Microbiome preterm birth DREAM challenge: Crowdsourcing machine learning approaches to advance preterm birth research.
Ontology highlight
ABSTRACT:
SUBMITTER: Golob JL
PROVIDER: S-EPMC10829755 | biostudies-literature | 2024 Jan
REPOSITORIES: biostudies-literature
Publications
Microbiome preterm birth DREAM challenge: Crowdsourcing machine learning approaches to advance preterm birth research.
Golob Jonathan L JL Oskotsky Tomiko T TT Tang Alice S AS Roldan Alennie A Chung Verena V Ha Connie W Y CWY Wong Ronald J RJ Flynn Kaitlin J KJ Parraga-Leo Antonio A Wibrand Camilla C Minot Samuel S SS Oskotsky Boris B Andreoletti Gaia G Kosti Idit I Bletz Julie J Nelson Amber A Gao Jifan J Wei Zhoujingpeng Z Chen Guanhua G Tang Zheng-Zheng ZZ Novielli Pierfrancesco P Romano Donato D Pantaleo Ester E Amoroso Nicola N Monaco Alfonso A Vacca Mirco M De Angelis Maria M Bellotti Roberto R Tangaro Sabina S Kuntzleman Abigail A Bigcraft Isaac I Techtmann Stephen S Bae Daehun D Kim Eunyoung E Jeon Jongbum J Joe Soobok S Theis Kevin R KR Ng Sherrianne S Lee Yun S YS Diaz-Gimeno Patricia P Bennett Phillip R PR MacIntyre David A DA Stolovitzky Gustavo G Lynch Susan V SV Albrecht Jake J Gomez-Lopez Nardhy N Romero Roberto R Stevenson David K DK Aghaeepour Nima N Tarca Adi L AL Costello James C JC Sirota Marina M
Cell reports. Medicine 20231221 1
Every year, 11% of infants are born preterm with significant health consequences, with the vaginal microbiome a risk factor for preterm birth. We crowdsource models to predict (1) preterm birth (PTB; <37 weeks) or (2) early preterm birth (ePTB; <32 weeks) from 9 vaginal microbiome studies representing 3,578 samples from 1,268 pregnant individuals, aggregated from public raw data via phylogenetic harmonization. The predictive models are validated on two independent unpublished datasets representi ...[more]
Similar Datasets
Project description:Globally, every year about 11% of infants are born preterm, defined as a birth prior to 37 weeks of gestation, with significant and lingering health consequences. Multiple studies have related the vaginal microbiome to preterm birth. We present a crowdsourcing approach to predict: (a) preterm or (b) early preterm birth from 9 publicly available vaginal microbiome studies representing 3,578 samples from 1,268 pregnant individuals, aggregated from raw sequences via an open-source tool, MaLiAmPi. We validated the crowdsourced models on novel datasets representing 331 samples from 148 pregnant individuals. From 318 DREAM challenge participants we received 148 and 121 submissions for our two separate prediction sub-challenges with top-ranking submissions achieving bootstrapped AUROC scores of 0.69 and 0.87, respectively. Alpha diversity, VALENCIA community state types, and composition (via phylotype relative abundance) were important features in the top performing models, most of which were tree based methods. This work serves as the foundation for subsequent efforts to translate predictive tests into clinical practice, and to better understand and prevent preterm birth.
| S-EPMC10029035 | biostudies-literature
Project description:Preterm birth prediction is essential for improving neonatal outcomes. While many machine learning techniques have been applied to predict preterm birth using health records, inflammatory markers, and vaginal microbiome data, the role of prenatal oral microbiome remains unclear. This study aimed to compare oral microbiome compositions between a preterm and a full-term birth group, identify oral microbiome associated with preterm birth, and develop a preterm birth prediction model using machine learning of oral microbiome compositions. Participants included singleton pregnant women admitted to Jeonbuk National University Hospital between 2019 and 2021. Subjects were divided into a preterm and a full-term birth group based on pregnancy outcomes. Oral microbiome samples were collected using mouthwash within 24 h before delivery and 16S ribosomal RNA sequencing was performed to analyze taxonomy. Differentially abundant taxa were identified using DESeq2. A random forest classifier was applied to predict preterm birth based on the oral microbiome. A total of 59 women participated in this study, with 30 in the preterm birth group and 29 in the full-term birth group. There was no significant difference in maternal clinical characteristics between the preterm and the full-birth group. Twenty-five differentially abundant taxa were identified, including 22 full-term birth-enriched taxa and 3 preterm birth-enriched taxa. The random forest classifier achieved high balanced accuracies (0.765 ± 0.071) using the 9 most important taxa. Our study identified 25 differentially abundant taxa that could differentiate preterm and full-term birth groups. A preterm birth prediction model was developed using machine learning of oral microbiome compositions in mouthwash samples. Findings of this study suggest the potential of using oral microbiome for predicting preterm birth. Further multi-center and larger studies are required to validate our results before clinical applications.
| S-EPMC10689490 | biostudies-literature
Project description:Consumer wearables and sensors are a rich source of data about patients' daily disease and symptom burden, particularly in the case of movement disorders like Parkinson's disease (PD). However, interpreting these complex data into so-called digital biomarkers requires complicated analytical approaches, and validating these biomarkers requires sufficient data and unbiased evaluation methods. Here we describe the use of crowdsourcing to specifically evaluate and benchmark features derived from accelerometer and gyroscope data in two different datasets to predict the presence of PD and severity of three PD symptoms: tremor, dyskinesia, and bradykinesia. Forty teams from around the world submitted features, and achieved drastically improved predictive performance for PD status (best AUROC = 0.87), as well as tremor- (best AUPR = 0.75), dyskinesia- (best AUPR = 0.48) and bradykinesia-severity (best AUPR = 0.95).
| S-EPMC7979931 | biostudies-literature
Project description:Microbiome data predictive analysis within a machine learning (ML) workflow presents numerous domain-specific challenges involving preprocessing, feature selection, predictive modeling, performance estimation, model interpretation, and the extraction of biological information from the results. To assist decision-making, we offer a set of recommendations on algorithm selection, pipeline creation and evaluation, stemming from the COST Action ML4Microbiome. We compared the suggested approaches on a multi-cohort shotgun metagenomics dataset of colorectal cancer patients, focusing on their performance in disease diagnosis and biomarker discovery. It is demonstrated that the use of compositional transformations and filtering methods as part of data preprocessing does not always improve the predictive performance of a model. In contrast, the multivariate feature selection, such as the Statistically Equivalent Signatures algorithm, was effective in reducing the classification error. When validated on a separate test dataset, this algorithm in combination with random forest modeling, provided the most accurate performance estimates. Lastly, we showed how linear modeling by logistic regression coupled with visualization techniques such as Individual Conditional Expectation (ICE) plots can yield interpretable results and offer biological insights. These findings are significant for clinicians and non-experts alike in translational applications.
| S-EPMC10556866 | biostudies-literature
Project description:BackgroundThe etiology of preterm birth (PTB) is heterogeneous and not yet well known. Maternal periodontal disease has been investigated for decades and is a known risk factor for adverse pregnancy outcomes. However, no particular bacterial species or higher taxonomic order has been found as causative of PTB, leading to studies of the whole oral microbiome. In order to determine if and how the composition of the oral microbiome is associated with PTB, we performed a large case-control study including women with term (TB) and PTB.MethodsWe compared oral microbiomes in PTB to TB, to examine differences in the microbial richness, diversity, and differential abundance of specific taxa. We obtained oral swab samples from 152 Caucasian pregnant women who were classified as either PTB (≤36 6/7 weeks, n = 61) or TB (≥38 0/7 weeks, n = 91) in exclusion of any other major medical or obstetric conditions. The oral microbiomes of these women were characterized by 16S ribosomal RNA (rRNA) gene sequencing of the V3-V4 region on the MiSeq platform.ResultsThe dominant microorganisms at the phylum level in all pregnant women regardless of birth week outcomes as belonging to Firmicutes, Proteobacteria, Bacteroidetes, Fusobacteria, and Actinobacteria. The phyla Firmicutes and Bacteroidetes were relatively more abundant in women with a PTB than in women with a TB, while Proteobacteria was less prevalent in women with a PTB. At the genus level, Veillonella, Prevotella, and Capnocytophaga were enriched in the PTB, and while many of the members of these genera could not be resolved to the species level, Veillonella massillensis was shown to be increased in the PTB group.ConclusionWe identified the genera Veillonella, Prevotella, and Capnocytophaga in the maternal oral microbiome as being associated with PTB independently of clinically apparent infection, uterine anomalies, and other pregnancy complications, including placenta previa, and placental abruption. The clarification of the role of those taxa in the etiology of PTB merits further research.
| S-EPMC10441245 | biostudies-literature
Project description:To predict preterm birth (PTB) in multiparous women, comparing machine learning approaches with traditional logistic regression. A population-based cohort study was conducted using data from the Ontario Better Outcomes Registry and Network (BORN). The cohort included all multiparous women who delivered a singleton birth at 20-42 weeks' gestation in an Ontario hospital between April 1, 2012 and March 31, 2014. The primary outcome was PTB < 37 weeks, with spontaneous PTB analyzed as a secondary outcome. Stepwise logistic regression and the Boruta machine learning were used to select the important variables during the first and second trimester. For building prediction models, the whole data set were divided for the two independent parts: two-third for training the classifiers (Logistic regression, random forests, decision trees, and artificial neural networks) and one-third for model validation. Then, the training data set were balanced by random over sampling technique. The best hyper parameters were obtained by the tenfold cross validation. The performance of all models was evaluated by sensitivity, specificity, positive predictive value, negative predictive value, and the area under the receiver operating characteristics (AUC). The cohort included 145,846 births, of which 8125 (5.57%) were preterm. In first-trimester models, the strongest predictors of PTB were previous PTB, preexisting diabetes, and abnormal pregnancy-associated plasma protein-A. In the testing data set, the highest predictive ability was seen for artificial neural networks, with an area under the receiver operating characteristic curve (AUC) of 68.8% (95% CI 67.6-70.1%). In second-trimester models, addition of infant sex, attendance at first-trimester appointment, medication exposure, and abnormal alpha-fetoprotein concentrations increased the AUC to 72.1% (95% CI 71.1-73.1%) with logistic regression. With the inclusion of the variable complications during pregnancy, the AUC increased to 80.5% (95% CI 79.6-81.5%) using logistic regression. For both overall and spontaneous PTB, during both the first and second trimesters, models yielded negative predictive values of 97%. Overall, machine learning and logistic regression produced similar performance for prediction of PTB. For overall and spontaneous PTB, both first- and second-trimester models provided negative predictive values of ~ 97%, higher than that of fetal fibronectin.
| S-EPMC11415355 | biostudies-literature
Project description:The incidence of preterm birth exceeds 10% worldwide. There are significant disparities in the frequency of preterm birth among populations within countries, and women of African ancestry disproportionately bear the burden of risk in the United States. In the present study, we report a community resource that includes 'omics' data from approximately 12,000 samples as part of the integrative Human Microbiome Project. Longitudinal analyses of 16S ribosomal RNA, metagenomic, metatranscriptomic and cytokine profiles from 45 preterm and 90 term birth controls identified harbingers of preterm birth in this cohort of women predominantly of African ancestry. Women who delivered preterm exhibited significantly lower vaginal levels of Lactobacillus crispatus and higher levels of BVAB1, Sneathia amnii, TM7-H1, a group of Prevotella species and nine additional taxa. The first representative genomes of BVAB1 and TM7-H1 are described. Preterm-birth-associated taxa were correlated with proinflammatory cytokines in vaginal fluid. These findings highlight new opportunities for assessment of the risk of preterm birth.
| S-EPMC6750801 | biostudies-literature
Project description:Identification of pregnancies at risk of preterm birth (PTB), the leading cause of newborn deaths, remains challenging given the syndromic nature of the disease. We report a longitudinal multi-omics study coupled with a DREAM challenge to develop predictive models of PTB. The findings indicate that whole-blood gene expression predicts ultrasound-based gestational ages in normal and complicated pregnancies (r = 0.83) and, using data collected before 37 weeks of gestation, also predicts the delivery date in both normal pregnancies (r = 0.86) and those with spontaneous preterm birth (r = 0.75). Based on samples collected before 33 weeks in asymptomatic women, our analysis suggests that expression changes preceding preterm prelabor rupture of the membranes are consistent across time points and cohorts and involve leukocyte-mediated immunity. Models built from plasma proteomic data predict spontaneous preterm delivery with intact membranes with higher accuracy and earlier in pregnancy than transcriptomic models (AUROC = 0.76 versus AUROC = 0.6 at 27-33 weeks of gestation).
| S-EPMC8233692 | biostudies-literature
Project description:BackgroundWe examined over a million California birth records for 2010 through 2021 to investigate whether disparities in preterm birth (PTB) by nativity and race support the widely held but hitherto unsubstantiated belief that genetic differences explain the persistent Black-White disparity in PTB.MethodsWe examined PTB rates and risk ratios among African-, Caribbean-, and U.S.-born Black women compared to U.S.-born White women. Multivariate analyses adjusted for maternal age, education, number of live births, delivery payer, trimester of prenatal care initiation, pre-pregnancy BMI, smoking, and prevalence of poverty in a woman's residence census tract; and for paternal education.ResultsIn adjusted analyses, African-born Black women's PTB rates were no different from those of U.S.-born White women.DiscussionThe results add to prior evidence making a genetic etiology for the racial disparity in PTB unlikely. If genetic differences tied to "race" explained the Black-White disparity in PTB among U.S.-born women, the African immigrants in this study would have had higher rates of PTB, not the lower rates observed. Multiple explanations for the observed patterns and their implications are discussed. Failure to distinguish causes of PTB from causes of the racial disparity in PTB have likely contributed to erroneous attribution of the racial disparity to genetic differences. Based on the literature, unmeasured experiences of racism, including racism-related stress and adverse environmental exposures, are plausible explanations for the PTB disparity between Black and White U.S.-born women. The favorable birth outcomes of African-born Black immigrants may reflect less exposure to racism during sensitive life periods, e.g., childhood, when they were in African countries, where Black people are in the racial majority.
| S-EPMC10794556 | biostudies-literature
Project description:Previous studies have investigated the associations between the vaginal microbiome and preterm birth, with the aim of determining whether differences in community patterns meaningfully alter risk and could therefore be the target of intervention. We report on vaginal microbial analysis of a nested case-control subset of the Pregnancy, Infection, and Nutrition (PIN) Study, including 464 White women (375 term birth and 89 spontaneous preterm birth, sPTB) and 360 Black women (276 term birth and 84 sPTB). We found that the microbiome of Black women has higher alpha-diversity, higher abundance of Lactobacillus iners, and lower abundance of Lactobacillus crispatus. However, among women who douche, there were no significant differences in microbiome by race. The sPTB-associated microbiome exhibited a lower abundance of L. crispatus, while alpha diversity and L. iners were not significantly associated with sPTB. For each order of magnitude increase in the normalized relative abundance of L. crispatus, multivariable adjusted odds of sPTB decreased by approximately 20% (odds ratio, 0.81; 95% confidence interval, 0.70, 0.94). When we considered the impact of douching, associations between the microbiome and sPTB were limited to women who do not douche. We also observed strong intercorrelations between a range of maternal factors, including poverty, education, marital status, age, douching, and race, with microbiome effect sizes in the range of 1.8 to 5.2% in univariate models. Therefore, race may simply be a proxy for other socially driven factors that differentiate microbiome community structures. Future work will continue to refine reliable microbial biomarkers for preterm birth across diverse cohorts. IMPORTANCE Approximately 10% of all pregnancies in the United States end in preterm birth, and over 14% of pregnancies end in preterm birth among Black women. Knowledge on the associations between vaginal microbiome and preterm birth is important for understanding the potential cause and assessing risk of preterm birth. Our study is one of the largest studies performed to date to investigate the associations between vaginal microbiome and spontaneous preterm birth (sPTB), with stratified design for Black and White women. We found that the vaginal microbiome was different between Black and White women. The vaginal microbiome was associated with sPTB, and a lower abundance of L. crispatus increased the risk of sPTB independent of racial differences in microbial community structures. Furthermore, we also found that vaginal douching obscured the associations between vaginal microbiome, race, and preterm birth, suggesting that vaginal douching is an important factor to consider in future studies.
| S-EPMC9238383 | biostudies-literature