Ivana Maric

All Publications

• Discovery of sparse, reliable omic biomarkers with Stabl. Nature biotechnology Hédou, J., Marić, I., Bellan, G., Einhaus, J., Gaudillière, D. K., Ladant, F. X., Verdonk, F., Stelzer, I. A., Feyaerts, D., Tsai, A. S., Ganio, E. A., Sabayev, M., Gillard, J., Amar, J., Cambriel, A., Oskotsky, T. T., Roldan, A., Golob, J. L., Sirota, M., Bonham, T. A., Sato, M., Diop, M., Durand, X., Angst, M. S., Stevenson, D. K., Aghaeepour, N., Montanari, A., Gaudillière, B. 2024

  Abstract

  Adoption of high-content omic technologies in clinical studies, coupled with computational methods, has yielded an abundance of candidate biomarkers. However, translating such findings into bona fide clinical biomarkers remains challenging. To facilitate this process, we introduce Stabl, a general machine learning method that identifies a sparse, reliable set of biomarkers by integrating noise injection and a data-driven signal-to-noise threshold into multivariable predictive modeling. Evaluation of Stabl on synthetic datasets and five independent clinical studies demonstrates improved biomarker sparsity and reliability compared to commonly used sparsity-promoting regularization methods while maintaining predictive performance; it distills datasets containing 1,400-35,000 features down to 4-34 candidate biomarkers. Stabl extends to multi-omic integration tasks, enabling biological interpretation of complex predictive models, as it hones in on a shortlist of proteomic, metabolomic and cytometric events predicting labor onset, microbial biomarkers of pre-term birth and a pre-operative immune signature of post-surgical infections. Stabl is available at https://github.com/gregbellan/Stabl .

  View details for DOI 10.1038/s41587-023-02033-x

  View details for PubMedID 38168992

  View details for PubMedCentralID 7003173

• Early prediction and longitudinal modeling of preeclampsia from multiomics. Patterns (New York, N.Y.) Maric, I., Contrepois, K., Moufarrej, M. N., Stelzer, I. A., Feyaerts, D., Han, X., Tang, A., Stanley, N., Wong, R. J., Traber, G. M., Ellenberger, M., Chang, A. L., Fallahzadeh, R., Nassar, H., Becker, M., Xenochristou, M., Espinosa, C., De Francesco, D., Ghaemi, M. S., Costello, E. K., Culos, A., Ling, X. B., Sylvester, K. G., Darmstadt, G. L., Winn, V. D., Shaw, G. M., Relman, D. A., Quake, S. R., Angst, M. S., Snyder, M. P., Stevenson, D. K., Gaudilliere, B., Aghaeepour, N. 2022; 3 (12): 100655

  Abstract

  Preeclampsia is a complex disease of pregnancy whose physiopathology remains unclear. We developed machine-learning models for early prediction of preeclampsia (first 16weeks of pregnancy) and over gestation by analyzing six omics datasets from a longitudinal cohort of pregnant women. For early pregnancy, a prediction model using nine urine metabolites had the highest accuracy and was validated on an independent cohort (area under the receiver-operating characteristic curve [AUC]= 0.88, 95% confidence interval [CI] [0.76, 0.99] cross-validated; AUC= 0.83, 95% CI [0.62,1] validated). Univariate analysis demonstrated statistical significance of identified metabolites. An integrated multiomics model further improved accuracy (AUC= 0.94). Several biological pathways were identified including tryptophan, caffeine, and arachidonic acid metabolisms. Integration with immune cytometry data suggested novel associations between immune and proteomic dynamics. While further validation in a larger population is necessary, these encouraging results can serve as a basis for a simple, early diagnostic test for preeclampsia.

  View details for DOI 10.1016/j.patter.2022.100655

  View details for PubMedID 36569558

• Decreased Mortality Rate Among COVID-19 Patients Prescribed Statins: Data From Electronic Health Records in the US Frontiers in Medicine Maric, I., Oskotsky, T., Kosti, I., Le, B., Wong, R. J., Shaw, G. M., Sirota, M., Stevenson, D. K. 2021; 8

  View details for DOI 10.3389/fmed.2021.639804

• Data-Driven Modeling of Pregnancy-Related Complications. Trends in molecular medicine Espinosa, C. n., Becker, M. n., Marić, I. n., Wong, R. J., Shaw, G. M., Gaudilliere, B. n., Aghaeepour, N. n., Stevenson, D. K. 2021

  Abstract

  A healthy pregnancy depends on complex interrelated biological adaptations involving placentation, maternal immune responses, and hormonal homeostasis. Recent advances in high-throughput technologies have provided access to multiomics biological data that, combined with clinical and social data, can provide a deeper understanding of normal and abnormal pregnancies. Integration of these heterogeneous datasets using state-of-the-art machine-learning methods can enable the prediction of short- and long-term health trajectories for a mother and offspring and the development of treatments to prevent or minimize complications. We review advanced machine-learning methods that could: provide deeper biological insights into a pregnancy not yet unveiled by current methodologies; clarify the etiologies and heterogeneity of pathologies that affect a pregnancy; and suggest the best approaches to address disparities in outcomes affecting vulnerable populations.

  View details for DOI 10.1016/j.molmed.2021.01.007

  View details for PubMedID 33573911

• Mortality Risk Among Patients With COVID-19 Prescribed Selective Serotonin Reuptake Inhibitor Antidepressants. JAMA network open Oskotsky, T., Maric, I., Tang, A., Oskotsky, B., Wong, R. J., Aghaeepour, N., Sirota, M., Stevenson, D. K. 2021; 4 (11): e2133090

  Abstract

  Antidepressant use may be associated with reduced levels of several proinflammatory cytokines suggested to be involved with the development of severe COVID-19. An association between the use of selective serotonin reuptake inhibitors (SSRIs)-specifically fluoxetine hydrochloride and fluvoxamine maleate-with decreased mortality among patients with COVID-19 has been reported in recent studies; however, these studies had limited power due to their small size.To investigate the association of SSRIs with outcomes in patients with COVID-19 by analyzing electronic health records (EHRs).This retrospective cohort study used propensity score matching by demographic characteristics, comorbidities, and medication indication to compare SSRI-treated patients with matched control patients not treated with SSRIs within a large EHR database representing a diverse population of 83 584 patients diagnosed with COVID-19 from January to September 2020 and with a duration of follow-up of as long as 8 months in 87 health care centers across the US.Selective serotonin reuptake inhibitors and specifically (1) fluoxetine, (2) fluoxetine or fluvoxamine, and (3) other SSRIs (ie, not fluoxetine or fluvoxamine).Death.A total of 3401 adult patients with COVID-19 prescribed SSRIs (2033 women [59.8%]; mean [SD] age, 63.8 [18.1] years) were identified, with 470 receiving fluoxetine only (280 women [59.6%]; mean [SD] age, 58.5 [18.1] years), 481 receiving fluoxetine or fluvoxamine (285 women [59.3%]; mean [SD] age, 58.7 [18.0] years), and 2898 receiving other SSRIs (1733 women [59.8%]; mean [SD] age, 64.7 [18.0] years) within a defined time frame. When compared with matched untreated control patients, relative risk (RR) of mortality was reduced among patients prescribed any SSRI (497 of 3401 [14.6%] vs 1130 of 6802 [16.6%]; RR, 0.92 [95% CI, 0.85-0.99]; adjusted P = .03); fluoxetine (46 of 470 [9.8%] vs 937 of 7050 [13.3%]; RR, 0.72 [95% CI, 0.54-0.97]; adjusted P = .03); and fluoxetine or fluvoxamine (48 of 481 [10.0%] vs 956 of 7215 [13.3%]; RR, 0.74 [95% CI, 0.55-0.99]; adjusted P = .04). The association between receiving any SSRI that is not fluoxetine or fluvoxamine and risk of death was not statistically significant (447 of 2898 [15.4%] vs 1474 of 8694 [17.0%]; RR, 0.92 [95% CI, 0.84-1.00]; adjusted P = .06).These results support evidence that SSRIs may be associated with reduced severity of COVID-19 reflected in the reduced RR of mortality. Further research and randomized clinical trials are needed to elucidate the effect of SSRIs generally, or more specifically of fluoxetine and fluvoxamine, on the severity of COVID-19 outcomes.

  View details for DOI 10.1001/jamanetworkopen.2021.33090

  View details for PubMedID 34779847

• Towards personalized medicine in maternal and child health: integrating biologic and social determinants. Pediatric research Stevenson, D. K., Wong, R. J., Aghaeepour, N., Maric, I., Angst, M. S., Contrepois, K., Darmstadt, G. L., Druzin, M. L., Eisenberg, M. L., Gaudilliere, B., Gibbs, R. S., Gotlib, I. H., Gould, J. B., Lee, H. C., Ling, X. B., Mayo, J. A., Moufarrej, M. N., Quaintance, C. C., Quake, S. R., Relman, D. A., Sirota, M., Snyder, M. P., Sylvester, K. G., Hao, S., Wise, P. H., Shaw, G. M., Katz, M. 2020

  View details for DOI 10.1038/s41390-020-0981-8

  View details for PubMedID 32454518

• Early Prediction of Preeclampsia via Machine Learning American Journal of Obstetrics & Gynecology MFM Maric, I., Tsur, A., Aghaeepour, N., Montanari, A., Stevenson, D. K., Shaw, G. M., Winn, V. D. 2020; 2 (2)
• Deep representation learning identifies associations between physical activity and sleep patterns during pregnancy and prematurity. NPJ digital medicine Ravindra, N. G., Espinosa, C., Berson, E., Phongpreecha, T., Zhao, P., Becker, M., Chang, A. L., Shome, S., Marić, I., De Francesco, D., Mataraso, S., Saarunya, G., Thuraiappah, M., Xue, L., Gaudillière, B., Angst, M. S., Shaw, G. M., Herzog, E. D., Stevenson, D. K., England, S. K., Aghaeepour, N. 2023; 6 (1): 171

  Abstract

  Preterm birth (PTB) is the leading cause of infant mortality globally. Research has focused on developing predictive models for PTB without prioritizing cost-effective interventions. Physical activity and sleep present unique opportunities for interventions in low- and middle-income populations (LMICs). However, objective measurement of physical activity and sleep remains challenging and self-reported metrics suffer from low-resolution and accuracy. In this study, we use physical activity data collected using a wearable device comprising over 181,944 h of data across N = 1083 patients. Using a new state-of-the art deep learning time-series classification architecture, we develop a 'clock' of healthy dynamics during pregnancy by using gestational age (GA) as a surrogate for progression of pregnancy. We also develop novel interpretability algorithms that integrate unsupervised clustering, model error analysis, feature attribution, and automated actigraphy analysis, allowing for model interpretation with respect to sleep, activity, and clinical variables. Our model performs significantly better than 7 other machine learning and AI methods for modeling the progression of pregnancy. We found that deviations from a normal 'clock' of physical activity and sleep changes during pregnancy are strongly associated with pregnancy outcomes. When our model underestimates GA, there are 0.52 fewer preterm births than expected (P = 1.01e - 67, permutation test) and when our model overestimates GA, there are 1.44 times (P = 2.82e - 39, permutation test) more preterm births than expected. Model error is negatively correlated with interdaily stability (P = 0.043, Spearman's), indicating that our model assigns a more advanced GA when an individual's daily rhythms are less precise. Supporting this, our model attributes higher importance to sleep periods in predicting higher-than-actual GA, relative to lower-than-actual GA (P = 1.01e - 21, Mann-Whitney U). Combining prediction and interpretability allows us to signal when activity behaviors alter the likelihood of preterm birth and advocates for the development of clinical decision support through passive monitoring and exercise habit and sleep recommendations, which can be easily implemented in LMICs.

  View details for DOI 10.1038/s41746-023-00911-x

  View details for PubMedID 37770643

  View details for PubMedCentralID 3796350

• Persistent Bacterial Vaginosis and Risk for Spontaneous Preterm Birth. American journal of perinatology Blumenfeld, Y. J., Marić, I., Stevenson, D. K., Gibbs, R. S., Shaw, G. M. 2023

  Abstract

   The aim of this study was to determine the association between persistent bacterial vaginosis (BV) in pregnancy and risk for spontaneous preterm birth (sPTB). Retrospective data from IBM MarketScan Commercial Database were analyzed. Women aged between 12 and 55 years with singleton gestations were included and linked to an outpatient medications database and medications prescribed during the pregnancy were analyzed. BV in pregnancy was determined based on both a diagnosis of BV and treatment with metronidazole and/or clindamycin, and persistent treatment of BV was defined as BV in more than one trimester or BV requiring more than one antibiotic prescription. Odds ratios were calculated comparing sPTB frequencies in those with BV, or persistent BV, to women without BV in pregnancy. Survival analysis using Kaplan-Meier curves for the gestational age at delivery was also performed. Among a cohort of 2,538,606 women, 216,611 had an associated International Classification of Diseases, 9th Revision or 10th Revision code for diagnosis of BV alone, and 63,817 had both a diagnosis of BV and were treated with metronidazole and/or clindamycin. Overall, the frequency of sPTB among women treated with BV was 7.5% compared with 5.7% for women without BV who did not receive antibiotics. Relative to those without BV in pregnancy, odds ratios for sPTB were highest in those treated for BV in both the first and second trimester (1.66 [95% confidence interval [CI]: 1.52, 1.81]) or those with three or more prescriptions in pregnancy (1.48 [95% CI: 1.35, 1.63]. Persistent BV may have a higher risk for sPTB than a single episode of BV in pregnancy.· Persistent BV beyond one trimester may increase the risk for sPTB.. · Persistent BV requiring more than one prescription may increase the risk for sPTB.. · Almost half of antibiotic prescriptions treating BV in pregnancy are filled after 20 weeks gestation..

  View details for DOI 10.1055/s-0043-1770703

  View details for PubMedID 37379861

• Development of a Urine Metabolomics Biomarker-Based Prediction Model for Preeclampsia during Early Pregnancy. Metabolites Zhang, Y., Sylvester, K. G., Jin, B., Wong, R. J., Schilling, J., Chou, C. J., Han, Z., Luo, R. Y., Tian, L., Ladella, S., Mo, L., Maric, I., Blumenfeld, Y. J., Darmstadt, G. L., Shaw, G. M., Stevenson, D. K., Whitin, J. C., Cohen, H. J., McElhinney, D. B., Ling, X. B. 2023; 13 (6)

  Abstract

  Preeclampsia (PE) is a condition that poses a significant risk of maternal mortality and multiple organ failure during pregnancy. Early prediction of PE can enable timely surveillance and interventions, such as low-dose aspirin administration. In this study, conducted at Stanford Health Care, we examined a cohort of 60 pregnant women and collected 478 urine samples between gestational weeks 8 and 20 for comprehensive metabolomic profiling. By employing liquid chromatography mass spectrometry (LCMS/MS), we identified the structures of seven out of 26 metabolomics biomarkers detected. Utilizing the XGBoost algorithm, we developed a predictive model based on these seven metabolomics biomarkers to identify individuals at risk of developing PE. The performance of the model was evaluated using 10-fold cross-validation, yielding an area under the receiver operating characteristic curve of 0.856. Our findings suggest that measuring urinary metabolomics biomarkers offers a noninvasive approach to assess the risk of PE prior to its onset.

  View details for DOI 10.3390/metabo13060715

  View details for PubMedID 37367874

• Multiomic signals associated with maternal epidemiological factors contributing to preterm birth in low- and middle-income countries. Science advances Espinosa, C. A., Khan, W., Khanam, R., Das, S., Khalid, J., Pervin, J., Kasaro, M. P., Contrepois, K., Chang, A. L., Phongpreecha, T., Michael, B., Ellenberger, M., Mehmood, U., Hotwani, A., Nizar, A., Kabir, F., Wong, R. J., Becker, M., Berson, E., Culos, A., De Francesco, D., Mataraso, S., Ravindra, N., Thuraiappah, M., Xenochristou, M., Stelzer, I. A., Marić, I., Dutta, A., Raqib, R., Ahmed, S., Rahman, S., Hasan, A. S., Ali, S. M., Juma, M. H., Rahman, M., Aktar, S., Deb, S., Price, J. T., Wise, P. H., Winn, V. D., Druzin, M. L., Gibbs, R. S., Darmstadt, G. L., Murray, J. C., Stringer, J. S., Gaudilliere, B., Snyder, M. P., Angst, M. S., Rahman, A., Baqui, A. H., Jehan, F., Nisar, M. I., Vwalika, B., Sazawal, S., Shaw, G. M., Stevenson, D. K., Aghaeepour, N. 2023; 9 (21): eade7692

  Abstract

  Preterm birth (PTB) is the leading cause of death in children under five, yet comprehensive studies are hindered by its multiple complex etiologies. Epidemiological associations between PTB and maternal characteristics have been previously described. This work used multiomic profiling and multivariate modeling to investigate the biological signatures of these characteristics. Maternal covariates were collected during pregnancy from 13,841 pregnant women across five sites. Plasma samples from 231 participants were analyzed to generate proteomic, metabolomic, and lipidomic datasets. Machine learning models showed robust performance for the prediction of PTB (AUROC = 0.70), time-to-delivery (r = 0.65), maternal age (r = 0.59), gravidity (r = 0.56), and BMI (r = 0.81). Time-to-delivery biological correlates included fetal-associated proteins (e.g., ALPP, AFP, and PGF) and immune proteins (e.g., PD-L1, CCL28, and LIFR). Maternal age negatively correlated with collagen COL9A1, gravidity with endothelial NOS and inflammatory chemokine CXCL13, and BMI with leptin and structural protein FABP4. These results provide an integrated view of epidemiological factors associated with PTB and identify biological signatures of clinical covariates affecting this disease.

  View details for DOI 10.1126/sciadv.ade7692

  View details for PubMedID 37224249

• Postpartum long-acting reversible contraception among privately insured: national analysis 2007-2016, by term and preterm birth. Contraception Shaw, J. G., Goldthwaite, L. M., Marić, I., Shaw, K. A., Stevenson, D. K., Shaw, G. M. 2023: 110065

  Abstract

  To investigate postpartum long-acting reversible contraception (LARC) use among privately insured women, with specific consideration of use after preterm delivery.We used the national IBM® MarketScan® Commercial Database to identify singleton deliveries from 2007-2016, spontaneous preterm birth, and follow up ≤12 weeks postpartum. We assessed ≤12 week postpartum LARC placement overall and after spontaneous preterm deliveries, across study years. We examined timing of placement, rates of postpartum follow-up, and state-level variation in postpartum LARC.Among 3,132,107 singleton deliveries, 6.6% were spontaneous preterm. Over the time period, total postpartum LARC use increased: 4.8% to 11.7% for intrauterine devices (IUDs), 0.2% to 2.4% for implants. In 2016, those who experienced a spontaneous preterm birth were less likely to initiate postpartum IUDs compared to their peers (10.2% vs 11.8%, p<0.001), minimally more likely to initiate implants (2.7% vs 2.4%, p=0.04) and more likely to present for postpartum care (61.7% vs 55.9%, p<0.001). LARC placement prior to hospital discharge was rare (preterm: 8 per 10,000 deliveries vs all others: 6.3 per 10,000 deliveries, p=0.002). State level analysis showed wide variation in postpartum LARC (range 6%-32%).While postpartum LARC use increased among the privately insured 2007 to 2016, few received LARC prior to hospital discharge. Those experiencing preterm birth were no more likely to receive inpatient LARC. Postpartum follow-up remained low and regional variation of LARC was high, highlighting the need for efforts to remove barriers to inpatient postpartum LARC for all who desire it-public and privately insured alike.Among the half of U.S. births that are privately insured, postpartum LARC is increasing after both term and preterm births, yet exceedingly few (<0.1%) received LARC prior to hospital discharge.

  View details for DOI 10.1016/j.contraception.2023.110065

  View details for PubMedID 37210023

• Target-agnostic drug prediction integrated with medical record analysis uncovers differential associations of statins with increased survival in COVID-19 patients. PLoS computational biology Sperry, M. M., Oskotsky, T. T., Marić, I., Kaushal, S., Takeda, T., Horvath, V., Powers, R. K., Rodas, M., Furlong, B., Soong, M., Prabhala, P., Goyal, G., Carlson, K. E., Wong, R. J., Kosti, I., Le, B. L., Logue, J., Hammond, H., Frieman, M., Stevenson, D. K., Ingber, D. E., Sirota, M., Novak, R. 2023; 19 (5): e1011050

  Abstract

  Drug repurposing requires distinguishing established drug class targets from novel molecule-specific mechanisms and rapidly derisking their therapeutic potential in a time-critical manner, particularly in a pandemic scenario. In response to the challenge to rapidly identify treatment options for COVID-19, several studies reported that statins, as a drug class, reduce mortality in these patients. However, it is unknown if different statins exhibit consistent function or may have varying therapeutic benefit. A Bayesian network tool was used to predict drugs that shift the host transcriptomic response to SARS-CoV-2 infection towards a healthy state. Drugs were predicted using 14 RNA-sequencing datasets from 72 autopsy tissues and 465 COVID-19 patient samples or from cultured human cells and organoids infected with SARS-CoV-2. Top drug predictions included statins, which were then assessed using electronic medical records containing over 4,000 COVID-19 patients on statins to determine mortality risk in patients prescribed specific statins versus untreated matched controls. The same drugs were tested in Vero E6 cells infected with SARS-CoV-2 and human endothelial cells infected with a related OC43 coronavirus. Simvastatin was among the most highly predicted compounds (14/14 datasets) and five other statins, including atorvastatin, were predicted to be active in > 50% of analyses. Analysis of the clinical database revealed that reduced mortality risk was only observed in COVID-19 patients prescribed a subset of statins, including simvastatin and atorvastatin. In vitro testing of SARS-CoV-2 infected cells revealed simvastatin to be a potent direct inhibitor whereas most other statins were less effective. Simvastatin also inhibited OC43 infection and reduced cytokine production in endothelial cells. Statins may differ in their ability to sustain the lives of COVID-19 patients despite having a shared drug target and lipid-modifying mechanism of action. These findings highlight the value of target-agnostic drug prediction coupled with patient databases to identify and clinically evaluate non-obvious mechanisms and derisk and accelerate drug repurposing opportunities.

  View details for DOI 10.1371/journal.pcbi.1011050

  View details for PubMedID 37146076

• Large-scale correlation network construction for unraveling the coordination of complex biological systems. Nature computational science Becker, M., Nassar, H., Espinosa, C., Stelzer, I. A., Feyaerts, D., Berson, E., Bidoki, N. H., Chang, A. L., Saarunya, G., Culos, A., De Francesco, D., Fallahzadeh, R., Liu, Q., Kim, Y., Marić, I., Mataraso, S. J., Payrovnaziri, S. N., Phongpreecha, T., Ravindra, N. G., Stanley, N., Shome, S., Tan, Y., Thuraiappah, M., Xenochristou, M., Xue, L., Shaw, G., Stevenson, D., Angst, M. S., Gaudilliere, B., Aghaeepour, N. 2023; 3 (4): 346-359

  Abstract

  Advanced measurement and data storage technologies have enabled high-dimensional profiling of complex biological systems. For this, modern multiomics studies regularly produce datasets with hundreds of thousands of measurements per sample, enabling a new era of precision medicine. Correlation analysis is an important first step to gain deeper insights into the coordination and underlying processes of such complex systems. However, the construction of large correlation networks in modern high-dimensional datasets remains a major computational challenge owing to rapidly growing runtime and memory requirements. Here we address this challenge by introducing CorALS (Correlation Analysis of Large-scale (biological) Systems), an open-source framework for the construction and analysis of large-scale parametric as well as non-parametric correlation networks for high-dimensional biological data. It features off-the-shelf algorithms suitable for both personal and high-performance computers, enabling workflows and downstream analysis approaches. We illustrate the broad scope and potential of CorALS by exploring perspectives on complex biological processes in large-scale multiomics and single-cell studies.

  View details for DOI 10.1038/s43588-023-00429-y

  View details for PubMedID 38116462

  View details for PubMedCentralID PMC10727505

• Breath: The Exhaust of Metabolism. The Journal of pediatrics Stevenson, D. K., Marić, I., Wong, R. J. 2023

  View details for DOI 10.1016/j.jpeds.2023.03.002

  View details for PubMedID 36925060

• Data-driven longitudinal characterization of neonatal health and morbidity. Science translational medicine De Francesco, D., Reiss, J. D., Roger, J., Tang, A. S., Chang, A. L., Becker, M., Phongpreecha, T., Espinosa, C., Morin, S., Berson, E., Thuraiappah, M., Le, B. L., Ravindra, N. G., Payrovnaziri, S. N., Mataraso, S., Kim, Y., Xue, L., Rosenstein, M. G., Oskotsky, T., Marić, I., Gaudilliere, B., Carvalho, B., Bateman, B. T., Angst, M. S., Prince, L. S., Blumenfeld, Y. J., Benitz, W. E., Fuerch, J. H., Shaw, G. M., Sylvester, K. G., Stevenson, D. K., Sirota, M., Aghaeepour, N. 2023; 15 (683): eadc9854

  Abstract

  Although prematurity is the single largest cause of death in children under 5 years of age, the current definition of prematurity, based on gestational age, lacks the precision needed for guiding care decisions. Here, we propose a longitudinal risk assessment for adverse neonatal outcomes in newborns based on a deep learning model that uses electronic health records (EHRs) to predict a wide range of outcomes over a period starting shortly before conception and ending months after birth. By linking the EHRs of the Lucile Packard Children's Hospital and the Stanford Healthcare Adult Hospital, we developed a cohort of 22,104 mother-newborn dyads delivered between 2014 and 2018. Maternal and newborn EHRs were extracted and used to train a multi-input multitask deep learning model, featuring a long short-term memory neural network, to predict 24 different neonatal outcomes. An additional cohort of 10,250 mother-newborn dyads delivered at the same Stanford Hospitals from 2019 to September 2020 was used to validate the model. Areas under the receiver operating characteristic curve at delivery exceeded 0.9 for 10 of the 24 neonatal outcomes considered and were between 0.8 and 0.9 for 7 additional outcomes. Moreover, comprehensive association analysis identified multiple known associations between various maternal and neonatal features and specific neonatal outcomes. This study used linked EHRs from more than 30,000 mother-newborn dyads and would serve as a resource for the investigation and prediction of neonatal outcomes. An interactive website is available for independent investigators to leverage this unique dataset: https://maternal-child-health-associations.shinyapps.io/shiny_app/.

  View details for DOI 10.1126/scitranslmed.adc9854

  View details for PubMedID 36791208

• Omics approaches: interactions at the maternal-fetal interface and origins of child health and disease. Pediatric research Ozen, M., Aghaeepour, N., Maric, I., Wong, R. J., Stevenson, D. K., Jantzie, L. L. 2022

  Abstract

  Immunoperinatology is an emerging field. Transdisciplinary efforts by physicians, physician-scientists, basic science researchers, and computational biologists have made substantial advancements by identifying unique immunologic signatures of specific diseases, discovering innovative preventative or treatment strategies, and establishing foundations for individualized neonatal intensive care of the most vulnerable neonates. In this review, we summarize the immunobiology and immunopathology of pregnancy, highlight omics approaches to study the maternal-fetal interface, and their contributions to pregnancy health. We examined the importance of transdisciplinary, multiomic (such as genomics, transcriptomics, proteomics, metabolomics, and immunomics) and machine-learning strategies in unraveling the mechanisms of adverse pregnancy, neonatal, and childhood outcomes and how they can guide the development of novel therapies to improve maternal and neonatal health. IMPACT: Discuss immunoperinatology research from the lens of omics and machine-learning approaches. Identify opportunities for omics-based approaches to delineate infection/inflammation-associated maternal, neonatal, and later life adverse outcomes (e.g., histologic chorioamnionitis [HCA]).

  View details for DOI 10.1038/s41390-022-02335-x

  View details for PubMedID 36216868

• Information Theoretic Perspectives on 5G Systems and Beyond edited by Maric, I., Shamai (Shitz), S., Simeone, O. Cambridge University Press. 2022
• Multiomics Characterization of Preterm Birth in Low- and Middle-Income Countries (vol 3, e2029655, 2020) JAMA NETWORK OPEN Jehan, F., Sazawal, S., Baqui, A. H. 2021; 4 (2)

  View details for DOI 10.1001/jamanetworkopen.2021.0399

  View details for Web of Science ID 000617274100003

• DECREASED MORTALITY RATE AMONG COVID-19 PATIENTS USING STATINS: DATA FROM US ELECTRONIC HEALTH RECORDS Oskotsky, T., Maric, I., Kosti, I., Le, B. L., Wong, R. J., Shaw, G. M., Sirota, M., Stevenson, D. K. BMJ PUBLISHING GROUP. 2021: 219–20

  View details for DOI 10.1136/jim-2021-WRMC.262

  View details for Web of Science ID 000608729100280

• PERSISTENT BACTERIAL VAGINOSIS AND RISK FOR SPONTANEOUS PRETERM BIRTH Blumenfeld, Y. J., Maric, I., Stevenson, D. K., Shaw, G. M. BMJ PUBLISHING GROUP. 2021: 127–28

  View details for DOI 10.1136/jim-2021-WRMC.58

  View details for Web of Science ID 000608729100076

• Integration of mechanistic immunological knowledge into a machine learning pipeline improves predictions NATURE MACHINE INTELLIGENCE Culos, A., Tsai, A. S., Stanley, N., Becker, M., Ghaemi, M. S., McIlwain, D. R., Fallahzadeh, R., Tanada, A., Nassar, H., Espinosa, C., Xenochristou, M., Ganio, E., Peterson, L., Han, X., Stelzer, I. A., Ando, K., Gaudilliere, D., Phongpreecha, T., Maric, I., Chang, A. L., Shaw, G. M., Stevenson, D. K., Bendall, S., Davis, K. L., Fantl, W., Nolan, G. P., Hastie, T., Tibshirani, R., Angst, M. S., Gaudilliere, B., Aghaeepour, N. 2020

  View details for DOI 10.1038/s42256-020-00232-8

  View details for Web of Science ID 000579336000001

• VoPo leverages cellular heterogeneity for predictive modeling of single-cell data. Nature communications Stanley, N. n., Stelzer, I. A., Tsai, A. S., Fallahzadeh, R. n., Ganio, E. n., Becker, M. n., Phongpreecha, T. n., Nassar, H. n., Ghaemi, S. n., Maric, I. n., Culos, A. n., Chang, A. L., Xenochristou, M. n., Han, X. n., Espinosa, C. n., Rumer, K. n., Peterson, L. n., Verdonk, F. n., Gaudilliere, D. n., Tsai, E. n., Feyaerts, D. n., Einhaus, J. n., Ando, K. n., Wong, R. J., Obermoser, G. n., Shaw, G. M., Stevenson, D. K., Angst, M. S., Gaudilliere, B. n., Aghaeepour, N. n. 2020; 11 (1): 3738

  Abstract

  High-throughput single-cell analysis technologies produce an abundance of data that is critical for profiling the heterogeneity of cellular systems. We introduce VoPo (https://github.com/stanleyn/VoPo), a machine learning algorithm for predictive modeling and comprehensive visualization of the heterogeneity captured in large single-cell datasets. In three mass cytometry datasets, with the largest measuring hundreds of millions of cells over hundreds of samples, VoPo defines phenotypically and functionally homogeneous cell populations. VoPo further outperforms state-of-the-art machine learning algorithms in classification tasks, and identified immune-correlates of clinically-relevant parameters.

  View details for DOI 10.1038/s41467-020-17569-8

  View details for PubMedID 32719375

• Changes in pregnancy-related serum biomarkers early in gestation are associated with later development of preeclampsia. PloS one Hao, S. n., You, J. n., Chen, L. n., Zhao, H. n., Huang, Y. n., Zheng, L. n., Tian, L. n., Maric, I. n., Liu, X. n., Li, T. n., Bianco, Y. K., Winn, V. D., Aghaeepour, N. n., Gaudilliere, B. n., Angst, M. S., Zhou, X. n., Li, Y. M., Mo, L. n., Wong, R. J., Shaw, G. M., Stevenson, D. K., Cohen, H. J., Mcelhinney, D. B., Sylvester, K. G., Ling, X. B. 2020; 15 (3): e0230000

  Abstract

  Placental protein expression plays a crucial role during pregnancy. We hypothesized that: (1) circulating levels of pregnancy-associated, placenta-related proteins throughout gestation reflect the temporal progression of the uncomplicated, full-term pregnancy, and can effectively estimate gestational ages (GAs); and (2) preeclampsia (PE) is associated with disruptions in these protein levels early in gestation; and can identify impending PE. We also compared gestational profiles of proteins in the human and mouse, using pregnant heme oxygenase-1 (HO-1) heterozygote (Het) mice, a mouse model reflecting PE-like symptoms.Serum levels of placenta-related proteins-leptin (LEP), chorionic somatomammotropin hormone like 1 (CSHL1), elabela (ELA), activin A, soluble fms-like tyrosine kinase 1 (sFlt-1), and placental growth factor (PlGF)-were quantified by ELISA in blood serially collected throughout human pregnancies (20 normal subjects with 66 samples, and 20 subjects who developed PE with 61 samples). Multivariate analysis was performed to estimate the GA in normal pregnancy. Mean-squared errors of GA estimations were used to identify impending PE. The human protein profiles were then compared with those in the pregnant HO-1 Het mice.An elastic net-based gestational dating model was developed (R2 = 0.76) and validated (R2 = 0.61) using serum levels of the 6 proteins measured at various GAs from women with normal uncomplicated pregnancies. In women who developed PE, the model was not (R2 = -0.17) associated with GA. Deviations from the model estimations were observed in women who developed PE (P = 0.01). The model developed with 5 proteins (ELA excluded) performed similarly from sera from normal human (R2 = 0.68) and WT mouse (R2 = 0.85) pregnancies. Disruptions of this model were observed in both human PE-associated (R2 = 0.27) and mouse HO-1 Het (R2 = 0.30) pregnancies. LEP outperformed sFlt-1 and PlGF in differentiating impending PE at early human and late mouse GAs.Serum placenta-related protein profiles are temporally regulated throughout normal pregnancies and significantly disrupted in women who develop PE. LEP changes earlier than the well-established biomarkers (sFlt-1 and PlGF). There may be evidence of a causative action of HO-1 deficiency in LEP upregulation in a PE-like murine model.

  View details for DOI 10.1371/journal.pone.0230000

  View details for PubMedID 32126118

• Increased Carbon Monoxide Washout Rates in Newborn Infants. Neonatology Stevenson, D. K., Wong, R. J., Ostrander, C. R., Maric, I., Vreman, H. J., Cohen, R. S. 2019: 1–5

  Abstract

  BACKGROUND: Endogenous carbon monoxide (CO) production is primarily due to heme degradation, which also results in the equimolar production of bilirubin. Thus, estimates of total body CO production can serve as indices of total body bilirubin formation. The elimination rate of CO from a person's body (CO washout rate) after exposure to an elevated ambient CO concentration is determined by a variety of factors, and is very different between babies and adults.OBJECTIVE: We determined CO washout rates for babies using a simplified technique to measure total body CO excretion rates (VeCO).METHODS: Using a simplified technique, we measured the times to reach an approximate steady state after a change in ambient CO concentration (decay time constant) and CO washout rates in normal newborn infants using non-linear least squares curve fitting.RESULTS: We found a mean CO washout time of 18.7 ± 4.2 min and a CO equilibration (decay time) constant of 0.12 ± 0.04/min (0.08-0.21) for newborn infants.CONCLUSIONS: We confirm that CO washout rates for babies are much faster than those for adults. Therefore, measurements of carboxyhemoglobin (COHb) or end-tidal CO (ETCO), corrected for ambient CO, (COHbc and ETCOc, respectively) can be used as surrogates for VeCO and can provide accurate estimates of endogenous CO (VCO) and bilirubin production rates under normal environmental conditions. Such measurements can be used to identify infants with severe hyperbilirubinemia due to hemolysis and thus at high risk for bilirubin neurotoxicity.

  View details for DOI 10.1159/000503635

  View details for PubMedID 31634890

• Data-driven queries between medications and spontaneous preterm birth among 2.5 million pregnancies. Birth defects research Marić, I. n., Winn, V. D., Borisenko, E. n., Weber, K. A., Wong, R. J., Aziz, N. n., Blumenfeld, Y. J., El-Sayed, Y. Y., Stevenson, D. K., Shaw, G. M. 2019

  Abstract

  Our goal was to develop an approach that can systematically identify potential associations between medication prescribed in pregnancy and spontaneous preterm birth (sPTB) by mining large administrative "claims" databases containing hundreds of medications. One such association that we illustrate emerged with antiviral medications used for herpes treatment.IBM MarketScan® databases (2007-2016) were used. A pregnancy cohort was established using International Classification of Diseases (ICD-9/10) codes. Multiple hypothesis testing and the Benjamini-Hochberg procedure that limited false discovery rate at 5% revealed, among 863 medications, five that showed odds ratios (ORs) <1. The statistically strongest was an association between antivirals and sPTB that we illustrate as a real example of our approach, specifically for treatment of genital herpes (GH). Three groups of women were identified based on diagnosis of GH and treatment during the first 36 weeks of pregnancy: (a) GH without treatment; (b) GH treated with antivirals; (c) no GH or treatment.We identified 2,538,255 deliveries. 0.98% women had a diagnosis of GH. Among them, 60.0% received antiviral treatment. Women with treated GH had OR < 1, (OR [95% CI] = 0.91 [0.85, 0.98]). In contrast, women with untreated GH had a small increased risk of sPTB (OR [95% CI] =1.22 [1.14, 1.32]).Data-driven approaches can effectively generate new hypotheses on associations between medications and sPTB. This analysis led us to examine the association with GH treatment. While unknown confounders may impact these findings, our results indicate that women with untreated GH have a modest increased risk of sPTB.

  View details for DOI 10.1002/bdr2.1580

  View details for PubMedID 31433567

• Maternal Height and Risk of Preeclampsia among Race/Ethnic Groups. American journal of perinatology Maric, I., Mayo, J. A., Druzin, M. L., Wong, R. J., Winn, V. D., Stevenson, D. K., Shaw, G. M. 2018

  Abstract

  OBJECTIVE: Shorter maternal height has been associated with preeclampsia risk in several populations. It has been less evident whether an independent contribution to the risk exists from maternal height consistently across different races/ethnicities. We investigated associations between maternal height and risk of preeclampsia for different races/ethnicities.STUDY DESIGN: California singleton live births from 2007 to 2011 were analyzed. Logistic regression was used to estimate adjusted odds ratios for the association between height and preeclampsia after stratification by race/ethnicity. To determine the contribution of height that is as independent of body composition as possible, we performed one analysis adjusted for body mass index (BMI) and the other for weight. Additional analyses were performed stratified by parity, and the presence of preexisting/gestational diabetes and autoimmune conditions.RESULTS: Among 2,138,012 deliveries, 3.1% preeclampsia/eclampsia cases were observed. The analysis, adjusted for prepregnancy weight, revealed an inverse relation between maternal height and risk of mild and severe preeclampsia/eclampsia. When the analysis was adjusted for BMI, an inverse relation between maternal height was observed for severe preeclampsia/eclampsia. These associations were observed for each race/ethnicity.CONCLUSION: Using a large and diverse cohort, we demonstrated that shorter height, irrespective of prepregnancy weight or BMI, is associated with an increased risk of severe preeclampsia/eclampsia across different races/ethnicities.

  View details for PubMedID 30396225

• Residential agricultural pesticide exposures and risks of preeclampsia ENVIRONMENTAL RESEARCH Shaw, G. M., Yang, W., Roberts, E. M., Aghaeepour, N., Mayo, J. A., Weber, K. A., Maric, I., Carmichael, S. L., Winn, V. D., Stevenson, D. K., English, P. B. 2018; 164: 546–55

  View details for DOI 10.1016/j.envres.2018.03.020

  View details for Web of Science ID 000431387100065

• Capacity-Achieving Rate-Compatible Polar Codes IEEE Trans. Information Theory Hong, S., Hui, D., Maric, I. 2017; 63 (12): 7620-7632
• Capacity-Achieving Rate-Compatible Polar Codes for General Channels IEEE Wireless Communications and Networking Conference Workshops (WCNCW) Mondelli, M., Hassani, H., Marić, I., Hui, D., Hong, S. 2017

  View details for DOI 10.1109/WCNCW.2017.7919107

• Short Message Noisy Network Coding With Sliding-Window Decoding for Half-Duplex Multihop Relay Networks IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS Hong, S., Maric, I., Hui, D. 2016; 15 (10): 6676-6689

  View details for DOI 10.1109/TWC.2016.2587627

  View details for Web of Science ID 000386068800012

• Capacity-Achieving Rate-Compatible Polar Codes 2016 IEEE INTERNATIONAL SYMPOSIUM ON INFORMATION THEORY Hong, S., Hui, D., Maric, I. 2016: 41-45

  View details for Web of Science ID 000390098700009

• Diversity-Multiplexing Tradeoff for the Interference Channel With a Relay IEEE TRANSACTIONS ON INFORMATION THEORY Zahavi, D., Zhang, L., Maric, I., Dabora, R., Goldsmith, A. J., Cui, S. 2015; 61 (2): 963-982

  View details for DOI 10.1109/TIT.2014.2381664

  View details for Web of Science ID 000348298400020

• On the Achievable Rates of Multihop Virtual Full-Duplex Relay Channels 2015 IEEE INTERNATIONAL SYMPOSIUM ON INFORMATION THEORY (ISIT) Hong, S., Maric, I., Hui, D., Caire, G. 2015: 2246-2250

  View details for Web of Science ID 000380904702060

• Multihop Virtual Full-Duplex Relay Channels 2015 IEEE INFORMATION THEORY WORKSHOP (ITW) Hong, S., Maric, I., Hui, D., Caire, G. 2015

  View details for Web of Science ID 000381501800073

• Enhanced Relay Cooperation via Rate Splitting CONFERENCE RECORD OF THE 2014 FORTY-EIGHTH ASILOMAR CONFERENCE ON SIGNALS, SYSTEMS & COMPUTERS Maric, I., Hui, D. 2014: 225-229

  View details for Web of Science ID 000370964900041

• Capacity Bounds and Exact Results for the Cognitive Z-Interference Channel IEEE TRANSACTIONS ON INFORMATION THEORY Liu, N., Maric, I., Goldsmith, A. J., Shamai (Shitz), S. 2013; 59 (2): 886-893

  View details for DOI 10.1109/TIT.2012.2225138

  View details for Web of Science ID 000313728800011

• Low Latency Communications 2013 INFORMATION THEORY AND APPLICATIONS WORKSHOP (ITA) Maric, I. 2013

  View details for Web of Science ID 000321214400036

• Capacity of Cognitive Radio Networks Multiantenna Channels: Principles of Cognitive Radio Goldsmith, A., Maric, I. Cambridge Press. 2013
• Relaying in the Presence of Interference: Achievable Rates, Interference Forwarding, and Outer Bounds IEEE TRANSACTIONS ON INFORMATION THEORY Maric, I., Dabora, R., Goldsmith, A. J. 2012; 58 (7): 4342-4354

  View details for DOI 10.1109/TIT.2012.2191710

  View details for Web of Science ID 000305575000019

• Multihop Analog Network Coding via Amplify-and-Forward: The High SNR Regime IEEE TRANSACTIONS ON INFORMATION THEORY Maric, I., Goldsmith, A., Medard, M. 2012; 58 (2): 793-803

  View details for DOI 10.1109/TIT.2011.2173721

  View details for Web of Science ID 000300246900020

• Bandwidth and Power Allocation for Cooperative Strategies in Gaussian Relay Networks IEEE TRANSACTIONS ON INFORMATION THEORY Maric, I., Yates, R. D. 2010; 56 (4): 1880-1889

  View details for DOI 10.1109/TIT.2010.2040875

  View details for Web of Science ID 000275999500033

• Breaking Spectrum Gridlock With Cognitive Radios: An Information Theoretic Perspective PROCEEDINGS OF THE IEEE Goldsmith, A., Jafar, S. A., Maric, I., Srinivasa, S. 2009; 97 (5): 894-914

  View details for DOI 10.1109/JPROC.2009.2015717

  View details for Web of Science ID 000266174300008

• On the capacity of interference channels with one cooperating transmitter EUROPEAN TRANSACTIONS ON TELECOMMUNICATIONS Maric, I., Goldsmith, A., Kramer, G., Shamai (Shitz), S. 2008; 19 (4): 405-420

  View details for DOI 10.1002/ett.1298

  View details for Web of Science ID 000257012200006

• Discrete memoryless interference and broadcast channels with confidential, messages: Secrecy rate regions IEEE Information Theory and Applications Workshop Liu, R., Maric, I., Spasojevic, P., Yates, R. D. IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC. 2008: 2493–2507

  View details for DOI 10.1109/TIT.2008.921879

  View details for Web of Science ID 000258033700007

• On the Capacity of the Interference Channel with a Relay IEEE International Symposium on Information Theory Marric, I., Dabora, R., Goldsmith, A. IEEE. 2008: 554–558

  View details for Web of Science ID 000260364400112

• Capacity of interference channels with partial transmitter cooperation IEEE TRANSACTIONS ON INFORMATION THEORY Maric, I., Yates, R. D., Kramer, G. 2007; 53 (10): 3536-3548

  View details for DOI 10.1109/TIT.2007.904792

  View details for Web of Science ID 000249774000013

• Joint relaying and network coding in wireless networks IEEE International Symposium on Information Theory Katti, S., Maric, I., Goldsmith, A., Katabi, D., Medard, M. IEEE. 2007: 1101–1105

  View details for Web of Science ID 000257010201052

• Cooperative Communications Foundations and Trends in Networking Kramer, G., Maric, I., Yates, R. D. NOW Publishers Inc. 2006; 1
• Cooperative multicast for maximum network lifetime IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS Maric, I., Yates, R. D. 2005; 23 (1): 127-135

  View details for DOI 10.1109/JSAC.2004.837343

  View details for Web of Science ID 000226063100012

• Cooperative multihop broadcast for wireless networks IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS Maric, Yates, R. D. 2004; 22 (6): 1080–88

  View details for DOI 10.1109/JSAC.2004.830912

  View details for Web of Science ID 000223150700013