM, the terms “scheduled” or “dynamic” refer to variations in the assay, where transitions for a particular analyte are acquired only within 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 test selectivity, because retention time is a characteristic dependent on the physical properties of the analyte. A single analyte can also be monitored using more than one transition. Finally, different standards corresponding to the analyte of interest (e.g., the same amino acid sequence) but with the inclusion of stable isotopes may be included in the assay. Stable isotope standards (SIS) can be incorporated into the assay at precise levels and used to quantify the corresponding unknown analyte. Additional levels of specificity are contributed by the co-elution of the unknown analyte and its corresponding SIS, as well as the characteristics of their transitions (e.g., the similarity in the ratio of levels of two unknown transitions and the ratio of two transitions of their 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 decomposition (PSD); MALDI-TOF / TOF; surface-enhanced laser desorption / ionization time-of-flight (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) nThis may include: Fragmentation of peptide ions in a tandem MS (MS / MS) configuration can be achieved using established methods in the art, such as collision-induced dissociation (CID). As described herein, detection and quantification of biomarkers by mass spectrometry may include, for example, multiple reaction monitoring (MRM) as described by Kuhn et al., Proteomics Vol. 4: pp. 1175-1186 (2004). Acquisition of a scheduled multiple reaction monitoring (scheduled MRM) mode during LC-MS / MS analysis enhances the sensitivity and accuracy of peptide quantification. Anderson and Hunter, Molecular and Cellular Proteomics Vol. 5 (No. 4): p. 573 (2006). As described herein, mass spectrometry-based assays can be advantageously combined with methods for separating or fractionating upstream peptides or proteins, such as chromatography and other methods described below herein. 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] Those skilled in the art will understand that several methods can be used to determine the amount of a biomarker (including mass spectrometry approaches, e.g., MS / MS, LC-MS / MS, multiple reaction monitoring (MRM), or SRM, and product ion monitoring (PIM), as well as antibody-based methods, e.g., immunoassays, e.g., Western blotting, enzyme-linked immunosorbent assay (ELISA), immunoprecipitation, immunohistochemistry, immunofluorescence, radioimmunoassay, dot blotting, and FACS). Thus, in some embodiments, determining the level of at least one biomarker involves using an immunoassay and / or mass spectrometry. In additional embodiments, the mass spectrometry is selected from MS, MS / MS, LC-MS / MS, SRM, PIM, and other such methods known in the art. In other embodiments, LC-MS / MS further includes 1D LC-MS / MS, 2D LC-MS / MS, or 3D LC-MS / MS. Immunoassay techniques and protocols are generally known to those skilled in the art (Price and Newman, Principles and Practice). (of Immunoassay, 2nd edition, Grove's Dictionaries, 1997 and Gosling, Immunoassays: A Practical Approach, Oxford University Press, 2000). Various immunoassay techniques, including competitive and non-competitive immunoassays, can be used (Self et al., Curr. Opin. Biotechnol., Vol. 7: pp. 60-65 (1996)).

[0096] In further embodiments, the immunoassay is selected from Western blotting, ELISA, immunoprecipitation, immunohistochemistry, immunofluorescence, radioimmunoassay (RIA), dot blotting, and FACS. In certain embodiments, the immunoassay is ELISA. In further embodiments, the ELISA is direct ELISA (enzyme-linked immunosorbent assay), indirect ELISA, sandwich ELISA, competitive ELISA, multiplex ELISA, ELISPOT technology, and other similar technologies 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, ELISA is performed using antibodies, but they can be performed using any capture agent that can specifically bind to and detect one or more biomarkers of the present invention. Multiplex ELISA typically enables the simultaneous detection of two or more analytes within a single compartment (e.g., a microplate well) using multiple array addresses (Nielsen and Geierstanger 2004, J Immunol Methods Vol. 290: pp. 107-120 (2004) and Ling et al. 2007, Expert Rev Mol Diagn Vol. 7: pp. 87-98 (2007)).

[0097] In some embodiments, radioimmunoassays (RIAs) can be used to detect one or more biomarkers in the methods of the present invention. RIAs are competitive-based assays well known in the art, and are radioactively labeled (e.g., 125 I or 131(1) This involves mixing a known amount of the target analyte with an analyte-specific antibody, then adding an unlabeled analyte from the sample, and measuring the amount of the substituted labeled analyte (see, for example, An Introduction to Radioimmunoassay and Related Techniques, edited by Chard T, Elsevier Science 1995, ISBN 0444821198 for guidance).

[0098] Detectable labels can be used in the assays described herein for the direct or indirect detection of biomarkers in the methods of the present invention. A variety of detectable labels can be used, and the selection of a label depends on the required sensitivity, ease of conjugation with the antibody, stability requirements, and available equipment and disposal regulations. Those skilled in the art will be familiar with the selection of an appropriate detectable label based on the assay detection of biomarkers in the methods of the present invention. Appropriate detectable labels include, but are not limited to, fluorescent dyes (e.g., fluorescein, fluorescein isothiocyanate (FITC), Oregon Green®, rhodamine, Texas Red, tetrarhodimine isothiocyanate (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, different tagging methods using isotopic reagents (e.g., isotope-coded affinity tagging (ICAT) or isobaric tagging reagents, more recent variations using iTRAQ (Applied Biosystems, Foster City, Calif.), or tandem mass tagging, TMT (Thermo Scientific, Rockford, IL) (followed by multidimensional liquid chromatography (LC) and tandem mass spectrometry (MS / MS) analysis) can provide further methodologies when performing the methods of the present invention.

[0100] Chemiluminescent assays using chemiluminescent antibodies can be used for highly sensitive, non-radioactive detection at the protein level. Antibodies labeled with fluorescent dyes may 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 lysamine. Indirect labeling includes various enzymes well known in the art, such as horseradish peroxidase (HRP), alkaline phosphatase (AP), beta-galactosidase, and urease. Detection systems using appropriate substrates for horseradish peroxidase, alkaline phosphatase, and beta-galactosidase are well known in the art.

[0101] Signals from direct or indirect labeling can be used, for example, in a spectrophotometer to detect color from a chromogenic substrate; or in a radiation counter to detect radiation, for example. 125For the detection of I, a gamma counter or similar device may be used; or, the analysis may be performed using a fluorometer that detects fluorescence in the presence of light of a specific wavelength. For the detection of enzyme-bound 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 assay used to carry out the present invention can be automated or performed by a robot to simultaneously detect signals from multiple samples.

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

[0103] As described above, chromatography can also be used when carrying out the methods of the present invention. Chromatography encompasses methods for separating chemical substances and generally involves a process in which a mixture of analytes is carried by a mobile flow of liquid or gas ("mobile phase") and flows around or over a fixed liquid or solid phase ("stationary phase"), resulting in the separation of the analytes into components as a result of different distributions of the analytes between the mobile phase and the stationary phase. The stationary phase can typically be a finely divided solid, a sheet of filter material, or a thin film of liquid on the surface of a solid. Chromatography is well understood by those skilled in the art as an applicable technique 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., stationary phase deposited or packed in a column), preferably liquid chromatography, more preferably high-performance liquid chromatography (HPLC), or ultrafast / 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) (e.g., cation or anion exchange chromatography), hydrophilic interaction chromatography (HILIC), hydrophobic interaction chromatography (HIC), size exclusion chromatography (SEC) (including gel filtration chromatography or gel permeation chromatography), chromatofocusing, affinity chromatography (e.g., immunoaffinity, immobilized metal affinity chromatography), and the like. Chromatography, including single, bidimensional, or more-dimensional chromatography, can be used as a peptide fractionation method in combination with further peptide analysis methods, such as downstream mass spectrometry as described elsewhere in this specification.

[0105] Further peptide or polypeptide isolation, identification, or quantification methods may be optionally used in conjunction with any of the analytical methods described above to measure the biomarkers in this disclosure. Such methods include, but are not limited to, chemical extraction and resolution, isoelectric focusing (IEF) (including capillary isoelectric focusing (CIEF), capillary isokinetic electrophoresis (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 “scavenger” refers to a compound that can specifically bind to a target, particularly a biomarker. This term includes antibodies, antibody fragments, nucleic acid-based protein-binding reagents (e.g., aptamers, Slow Off-rate Modified Aptamer (SOMAmer®)), protein scavengers, natural ligands (i.e., hormones for their receptors or vice versa), small molecules, or variants thereof.

[0107] The capture agent may be configured to specifically bind to a target, particularly a biomarker. The capture agent may include, but is not limited to, organic molecules, such as polypeptides, polynucleotides, and other nonpolymer molecules identifiable to those skilled in the art. In 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 complex mixtures of biological media for use in the disclosed method.

[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); and Goding, Monoclonal Antibodies: Principles and Practice (2nd edition, 1986). Antibody capture agents can be any immunoglobulin or its derivatives, whether naturally occurring or produced entirely or partially synthetically. This term also includes all derivatives thereof that maintain specific binding ability. Antibody capture agents have a binding domain that is homologous or largely homologous to the immunoglobulin binding domain and can be derived from natural sources or produced partially or entirely synthetically. Antibody capture agents can be monoclonal antibodies or polyclonal antibodies. In some embodiments, the antibody is a single-chain antibody. Those skilled in the art will understand that antibodies can be provided in any of the various forms, including, for example, humanized, partially humanized, chimeric, chimeric humanized, etc. Antibody capture agents may include, but are not limited to, antibody fragments such as Fab, Fab', F(ab')2, scFv, Fv, dsFv diabodies, and Fd fragments. Antibody capture agents can be produced by any means. For example, antibody capture agents can be produced enzymatically or chemically by fragmentation of intact antibodies, and / or it can be produced recombinantly from a gene encoding a partial antibody sequence. Antibody capture agents may include single-chain antibody fragments. Alternatively, or in addition, antibody capture agents may include, for example, multiple chains linked together by disulfide bonds; and any functional fragments obtained from such molecules (in which case such fragments retain the specific binding characteristics of the parent antibody molecule). Due to their smaller size as functional components of the whole molecule, antibody fragments may offer advantages over intact antibodies for use in certain immunochemical techniques and experimental applications.

[0109] Suitable scavengers useful for carrying out the present invention also include aptamers. An aptamer is an oligonucleotide sequence that can specifically bind to its target via a unique three-dimensional (3-D) structure. An aptamer can contain any suitable number of nucleotides, and different aptamers can have the same or different numbers of nucleotides. An aptamer can be DNA or RNA or a chemically modified nucleic acid, and can be single-stranded, double-stranded, or contain double-stranded regions and may contain a higher-dimensional structure. An aptamer can also be a photoaptamer, 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 an aptamer scavenger may include the use of two or more aptamers that specifically bind to the same biomarker. An aptamer may include a tag. An aptamer 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 (No. 27): pp. 4117-4125 (2011). A useful scavenger in carrying out the method of the present invention also includes SOMAmer (Slow Off-Rate Modified Aptamer), known in the art, which has improved off-rate characteristics. Brody et al., J Mol Biol. 422 (No. 5): pp. 595-606 (2012). SOMAmer can be produced using any known method, including the SELEX method.

[0110] It is understood by those skilled in the art that biomarkers can be modified before analysis to improve their resolution or determine their identity. For example, biomarkers can be subjected to proteolytic digestion before 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 act as fingerprints for the biomarkers, thereby indirectly enabling their detection. This is particularly useful when there are biomarkers with similar molecular weights that can be confused with the biomarker in question. Proteolytic fragmentation is also useful for high molecular weight biomarkers because smaller biomarkers are more easily degraded 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, improving binding to anionic adsorbents and thus improving detection resolution. In yet another example, biomarkers can be modified by attaching tags of a specific molecular weight that specifically bind to molecular biomarkers, further distinguishing them. After optionally 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 subsequently probed for the presence of proteins. Alternatively, protein-binding molecules attached to microspheres, microparticles, microbeads, beads, or other particles can be used for the capture and detection of biomarkers. Protein-binding molecules can be antibodies, peptides, peptoids, aptamers, small molecule ligands, or other protein-binding scavengers attached to the surface of particles. Each protein-binding molecule may contain a unique detectable label that it encodes, which can distinguish it from other detectable labels attached to other protein-binding molecules and enable the detection of biomarkers in multiple assays. Examples include, but are not limited to, color-coded microspheres with known fluorescence intensities (see, e.g., microspheres produced by Luminex (Austin, Tex.) using xMAP technology); microspheres containing quantum dot nanocrystals, e.g., with different ratios and combinations of quantum dot colors (e.g., Qdot nanocrystals produced by Life Technologies (Carlsbad, Calif.)); glass-coated metal nanoparticles (e.g., see SERS nanotags produced by Nanoplex Technologies, Inc. (Mountain View, Calif.)); barcode materials (e.g., submicron-sized striped metal rods, e.g., Nanobarcode produced by Nanoplex Technologies, Inc.); code nanoparticles with color barcodes (e.g., see CellCard produced by Vitra Bioscience, vitrabio.com); glass nanoparticles with digital holographic code images (e.g., see CyVera microbeads produced by Illumina (San Diego, Calif.)); chemiluminescent dyes, combinations of dye compounds; and detectable beads of different sizes.

[0112] In another embodiment, a biochip 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, a protein biochip comprises a substrate having a surface. A capture reagent or adsorbent is attached to the surface of the substrate. Frequently, the surface comprises a plurality of addressable positions, each of which has a capture agent bound thereto. The capture agent may be a biological molecule, such as a polypeptide or nucleic acid, which specifically captures other biomarkers. Alternatively, the capture agent may be a chromatographic material, such as an anion exchanger or hydrophilic substance. Examples of protein biochips are well known in the art.

[0113] Measuring mRNA in a biological sample can be used as a substitute for detecting the levels of corresponding protein biomarkers in the biological sample. 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 produce cDNA from mRNA. The cDNA can then be used in a qPCR assay to produce fluorescence as the DNA amplification process progresses. By comparison with a standard curve, qPCR can produce absolute measurements, such as the mRNA copy number per cell. The expression levels of mRNA in samples have been measured using Northern blotting, microarrays, Invader assays, and RT-PCR combined with capillary electrophoresis. 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 preterm birth in pregnant women. The probability of preterm birth in pregnant women can be determined by detecting the expression levels of one or more biomarkers and / or determining the ratio of biomarkers. Such detection methods can be used, for example, for early diagnosis of a condition, to determine whether a subject is predisposed to preterm birth, to monitor the progression of preterm birth or the progression of a treatment protocol, to assess the severity of preterm birth, to predict the outcome of preterm birth and / or the prospect of recovery or full-term delivery, or to help determine appropriate treatment for preterm birth.

[0115] The quantification of biomarkers in biological samples can be determined, without limitation, by the methods described above and any other methods known in the art. The quantitative data thus obtained are then subjected to an analytical classification process. In such a process, the raw data are manipulated according to an algorithm 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 it can generate an algorithm using a different set of data by utilizing the guidelines provided herein.

[0116] In some embodiments, analyzing measurable features to determine the probability of preterm birth in pregnant women involves the use of a predictive model. In further embodiments, analyzing measurable features to determine the probability of preterm birth in pregnant women involves comparing the measurable features to a reference feature. As those skilled in the art will understand, such a comparison may be a direct comparison with the reference feature or an indirect comparison in which the reference feature is incorporated into the predictive model. In further embodiments, analyzing measurable features to determine the probability of preterm birth in pregnant women involves one or more of the following: linear discriminant analysis models, support vector machine classification algorithms, recursive feature exclusion models, predictive analysis of microarray models, logistic regression models, CART algorithms, Flextree algorithms, LART algorithms, random forest algorithms, MART algorithms, machine learning algorithms, penalized regression methods, or a combination thereof. In certain embodiments, the analysis includes logistic regression.

[0117] In analytical classification processes, quantitative data can be manipulated and sample classifications can be provided using one of various statistical analysis methods. Examples of useful methods include linear discriminant analysis, recursive feature exclusion, microarray predictive analysis, logistic regression, CART algorithm, FlexTree algorithm, LART algorithm, random forest algorithm, MART algorithm, and machine learning algorithms.

[0118] To construct a random forest for predicting GAB, a person skilled in the art can consider a set of k subjects (pregnant women) whose gestational age at birth (GAB) is known, and whose N analytes (transitions) are measured in blood samples taken several weeks before delivery. The regression tree begins with a root node containing all subjects. The mean GAB for all subjects can be calculated at the root node. The variance of GABs in the root node will be high because there is a mixture of women with different GABs. The root node is then partitioned into two branches, each branch containing women with similar GABs. The mean GABs for subjects in each branch are calculated again. The variance of GABs in each branch will be lower than in the root node because the subset of women in each branch has relatively similar GABs than those in the root node. The two branches are created by selecting thresholds for the analytes and the analytes that create branches with similar GABs. The analytics and thresholds are selected from all possible pairs of analytics and thresholds, typically with a random subset of analytics at each node. The procedure continues to recursively produce branches, creating leaves (terminal nodes) with very similar GABs for subjects. The predicted GAB at each terminal node is the average GAB for 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 that sets a threshold for determining 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 statistically significant difference is obtained by comparing the obtained dataset with the reference dataset. If so, the sample from which the dataset was obtained is classified as not belonging to the class of the reference dataset. Conversely, if such a comparison does not show a statistically significant difference from the reference dataset, the sample from which the dataset was obtained is classified as belonging to the class of the reference dataset.

[0120] The predictive power of a model can be evaluated according to its ability to provide a quality metric (e.g., AUROC (Area Under the ROC Curve) or precision for a specific value or range of values). Measurements of the Area Under the Curve are useful for comparing the precision of classifiers across the full data range. Classifiers with a larger AUC have a greater ability to accurately classify an unknown between two groups of interest. In some embodiments, the desired quality threshold is a predictive model that classifies samples with precision of 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. As an alternative measure, the desired quality threshold may refer to a predictive model that classifies samples with AUC of 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.

[0121] As is well known in the art, the relative sensitivity and specificity of a predictive model can be adjusted to favor either a selection metric or a sensitivity metric, where the two metrics are inversely proportional. The constraints 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 performed. One or both of the sensitivity and specificity may 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] Raw data can initially be analyzed by measuring values ​​for each biomarker, typically in three or more sets of three measurements. The data can be manipulated; for example, raw data can be transformed using standard curves, as well as the mean of three measurements used to calculate the mean and standard deviation for each patient. These values ​​can be transformed before being used in a model (e.g., logarithmic transformation, Box-Cox transformation) (Box and Cox, Royal Stat. Soc., Series B, Vol. 26: pp. 211-246 (1964)). The data is then fed into a predictive model that classifies samples according to their condition. The resulting information can be communicated to the patient or healthcare provider.

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

[0124] In one embodiment, hierarchical clustering is performed in the induction of a 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 the standard in medical applications (Singer, Recursive Partitioning in the Health Sciences, Springer (1999)) for converting arbitrary qualitative features into quantitative features; Hotelling T 2 Regarding statistics, they can be selected by the significance level achieved, evaluated by sample reuse methods; and modified by the appropriate application of the lasso method. Problems in prediction become problems in regression without loss of predictive power by actually using the Gini criterion appropriately for classification when evaluating the quality of regression.

[0125] This approach led to something called FlexTree (Huang, Proc. Nat. Acad. Sci. USA Vol. 101: pp. 10529-10534 (2004)). FlexTree is very well implemented and useful in simulations and when applied to multiple forms of data. 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). This name, like CART and FlexTree, reflects the implementation of the lasso through binary trees; the lasso described; and what is called LARS by Efron et al. (2004) Annals of Statistics Vol. 32: pp. 407-451 (2004). See also Huang et al., Proc. Natl. Acad. Sci. USA. Vol. 101 (No. 29): pp. 10529-34 (2004). Other analytical methods that can be used include logical regression. One method of logical regression: Ruczinski, Journal of Computational and Graphical Statistics Vol. 12: pp. 475-512 (2003). Logical regression is similar to CART in that its classifier can be displayed as a binary tree. It differs in that each node has a Boolean description of a feature, which is more general than the simple "and" description produced by CART.

[0126] Another approach is the nearest shrinking centroid approach (Tibshirani, Proc. Natl. Acad. Sci. USA Vol. 99: pp. 6567-72 (2002)). This technique is k-means-like but has the advantage of automatically selecting features and focusing attention on a small number of informatively valuable ones, similar to the lasso method, by shrinking the cluster centers. This approach is available as PAM software and is widely used. Two further pairs of algorithms that can be used are Random Forest (Breiman, Machine Learning Vol. 45: pp. 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 include predictors that “vote” on outcomes.

[0127] To provide a significant order, the false detection rate (FDR) can be determined. First, a set of null distributions of difference values ​​is generated. In one embodiment, the values ​​of the observed profiles are reordered to create a distribution of a set of correlation coefficients obtained by chance, thereby creating a suitable set of null distributions of correlation coefficients (Tusher et al., Proc. Natl. Acad. Sci. USA, Vol. 98, pp. 5116-5121 (2001)). The set of null distributions is obtained by reordering the values ​​of each profile for all available profiles; calculating pairwise correlation coefficients for all profiles; calculating the probability density function of the correlation coefficients for this reorder; and repeating this procedure N times (where N is a large number, usually 300). Using the N distribution, a suitable measure (mean, median, etc.) of the count of correlation coefficient values ​​is calculated where their values ​​exceed the (similar) values ​​obtained from the experimentally observed distribution of similar values ​​at a given significance level.

[0128] The FDR is the ratio of the expected number of false significant correlations (estimated from correlations greater than this selected Pearson correlation in the randomized dataset) to the number of correlations greater than this selected Pearson correlation (significant correlation) in the empirical data. This cutoff correlation value can be applied to correlations between experimental profiles. Using the above distribution, a level of confidence is chosen for significance. This is used to determine the minimum correlation coefficient that exceeds the result that would have been obtained by chance. Using this method, a threshold for positive correlation, negative correlation, or both is obtained. Using this threshold, the user can filter observations of pairwise correlation coefficients and eliminate those that do not exceed the threshold. Furthermore, an estimate of the false positive rate can be obtained for a given threshold. For each of the individual “random correlation” distributions, it is possible to find out 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 mean 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 the temporal event analysis (survival analysis), where the event is the occurrence of preterm birth, and subjects without the event are considered censored at birth. Assuming a specific pregnancy outcome (preterm birth event or no event), and selecting the random time length observed for each patient, as well as proteomics 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 hazard ratios to be monotonically increasing, decreasing, or constant, and also has proportional hazards representation (as in the Cox model) and accelerated failure time representation. All standard tools available for obtaining approximate maximum likelihood estimators of regression coefficients and corresponding functions are available in this model.

[0130] In addition, Cox models can be used. In particular, because the analysis is significantly simplified by reducing the number of covariates to a manageable size, it offers the possibility of a non-parametric or semi-parametric approach to predicting time to preterm birth. These statistical tools are publicly known in the art and applicable to all forms of proteomics data. A set of biomarkers, clinical data, and genetic data is provided that can be easily determined and are highly informative regarding the probability of preterm birth and the predicted time to the preterm event in the pregnant woman. Furthermore, the algorithm provides information regarding the probability of preterm birth in the pregnant woman.

[0131] Therefore, those skilled in the art will understand that the probability of preterm birth according to the present invention can be determined using either quantitative or categorical variables. For example, when performing 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 preterm 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 quantitative variables, in particular the predicted gestational age at birth. The predicted gestational age at birth can then be used as a basis for predicting the risk of preterm birth. By first using quantitative variables and then converting the quantitative variables to categorical variables, the method of the present invention takes into account a sequence of measurements detected for the measurable features. For example, predicting the gestational age at birth, rather than making a binary prediction of preterm birth versus full term, allows for individualization of treatment to the pregnant woman. For example, an earlier predicted gestational age at birth would result in more intensive prenatal intervention, i.e., monitoring and treatment, than a predicted gestational age closer to full term.

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

[0133] In developing predictive models, it is often desirable to select a subset of markers, namely at least three, four, five, or six, up to a complete set of markers. Typically, the selection of marker subsets is made to provide what is necessary for quantitative sample analysis, such as reagent availability and quantification convenience, while maintaining a highly accurate predictive model. The selection of several informative markers for building a classification model requires the definition of performance metrics and user-defined thresholds for producing a model with useful predictive capability based on these metrics. For example, performance metrics could be AUC, predictive sensitivity and / or specificity, and the overall accuracy of the predictive model.

[0134] As will be understood by those skilled in the art, analytical classification processes can employ 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 exclusion, microarray predictive analysis, logistic regression, CART algorithm, FlexTree algorithm, LART algorithm, random forest algorithm, MART algorithm, and machine learning algorithms.

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

[0136] Table 1. Transitions with p-values ​​<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 used to distinguish preterm subjects from non-preterm subjects. 77 transitions with the highest AUROC area are shown. [Table 4-1] [Table 4-2]

[0140] Table 5. AUROC for Random Forest, Boosting, Lasso, and Logistic Regression models for a specific number of transitions allowed in the model, estimated by 100 rounds of bootstrap resampling. [Table 5]

[0141] Table 6. The top 15 transitions selected by each multivariate method, ranked by importance of the method. [Table 6]

[0142] In yet another embodiment, the present invention provides a kit for determining the probability of preterm birth, which can be used to detect N 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 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 isolated biomarkers selected from the group consisting of FLNWIK, FGFGGSTDSGPIR, LLELTGPK, VEHSDLSFSK, IEGNLIFDPNNYLPK, 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 isolated biomarkers selected from the group consisting of alpha-1B glycoprotein (A1BG), 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).

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

[0146] In one embodiment, the kit comprises a drug for measuring at least N levels of isolated biomarkers listed in Tables 1 to 63. The kit may include antibodies that specifically bind to these biomarkers, for example, the kit may include at least one of the following: 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 a drug for measuring levels of at least N isolated biomarkers listed in Tables 1 to 63. The kit may include antibodies that specifically bind to these biomarkers, and for example, the kit may include at least one antibody that specifically binds to alpha-1B glycoprotein (A1BG), 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), or inhibin beta E chain (INHBE).

[0148] The kit may include one or more containers for the compositions contained within the kit. The compositions may be in liquid form or lyophilized. Suitable containers for the compositions include, for example, bottles, vials, syringes, and test tubes. The containers may be made from a variety of materials, including glass or plastic. The kit may also include accompanying documentation, which may include instructions on how to determine the probability of preterm birth.

[0149] From the above description, it will be clear that modified and altered forms can be created for the present invention as described herein to be used in various ways and conditions. Such embodiments are also within the scope of the following claims.

[0150] Any enumeration of elements in any definition of a variable herein includes the definition of that variable as any single element or as a combination (or subcombination) of the enumerated elements. Any enumeration of embodiments herein includes that embodiment as any single embodiment or in combination with any other embodiment or parts thereof.

[0151] All patents and publications referenced herein are incorporated by reference to the same degree as each individual patent and publication is specifically and individually incorporated by reference.

[0152] The following embodiments are provided for illustrative purposes only, without limitation. [Examples]

[0153] (Example 1) Development of a sample set for the discovery and validation of biomarkers for premature birth A standard protocol was developed to manage the implementation of the Proteomic Assessment of Preterm Risk (PAPR) clinical study. This protocol also specified that samples and clinical information could be used to study other pregnancy complications in some of the subjects. Samples were obtained from women at 11 locations across the United States accredited by an Internal Review Board (IRB). After obtaining informed consent, serum and plasma samples, as well as relevant information on patients' demographic characteristics, past medical and pregnancy history, current pregnancy history, and concomitant medications, were obtained. Following delivery, data related to maternal and infant status and complications were collected. Serum and plasma samples were processed according to a protocol requiring 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 examined individually, and their condition was determined to be either spontaneous preterm or medically indicated preterm. Only spontaneous preterm cases were used for this analysis. To discover preterm birth biomarkers, 80 samples were analyzed in two gestational age groups: a) a late window consisting of samples from 23–28 weeks of gestation (13 cases, including 13 full-term controls matched within one week of sample collection and 14 random full-term controls); and b) an early window consisting of samples from 17–22 weeks of gestation (15 cases, including 15 full-term controls matched within one week of sample collection and 10 random full-term controls).

[0155] The samples were then depleted of large amounts of protein using the Human 14 Multiple Affinity Removal System (MARS 14), thereby removing the most abundant proteins that were treated as having no informational value for identifying disease-related changes in the serum proteome. For this purpose, an equal volume of each clinical or pooled human serum (HGS) sample was diluted with column buffer and filtered to remove precipitate. The filtered samples were depleted using a MARS-14 column (4.6 × 100 mm, catalog no. 5188-6558, Agilent Technologies). The samples were cooled to 4°C in an autosampler, the depletion column was run at room temperature, and the collected fractions were maintained at 4°C until further analysis. Unbound fractions were collected for further analysis.

[0156] A fixed amount of each clinical serum sample and a second of each HGS was diluted in ammonium bicarbonate buffer and depleted of 14 high and approximately 60 additional moderately rich proteins using a 10 mL bulk material (50% slurry, Sigma) IgY14-SuperMix (Sigma) hand-packed column. Shi et al., Methods, vol. 56 (no. 2): pp. 246-253 (2012). Samples were cooled to 4°C using an autosampler, the depletion column was run at room temperature, and the collected fractions were maintained at 4°C until further analysis. Unbound fractions were collected for further analysis.

[0157] Depleted serum samples were denatured with trifluoroethanol, reduced with dithiotreitol, alkylated with iodoacetamide, and then digested with trypsin in a 1:10 trypsin:protein ratio. Following trypsin digestion, the samples were desalted using a C18 column, and the eluate was lyophilized. The desalted samples were redissolved in a reconstitution solution containing five internal standard peptides.

[0158] The depleted and trypsin-digested samples were analyzed using scheduled multiple reaction monitoring (sMRM). The peptides were analyzed using Waters Nano Acquity. Using UPLC, the samples were separated on a 150 mm × 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). In the sMRM assay, 1708 transitions corresponding to 854 peptides and 236 proteins were measured. Chromatographic peaks were integrated using Rosetta Elucidator software (Ceiba Solutions).

[0159] Transitions were excluded from the analysis if their intensity area counts were less than 10,000, or if they were missing in samples where there were more than 3 per batch. Intensity area counts were logarithmically transformed, and the trend of the mass spectrometry run order and depletion batch effects were minimized using regression analysis.

[0160] (Example 2) Transition analysis for identifying preterm birth biomarkers I The purpose of these analyses was to examine the data collected in Example 1 in order to identify transitions and proteins that predict preterm birth. The specific analyses used were (i) Cox time event analysis and (ii) a model with preterm birth as the binary category dependent variable. The dependent variable for all Cox analyses was gestational age to the event (where the event is preterm birth). For the purposes of the Cox analysis, subjects with preterm birth had the event on their birthday. Subjects with full-term birth were censored on their birthday. Gestational age on the date of sample collection was a covariate in all Cox analyses.

[0161] Assay data were previously adjusted for run order and depletion batches and then logarithmically transformed. Values ​​for gestational age at sample collection were adjusted as follows: Transition values ​​were regressioned against gestational age at sample collection, using only controls (non-preterm subjects). Residuals from the regression were specified as the adjusted values. The adjusted values ​​were used as the binary categorical dependent variable in models involving preterm birth. Unadjusted values ​​were used in Cox analysis.

[0162] Univariate Cox proportional hazards analysis Univariate Cox proportional hazards analysis was performed to predict gestational age at birth, with gestational age at sample collection as a covariate. Table 1 shows transitions with p-values ​​less than 0.05. Five proteins exhibit multiple transitions among those 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 analysis was performed, using a stepped lasso model for variable selection to predict gestational age at birth, with gestational age at the sample collection date as a covariate. These analyses included a total of n=80 subjects and involved 28 PTB events. In the stepped variable selection analysis, the Akaike Information Criterion (AIC) was used as the termination criterion. Table 2 shows the transitions selected by the stepped AIC analysis. The coefficient of determination (R) for the stepped AIC model is shown. 2 ) is 0.86 (not corrected for multiple comparisons).

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

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

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

[0167] For each of the four methods (random forest, boosting, lasso, and logistic regression), each method was capable of selecting and ranking its own best 15 transitions. The inventors then constructed models using transitions 1 through 15. Each method independently reduced the number of nodes sequentially from 15 to 1. A recursive option was used to reduce the number of nodes at each step. Nodes were ranked at each step based on their importance from a nested cross-validation procedure to determine which nodes to remove. The least important nodes were eliminated. The importance metric for lasso and logistic regression is the z-value. For random forest and boosting, the importance of variables was calculated by changing the order of the out-of-bag data. For each tree, the classification error rate for the out-of-bag portion of the data was recorded. The error rate was then recalculated after changing the order of the values ​​for each variable (i.e., transitions). If the transitions were indeed important, there would have been a significant difference between the two error rates. The difference between the two error rates was then averaged across all trees and normalized by the standard deviation of the difference. The AUCs for these models are 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 to that method. These multivariate analyses suggest that models combining three or more transitions give a larger AUC than the 0.7 estimated by bootstrap.

[0168] In multivariate models, the random forest (RF), boosting, and lasso models yielded the best area under the AUROC curve. The following transitions were selected by these models because they were significant in the Cox univariate model and / or had a 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), with gestational age at sample collection as a covariate. In univariate Cox analysis, five proteins with multiple transitions were identified among those 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 multivariate Cox analysis, 24 transitions were selected in the stepwise AIC variable analysis, and three transitions were selected in the lasso model, including the top three proteins from 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 greater were identified. In the multivariate analysis, it is suggested that models combining three or more transitions estimated by bootstrap yield an AUC greater than 0.7.

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

[0172] The samples were processed in four batches, each consisting of 7 cases, 14 corresponding controls, and 3 HGS controls. Serum samples were depleted using MARS14, with 14 of the most abundant serum samples being depleted as described in Example 1. The depleted serum was then reduced with dithiothreitol, alkylated with iodoacetamide, and digested overnight at 37°C with trypsin in a 1:20 trypsin-to-protein ratio. Following trypsin digestion, the samples were desalted and lyophilized using an Empore C18 96-well solid-phase extraction plate (3M Company). The desalted samples were redissolved in a 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 Agilent 6490 Elution occurred in a triple quadrapole mass spectrometer using an acetonitrile gradient.

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

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

[0176] (Example 4) Study on premature birth biomarkers III: Shotgun identification Further studies employed a hypothesis-independent shotgun approach to identify and quantify additional biomarkers not present in our multiplexed hypothesis-dependent MRM assay. Unless otherwise noted, samples were processed as described in the preceding examples.

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

[0178] Using parent and fragment ion data, the human RefSeq database was searched using the Sequest algorithm (Eng et al., J. Am. Soc. Mass Spectrom 1994; Vol. 5: pp. 976-989) and the X!Tandem algorithm (Craig and Beavis, Bioinformatics 2004; Vol. 20: pp. 1466-1467). Sequest data was searched with a tolerance of 20 ppm for parent ions and 1 AMU for fragment ions. Two loss of trypsin cleavage was permitted, and modifications included static carboxyamide methylcysteine ​​and methionine oxidation. After the search, the data was 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 no Xcorr filter was available. Instead, the PeptideProphet algorithm (Keller et al., Anal. Chem 2002; Vol. 74: pp. 5383-5392) was used to validate each X!Tandem peptide spectral assignment, and the protein assignment was validated using the ProteinProphet algorithm (Nesvizhskii et al., Anal. Chem 2002; Vol. 74: pp. 5383-5392). The data was filtered to include only peptide spectral matches with a PeptideProphet probability of 0.9 or greater. After compiling the peptide and protein identifications, the spectral count data for each peptide was imported into DAnTE software (Polpitiya et al., Bioinformatics. 2008; Vol. 24: pp. 1556-1558). Logarithmically transformed data was concentrated to mean values ​​and missing values ​​were filtered by requiring that the peptide be identified in at least 4 cases and 4 controls. To determine the significance of the analytes, receiver operating characteristic (ROC) curves are constructed for each analyte, and the true positive rate (sensitivity) is plotted as a function of the false positive rate (1-specificity) for different thresholds separating the SPTB group from the full-term group.The area under the ROC curve (AUC) is equal to the probability that the classifier ranks a randomly selected positive case higher than a randomly selected negative case. Peptides with an AUC greater than or equal to 0.6, uniquely identified by Sequest or Xtandem, are found in Tables 8 and 9, respectively, while 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 for X!Tandem only (AUC > 0.6)

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 for both Tandem and Sequest (AUC > 0.6) [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 differently expressed proteins identified by the hypothesis-independent strategy described above were not already present in our MRM-MS assay and were candidates for inclusion in the MRM-MS assay. Two additional functionally interesting proteins (AFP, PGH1) were also selected for MRM development. Candidates were prioritized based on AUC and biological function, and novel pathways were prioritized. Sequences for each target protein were imported into Skyline software, thereby generating a list of trypsin peptides, m / z values ​​for parental and fragment ions, and instrument-specific collision energies (McLean et al. Bioinformatics (2010) Vol. 26 (No. 7): pp. 966-968; McLean et al. Anal. Chem (2010) Vol. 82 (No. 24): pp. 10116-10124).

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

[0184] After prioritizing parental 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 were collected using either QTRAP5500 (AB Sciex) or 6490QQQ (Agilent). Commercially available human female serum (from pregnant and non-pregnant donors) was depleted, processed into trypsin peptides as described above, and used to "scan" for the peptides of interest. In some cases, purified synthetic peptides were used for further optimization. For development, digested serum or purified synthetic peptides were separated at 100 μl / min using a 15-minute acetonitrile gradient on a 2.1 × 50 mM Poroshell 120 EC-C18 column (Agilent) at 40°C.

[0185] The MS / MS data was imported back into Skyline, where all chromatograms for each peptide were overlaid and used to identify the consensus peak corresponding to the peptide of interest, as well as the 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 the 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] Next, the top 2–10 transitions per peptide and up to 7 peptides per protein were selected for collision energy (CE) optimization on the Agilent 6490. Using Skyline or MassHunter Qual software, the optimized CE value for each transition was determined based on peak area relative to noise or signal. Two transitions with the maximum peak area per peptide and at least two peptides per protein were selected for the final MRM method. If a transition with a larger peak area had a lower m / z value with a higher background level or greater potential for interference, it was replaced with a transition with a lower peak area.

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

[0189] The above method was typical for most proteins. However, in some cases, differently expressed peptides identified by the shotgun method, for example, in protein families with high sequence identity, were not uniquely identified. In these cases, MRM methods were developed for each family member. It should also be noted that, for any given protein, peptides added to peptides found significant and fragment ions not observed in Orbitrap may 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 confirm biomarkers for preterm birth Further hypothesis-dependent discovery studies were conducted using the scheduled MRM assay used in Example 3, but this time enhanced with 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. The samples included approximately 30 cases and 60 corresponding controls from each of the three gestational periods (early, 17-22 weeks, mid-trim, 23-25 ​​weeks, and late, 26-28 weeks). Logarithmically transformed peak areas for each transition were corrected for run order and batch effects due to regression. The ability of each analyte to separate cases from controls was determined by calculating univariate AUC values ​​from the ROC curves. Ranked univariate AUC values ​​(0.6 or greater) are reported for individual gestational period window sample sets (Tables 12, 13, 15) and combinations of mid- and late-trim windows (Table 14). Multivariate classifiers were constructed using different subsets of analytes (listed below) with lasso and random forest methods. Significant transitions in the lasso method correspond to transitions with non-zero coefficients, and random forest analyte rankings were determined by the Gini importance value (the mean decrease in model precision if the variable were 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 the non-zero lasso coefficients or random forest Gini importance values ​​were summed for each analyte among panels with an AUC of 0.85 or greater.

[0191] Table 12. Individual statistics for the early window. [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 the medium-term window. [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 window [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 Window [Table 22]

[0202] Table 23. Lasso protein metaterm window [Table 23]

[0203] Table 24. Lasso Window (Overall Mid-Term) [Table 24]

[0204] Table 25. Lasso 32 Mid-to-Late Window [Table 25]

[0205] Table 26. Lasso 100 Mid-to-Late Window [Table 26]

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

[0207] Table 28. Lasso All-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 window (all and late stages) [Table 32]

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

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

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

[0215] Table 36. Random Forest Full Early Window [Table 36]

[0216] Table 37. Random Forest SummedGini Early Window [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 Window [Table 38]

[0218] Table 39. Random Forest 100 Medium-Term Window [Table 39]

[0219] Table 40. Random Forest Protein Midterm Window [Table 40]

[0220] Table 41. Random Forest Midterm Window [Table 41]

[0221] Table 42. Random Forest 32 Mid-to-Late Window [Table 42]

[0222] Table 43. Random Forest 100 Mid-to-Late Window [Table 43]

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

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

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

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

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

[0228] Table 49. Random Forest All-Time Window [Table 49]

[0229] Table 50. Selective transitions for early window [Table 50]

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

[0231] Table 52. Selective transitions for mid-to-late window [Table 52]

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

[0233] (Example 6) Research V for further improving preterm birth biomarkers Additional hypothesis-dependent discovery studies were conducted using a further improved, scheduled MRM assay. Less robust transitions were again eliminated, analytical performance was improved, and space was created for the inclusion of stable isotope-labeled standards (SIS) corresponding to the 79 analytes of interest identified in previous studies. SIS peptides have the same amino acid sequence, chromatographic and MS fragmentation behavior as their endogenous peptide counterparts, but differ in mass. Therefore, they can be used to reduce variability in LC-MS analysis and confirm analyte identity. The samples included approximately 60 spontaneous PTB cases (delivery before 37 weeks, day 0) and 180 full-term controls (delivery above or equal to 37 weeks, day 0). For each case, a "corresponding" control was designated within one day of blood collection, and two "random" controls were 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 previously described, except that trypsin digests were reconstituted in a solution containing SIS standards. The peak areas of the raw analytes were Box-Cox transformed, corrected for batch effects by execution order and regression, and used for univariate and multivariate statistical analysis. Univariate analysis included determining p-values ​​for adjusted peak areas for all analytes from t-tests considering case-for-case controls defined as either >37 weeks (Table 54) or >40 weeks (Table 55) of birth. Univariate analysis also included determining p-values ​​for linear models evaluating the dependence of the adjusted peak area of ​​each analyte on time to birth (gestational age at birth minus gestational age at blood collection) (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, and the results are summarized in Tables 58-61.

[0234] A multivariate random forest regression model was constructed using analytic 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 between predicted and actual GABs, and in terms of the mean absolute deviation (MAD) of the predicted GAB from the actual GAB. Accuracy was further evaluated by determining the area under the receiver operating feature curve (AUC) when the predicted GAB was used as a quantitative variable to classify subjects as full-term or preterm. The importance values ​​of the random forest were fitted to an empirical cumulative distribution function, and the probability (P) was calculated. We report the analytics by importance ranking (P>0.7) in the random forest model using adjusted analytic peak area values ​​(Table 62) and analytic / 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. The estimates are based on women enrolled in the Sera PAPR clinical trial, which provides the subjects used to develop the PTB prediction method.

[0236] Among women with a 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 a predicted GAB of j days plus or minus k days who actually gave birth before 37 weeks of 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 is less than the specified gestational age 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 gave birth before the 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 of (i) less than 37 weeks, (ii) 37–39 weeks, and (iii) 40 weeks or more.

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

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

[0240] Table 56. Univariate p-values ​​for adjusted peak area in the time-linear model to birth. [Table 56]

[0241] Table 57. Univariate p-values ​​for adjusted peak area in the linear model of gestational duration at birth. [Table 57-1] [Table 57-2] [Table 57-3]

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

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

[0244] Table 60. Univariate p-values ​​for peak area ratio in the time-linear model to childbirth. [Table 60-1] [Table 60-2]

[0245] Table 61. Univariate p-values ​​for peak area ratio in the linear model of gestational duration at birth. [Table 61]

[0246] Table 62. Random forest importance values ​​using adjusted peak area [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 above description, it will be clear that modified and altered forms can be created for the present invention as described herein to be used in various ways and conditions. Such embodiments are also within the scope of the following claims.

[0249] Any enumeration of elements in any definition of a variable herein includes the definition of that variable as any single element or as a combination (or subcombination) of the enumerated elements. Any enumeration of embodiments herein includes that embodiment as any single embodiment or in combination with any other embodiment or part thereof.

[0250] All patents and publications referenced herein are incorporated by reference to the same degree as each individual patent and publication is specifically and individually indicated as being incorporated by reference.

Claims

[Claim 1] The probability of premature birth.

Images