Delayed colonization of Bifidobacterium spp. and low prevalence of B. infantis among infants of Asian ancestry born in Singapore: insights from the GUSTO cohort study.

Background: The loss of ancestral microbes, or the "disappearing microbiota hypothesis" has been proposed to play a critical role in the rise of inflammatory and immune diseases in developed nations. The effect of this loss is most consequential during early-life, as initial colonizers of the newborn gut contribute significantly to the development of the immune system. Methods: In this longitudinal study (day 3, week 3, and month 3 post-birth) of infants of Asian ancestry born in Singapore, we studied how generational immigration status and common perinatal factors affect bifidobacteria and Bifidobacterium longum subsp. infantis (B. infantis) colonization. Cohort registry identifier: NCT01174875. Results: Our findings show that first-generation migratory status, perinatal antibiotics usage, and cesarean section birth, significantly influenced the abundance and acquisition of bifidobacteria in the infant gut. Most importantly, 95.6% of the infants surveyed in this study had undetectable B. infantis, an early and beneficial colonizer of infant gut due to its ability to metabolize the wide variety of human milk oligosaccharides present in breastmilk and its ability to shape the development of a healthy immune system. A comparative analysis of B. infantis in 12 countries by their GDP per capita showed a remarkably low prevalence of this microbe in advanced economies, especially Singapore. Conclusion: This study provides new insights into infant gut microbiota colonization, showing the impact of generational immigration on early-life gut microbiota acquisition. It also warrants the need to closely monitor the declining prevalence of beneficial microbes such as B. infantis in developed nations and its potential link to increasing autoimmune and allergic diseases.

Trigonelline is an NAD+ precursor that improves muscle function during ageing and is reduced in human sarcopenia.

Mitochondrial dysfunction and low nicotinamide adenine dinucleotide (NAD+) levels are hallmarks of skeletal muscle ageing and sarcopenia1-3, but it is unclear whether these defects result from local changes or can be mediated by systemic or dietary cues. Here we report a functional link between circulating levels of the natural alkaloid trigonelline, which is structurally related to nicotinic acid4, NAD+ levels and muscle health in multiple species. In humans, serum trigonelline levels are reduced with sarcopenia and correlate positively with muscle strength and mitochondrial oxidative phosphorylation in skeletal muscle. Using naturally occurring and isotopically labelled trigonelline, we demonstrate that trigonelline incorporates into the NAD+ pool and increases NAD+ levels in Caenorhabditis elegans, mice and primary myotubes from healthy individuals and individuals with sarcopenia. Mechanistically, trigonelline does not activate GPR109A but is metabolized via the nicotinate phosphoribosyltransferase/Preiss-Handler pathway5,6 across models. In C. elegans, trigonelline improves mitochondrial respiration and biogenesis, reduces age-related muscle wasting and increases lifespan and mobility through an NAD+-dependent mechanism requiring sirtuin. Dietary trigonelline supplementation in male mice enhances muscle strength and prevents fatigue during ageing. Collectively, we identify nutritional supplementation of trigonelline as an NAD+-boosting strategy with therapeutic potential for age-associated muscle decline.

Umbilical Cord Plasma Lysophospholipids and Triacylglycerols Associated with Birthweight Percentiles.

Dysregulated transplacental lipid transfer and fetal-placental lipid metabolism affect birthweight, as does maternal hyperglycemia. As the mechanisms are unclear, we aimed to identify the lipids in umbilical cord plasma that were most associated with birthweight. Seventy-five Chinese women with singleton pregnancies recruited into the GUSTO mother-offspring cohort were selected from across the glycemic range based on a mid-gestation 75 g oral glucose tolerance test, excluding pre-existing diabetes. Cord plasma samples collected at term delivery were analyzed using targeted liquid-chromatography tandem mass-spectrometry to determine the concentrations of 404 lipid species across 17 lipid classes. The birthweights were standardized for sex and gestational age by local references, and regression analyses were adjusted for the maternal age, BMI, parity, mode of delivery, insulin treatment, and fasting/2 h glucose, with a false discovery-corrected p < 0.05 considered significant. Ten lysophosphatidylcholines (LPCs) and two lysophosphatidylethanolamines were positively associated with the birthweight percentiles, while twenty-four triacylglycerols were negatively associated with the birthweight percentiles. The topmost associated lipid was LPC 20:2 [21.28 (95%CI 12.70, 29.87) percentile increase in the standardized birthweight with each SD-unit increase in log10-transformed concentration]. Within these same regression models, maternal glycemia did not significantly associate with the birthweight percentiles. Specific fetal circulating lysophospholipids and triacylglycerols associate with birthweight independently of maternal glycemia, but a causal relationship remains to be established.

Multigenerational adversity impacts on human gut microbiome composition and socioemotional functioning in early childhood.

Adversity exposures in the prenatal and postnatal period are associated with an increased risk for psychopathology, which can be perpetuated across generations. Nonhuman animal research highlights the gut microbiome as a putative biological mechanism underlying such generational risks. In a sample of 450 mother-child dyads living in Singapore, we examined associations between three distinct adversity exposures experienced across two generations-maternal childhood maltreatment, maternal prenatal anxiety, and second-generation children's exposure to stressful life events-and the gut microbiome composition of second-generation children at 2 y of age. We found distinct differences in gut microbiome profiles linked to each adversity exposure, as well as some nonaffected microbiome features (e.g., beta diversity). Remarkably, some of the microbial taxa associated with concurrent and prospective child socioemotional functioning shared overlapping putative functions with those affected by adversity, suggesting that the intergenerational transmission of adversity may have a lasting impact on children's mental health via alterations to gut microbiome functions. Our findings open up a new avenue of research into the underlying mechanisms of intergenerational transmission of mental health risks and the potential of the gut microbiome as a target for intervention.

A five-safes approach to a secure and scalable genomics data repository.

Genomic researchers increasingly utilize commercial cloud service providers (CSPs) to manage data and analytics needs. CSPs allow researchers to grow Information Technology (IT) infrastructure on demand to overcome bottlenecks when combining large datasets. However, without adequate security controls, the risk of unauthorized access may be higher for data stored on the cloud. Additionally, regulators are mandating data access patterns and specific security protocols for the storage and use of genomic data. While CSP provides tools for security and regulatory compliance, building the necessary controls required for cloud solutions is not trivial. Research Assets Provisioning and Tracking Online Repository (RAPTOR) by the Genome Institute of Singapore is a cloud-native genomics data repository and analytics platform that implements a "five-safes" framework to provide security and governance controls to data contributors and users, leveraging CSP for sharing and analysis of genomic datasets without the risk of security breaches or running afoul of regulations.

Safety, Tolerability, and Pharmacokinetics of ß-Cryptoxanthin Supplementation in Healthy Women: A Double-Blind, Randomized, Placebo-Controlled Clinical Trial.

BACKGROUND: ß-cryptoxanthin is a dietary carotenoid for which there have been few studies on the safety and pharmacokinetics following daily oral supplementation. METHODS: 90 healthy Asian women between 21 and 35 years were randomized into three groups: 3 and 6 mg/day oral ß-cryptoxanthin, and placebo. At 2, 4, and 8 weeks of supplementation, plasma carotenoid levels were measured. The effects of ß-cryptoxanthin on blood retinoid-dependent gene expression, mood, physical activity and sleep, metabolic parameters, and fecal microbial composition were investigated. RESULTS: ß-cryptoxanthin supplementation for 8 weeks (3 and 6 mg/day) was found to be safe and well tolerated. Plasma ß-cryptoxanthin concentration was significantly higher in the 6 mg/day group (9.0 ± 4.1 µmol/L) compared to 3 mg/day group (6.0 ± 2.6 µmol/L) (p < 0.03), and placebo (0.4 ± 0.1 µmol/L) (p < 0.001) after 8 weeks. Plasma all-trans retinol, α-cryptoxanthin, α-carotene, ß-carotene, lycopene, lutein, and zeaxanthin levels were not significantly changed. No effects were found on blood retinol-dependent gene expression, mood, physical activity and sleep, metabolic parameters, and fecal microbial composition. CONCLUSIONS: Oral ß-cryptoxanthin supplementation over 8 weeks lead to high plasma concentrations of ß-cryptoxanthin, with no impact on other carotenoids, and was well tolerated in healthy women.

Evidence for inefficient contraction and abnormal mitochondrial activity in sarcopenia using magnetic resonance spectroscopy.

BACKGROUND: Mitochondrial dysfunction has been implicated in sarcopenia. 31 P magnetic resonance spectroscopy (MRS) enables non-invasive measurement of adenosine triphosphate (ATP) synthesis rates to probe mitochondrial function. Here, we assessed muscle energetics in older sarcopenic and non-sarcopenic men and compared with muscle biopsy-derived markers of mitochondrial function. METHODS: Twenty Chinese men with sarcopenia (SARC, age = 73.1 ± 4.1 years) and 19 healthy aged and sex-matched controls (CON, age = 70.3 ± 4.2 years) underwent assessment of strength, physical performance, and magnetic resonance imaging. Concentrations of phosphocreatine (PCr), ATP and inorganic phosphate (Pi) as well as muscle pH were measured at rest and during an interleaved rest-exercise protocol to probe muscle mitochondrial function. Results were compared to biopsy-derived mitochondrial complex activity and expression to understand underlying metabolic perturbations. RESULTS: Despite matched muscle contractile power (strength/cross-sectional area), the ATP contractile cost was higher in SARC compared with CON (low-intensity exercise: 1.06 ± 0.59 vs. 0.57 ± 0.22, moderate: 0.93 ± 0.43 vs. 0.58 ± 0.68, high: 0.70 ± 0.57 vs. 0.43 ± 0.51 mmol L-1  min-1  bar-1  cm-2 , P = 0.003, <0.0001 and <0.0001, respectively). Post-exercise mitochondrial oxidative synthesis rates (a marker of mitochondrial function) tended to be longer in SARC but did not reach significance (17.3 ± 6.4 vs. 14.6 ± 6.5 mmol L-1  min-1 , P = 0.2). However, relative increases in end-exercise ADP in SARC (31.8 ± 9.9 vs. 24.0 ± 7.3 mmol L-1 , P = 0.008) may have been a compensatory mechanism. Mitochondrial complex activity was found to be associated with exercise-induced drops in PCr [citrate synthetase activity (CS), Spearman correlation rho = -0.42, P = 0.03] and end-exercise ADP (complex III, rho = -0.52, P = 0.01; CS rho = -0.45, P = 0.02; SDH rho = -0.45, P = 0.03), with CS also being strongly associated with the PCr recovery rate following low intensity exercise (rho = -0.47, P = 0.02), and the cost of contraction at high intensity (rho = -0.54, P = 0.02). Interestingly, at high intensity, the fractional contribution of oxidative phosphorylation to exercise was correlated with activity in complex II (rho = 0.5, P = 0.03), CS (rho = 0.47, P = 0.02) and SDH (rho = 0.46, P = 0.03), linking increased mitochondrial complex activity with increased ability to generate energy through oxidative pathways. CONCLUSIONS: This study used 31 P MRS to assess ATP utilization and resynthesis in sarcopenic muscle and demonstrated abnormal increases in the energy cost during exercise and perturbed mitochondrial energetics in recovery. Associations between mitochondrial complex activity and the fractional contribution to energy requirement during exercise indicate increased ability to generate energy oxidatively in those with better mitochondrial complex activity.

Plasma lipidomic profiling reveals metabolic adaptations to pregnancy and signatures of cardiometabolic risk: a preconception and longitudinal cohort study.

BACKGROUND: Adaptations in lipid metabolism are essential to meet the physiological demands of pregnancy and any aberration may result in adverse outcomes for both mother and offspring. However, there is a lack of population-level studies to define the longitudinal changes of maternal circulating lipids from preconception to postpartum in relation to cardiometabolic risk factors. METHODS: LC-MS/MS-based quantification of 689 lipid species was performed on 1595 plasma samples collected at three time points in a preconception and longitudinal cohort, Singapore PREconception Study of long-Term maternal and child Outcomes (S-PRESTO). We mapped maternal plasma lipidomic profiles at preconception (N = 976), 26-28 weeks' pregnancy (N = 337) and 3 months postpartum (N = 282) to study longitudinal lipid changes and their associations with cardiometabolic risk factors including pre-pregnancy body mass index, body weight changes and glycaemic traits. RESULTS: Around 56% of the lipids increased and 24% decreased in concentration in pregnancy before returning to the preconception concentration at postpartum, whereas around 11% of the lipids went through significant changes in pregnancy and their concentrations did not revert to the preconception concentrations. We observed a significant association of body weight changes with lipid changes across different physiological states, and lower circulating concentrations of phospholipids and sphingomyelins in pregnant mothers with higher pre-pregnancy BMI. Fasting plasma glucose and glycated haemoglobin (HbA1c) concentrations were lower whereas the homeostatic model assessment of insulin resistance (HOMA-IR), 2-h post-load glucose and fasting insulin concentrations were higher in pregnancy as compared to both preconception and postpartum. Association studies of lipidomic profiles with these glycaemic traits revealed their respective lipid signatures at three physiological states. Assessment of glycaemic traits in relation to the circulating lipids at preconception with a large sample size (n = 936) provided an integrated view of the effects of hyperglycaemia on plasma lipidomic profiles. We observed a distinct relationship of lipidomic profiles with different measures, with the highest percentage of significant lipids associated with HOMA-IR (58.9%), followed by fasting insulin concentration (56.9%), 2-h post-load glucose concentration (41.8%), HbA1c (36.7%), impaired glucose tolerance status (31.6%) and fasting glucose concentration (30.8%). CONCLUSIONS: We describe the longitudinal landscape of maternal circulating lipids from preconception to postpartum, and a comprehensive view of trends and magnitude of pregnancy-induced changes in lipidomic profiles. We identified lipid signatures linked with cardiometabolic risk traits with potential implications both in pregnancy and postpartum life. Our findings provide insights into the metabolic adaptations and potential biomarkers of modifiable risk factors in childbearing women that may help in better assessment of cardiometabolic health, and early intervention at the preconception period. TRIAL REGISTRATION: ClinicalTrials.gov, NCT03531658.

Timing of introduction of complementary foods, breastfeeding, and child cardiometabolic risk: a prospective multiethnic Asian cohort study.

BACKGROUND: The timing of introduction of complementary foods and the duration of breastfeeding (BF) have been independently associated with child overweight and obesity; however, their combined influence on body fat partitioning and cardiometabolic risk is unclear. OBJECTIVE: We investigated the associations of the timing of introduction of complementary foods, the duration of BF, and their interaction with child adiposity and cardiometabolic risk markers. METHODS: We analyzed data from 839 children in the prospective Growing Up in Singapore Towards healthy Outcomes (GUSTO) cohort. Mothers reported the age at which infants were first fed complementary foods and BF duration, classified as early (≤4 mo) versus typical (>4 mo) complementary feeding (CF) and short (≤4 mo) versus long (>4 mo) duration of any BF, respectively. We measured adiposity and cardiometabolic risk markers at the age of 6 y and examined their associations with infant feeding patterns using multiple regression, adjusting for sociodemographics, parents' body mass index (BMI), maternal factors, birth weight for gestational age, and infant weight gain. RESULTS: Of 839 children, 18% experienced early CF, whereas 54% experienced short BF. Short (vs. long) BF and early (vs. typical) CF were independently associated with higher z-scores of BMI [ß (95% confidence interval), short BF, 0.18 standard deviation score (SDS) (-0.01, 0.38); early CF, 0.34 SDS (0.11, 0.57)] and sum of skinfolds [short BF, 1.83 mm (0.05, 3.61); early CF, 2.73 mm (0.55, 4.91)]. Children who experienced both early CF and short BF (vs. typical CF-long BF) had synergistically higher diastolic blood pressure [1.41 mmHg (-0.15, 2.97), P-interaction = 0.023] and metabolic syndrome score [0.81 (0.16, 1.47), P-interaction = 0.081]. Early CF-long BF (vs. early CF-short BF) was associated with a lower systolic blood pressure [-3.74 mmHg (-7.01, -0.48)], diastolic blood pressure [-2.29 mmHg (-4.47, -0.11)], and metabolic syndrome score [-0.90 (-1.80, 0.00)]. CONCLUSIONS: A combination of early CF and short BF was associated with elevated child adiposity and cardiometabolic markers. Longer BF duration may protect against cardiometabolic risk associated with early CF. This trial was registered at clinicaltrials.gov as NCT01174875.

The Singapore National Precision Medicine Strategy.

Precision medicine promises to transform healthcare for groups and individuals through early disease detection, refining diagnoses and tailoring treatments. Analysis of large-scale genomic-phenotypic databases is a critical enabler of precision medicine. Although Asia is home to 60% of the world's population, many Asian ancestries are under-represented in existing databases, leading to missed opportunities for new discoveries, particularly for diseases most relevant for these populations. The Singapore National Precision Medicine initiative is a whole-of-government 10-year initiative aiming to generate precision medicine data of up to one million individuals, integrating genomic, lifestyle, health, social and environmental data. Beyond technologies, routine adoption of precision medicine in clinical practice requires social, ethical, legal and regulatory barriers to be addressed. Identifying driver use cases in which precision medicine results in standardized changes to clinical workflows or improvements in population health, coupled with health economic analysis to demonstrate value-based healthcare, is a vital prerequisite for responsible health system adoption.

Profiling Microbial Communities in Idiopathic Granulomatous Mastitis.

Idiopathic granulomatous mastitis (IGM) is a rare and benign inflammatory breast disease with ambiguous aetiology. Contrastingly, lactational mastitis (LM) is commonly diagnosed in breastfeeding women. To investigate IGM aetiology, we profiled the microbial flora of pus and skin in patients with IGM and LM. A total of 26 patients with IGM and 6 patients with LM were included in the study. The 16S rRNA sequencing libraries were constructed from 16S rRNA gene amplified from total DNA extracted from pus and skin swabs in patients with IGM and LM controls. Constructed libraries were multiplexed and paired-end sequenced on HiSeq4000. Metagenomic analysis was conducted using modified microbiome abundance analysis suite customised R-resource for paired pus and skin samples. Microbiome multivariable association analyses were performed using linear models. A total of 21 IGM and 3 LM paired pus and skin samples underwent metagenomic analysis. Bray−Curtis ecological dissimilarity distance showed dissimilarity across four sample types (IGM pus, IGM skin, LM pus, and LM skin; PERMANOVA, p < 0.001). No characteristic dominant genus was observed across the IGM samples. The IGM pus samples were more diverse than corresponding IGM skin samples (Shannon and Simpson index; Wilcoxon paired signed-rank tests, p = 0.022 and p = 0.07). Corynebacterium kroppenstedtii, reportedly associated with IGM in the literature, was higher in IGM pus samples than paired skin samples (Wilcoxon, p = 0.022). Three other species and nineteen genera were statistically significant in paired IGM pus−skin comparison after antibiotic treatment adjustment and multiple comparisons correction. Microbial profiles are unique between patients with IGM and LM. Inter-patient variability and polymicrobial IGM pus samples cannot implicate specific genus or species as an infectious cause for IGM.

Longitudinal characterization of determinants associated with obesogenic growth patterns in early childhood.

BACKGROUND: Longitudinal assessment of the determinants of obesogenic growth trajectories in childhood can suggest appropriate developmental windows for intervention. METHODS: Latent class growth mixture modelling was used to identify body mass index (BMI) z-score trajectories from birth to age 6 years in 994 children from a prospective mother-offspring cohort (Chinese, Indian and Malay ethnicities) based in Singapore. We evaluated the early-life determinants of the trajectories as well as their associations with cardiometabolic risk markers at age 6 years. RESULTS: Five BMI z-score trajectory patterns were identified, three within the healthy weight range, alongside early-acceleration and late-acceleration obesogenic trajectories. The early-acceleration pattern was characterized by elevated fetal abdominal circumference growth velocity, BMI acceleration immediately after birth and crossing of the obesity threshold by age 2 years. The late-acceleration pattern had normal fetal growth and BMI acceleration after infancy, and approached the obesity threshold by age 6 years. Abdominal fat, liver fat, insulin resistance and odds of pre-hypertension/hypertension were elevated in both groups. Indian ethnicity, high pre-pregnancy BMI, high polygenic risk scores for obesity and shorter breastfeeding duration were common risk factors for both groups. Malay ethnicity and low maternal educational attainment were uniquely associated with early BMI acceleration, whereas nulliparity and obesogenic eating behaviours in early childhood were uniquely associated with late BMI acceleration. CONCLUSION: BMI acceleration starting immediately after birth or after infancy were both linked to early cardiometabolic alterations. The determinants of these trajectories may be useful for developing early risk stratification and intervention approaches to counteract metabolic adversities linked to childhood obesity.

Analysis of clinically relevant variants from ancestrally diverse Asian genomes.

Asian populations are under-represented in human genomics research. Here, we characterize clinically significant genetic variation in 9051 genomes representing East Asian, South Asian, and severely under-represented Austronesian-speaking Southeast Asian ancestries. We observe disparate genetic risk burden attributable to ancestry-specific recurrent variants and identify individuals with variants specific to ancestries discordant to their self-reported ethnicity, mostly due to cryptic admixture. About 27% of severe recessive disorder genes with appreciable carrier frequencies in Asians are missed by carrier screening panels, and we estimate 0.5% Asian couples at-risk of having an affected child. Prevalence of medically-actionable variant carriers is 3.4% and a further 1.6% harbour variants with potential for pathogenic classification upon additional clinical/experimental evidence. We profile 23 pharmacogenes with high-confidence gene-drug associations and find 22.4% of Asians at-risk of Centers for Disease Control and Prevention Tier 1 genetic conditions concurrently harbour pharmacogenetic variants with actionable phenotypes, highlighting the benefits of pre-emptive pharmacogenomics. Our findings illuminate the diversity in genetic disease epidemiology and opportunities for precision medicine for a large, diverse Asian population.

Associations of maternal and foetoplacental factors with prehypertension/hypertension in early childhood.

OBJECTIVE: To evaluate whether characterization of maternal and foetoplacental factors beyond birthweight can enable early identification of children at risk of developing prehypertension/hypertension. METHODS: We recruited 693 mother-offspring dyads from the GUSTO prospective mother-offspring cohort. Prehypertension/hypertension at age 6 years was identified using the simplified paediatric threshold of 110/70 mmHg. We evaluated the associations of pregnancy complications (gestational diabetes, excessive/inadequate gestational weight gain, hypertensive disorders of pregnancy), foetal growth deceleration (decline in foetal abdominal circumference at least 0.67 standard deviations between second and third trimesters), high foetoplacental vascular resistance (third trimester umbilical artery systolic-to-diastolic ratio ≥90th centile), preterm birth, small-for-gestational age and neonatal kidney volumes with risk of prehypertension/hypertension at age 6 years, after adjusting for sex, ethnicity, maternal education and prepregnancy BMI. RESULTS: Pregnancy complications, small-for-gestational age, preterm birth, and low neonatal kidney volume were not associated with an increased risk of prehypertension/hypertension at age 6 years. In contrast, foetal growth deceleration was associated with a 72% higher risk [risk ratio (RR) = 1.72, 95% confidence interval (CI) 1.18-2.52]. High foetoplacental vascular resistance was associated with a 58% higher risk (RR = 1.58, 95% CI 0.96-2.62). Having both these characteristics, relative to having neither, was associated with over two-fold higher risk (RR = 2.55, 95% CI 1.26-5.16). Over 85% of the foetuses with either of these characteristics were born appropriate or large for gestational age. CONCLUSION: Foetal growth deceleration and high foetoplacental vascular resistance may be helpful in prioritizing high-risk children for regular blood pressure monitoring and preventive interventions, across the birthweight spectrum.

Fetal sex-specific epigenetic associations with prenatal maternal depressive symptoms.

Prenatal maternal mental health is a global health challenge with poorly defined biological mechanisms. We used maternal blood samples collected during the second trimester from a Singaporean longitudinal birth cohort study to examine the association between inter-individual genome-wide DNA methylation and prenatal maternal depressive symptoms. We found that (1) the maternal methylome was significantly associated with prenatal maternal depressive symptoms only in mothers with a female fetus; and (2) this sex-dependent association was observed in a comparable, UK-based birth cohort study. Qualitative analyses showed fetal sex-specific differences in genomic features of depression-related CpGs and genes mapped from these CpGs in mothers with female fetuses implicated in a depression-associated WNT/ß-catenin signaling pathway. These same genes also showed enriched expression in brain regions linked to major depressive disorder. We also found similar female-specific associations with fetal-facing placenta methylome. Our fetal sex-specific findings provide evidence for maternal-fetal interactions as a mechanism for intergenerational transmission.

Machine Learning-Derived Prenatal Predictive Risk Model to Guide Intervention and Prevent the Progression of Gestational Diabetes Mellitus to Type 2 Diabetes: Prediction Model Development Study.

BACKGROUND: The increasing prevalence of gestational diabetes mellitus (GDM) is concerning as women with GDM are at high risk of type 2 diabetes (T2D) later in life. The magnitude of this risk highlights the importance of early intervention to prevent the progression of GDM to T2D. Rates of postpartum screening are suboptimal, often as low as 13% in Asian countries. The lack of preventive care through structured postpartum screening in several health care systems and low public awareness are key barriers to postpartum diabetes screening. OBJECTIVE: In this study, we developed a machine learning model for early prediction of postpartum T2D following routine antenatal GDM screening. The early prediction of postpartum T2D during prenatal care would enable the implementation of effective strategies for diabetes prevention interventions. To our best knowledge, this is the first study that uses machine learning for postpartum T2D risk assessment in antenatal populations of Asian origin. METHODS: Prospective multiethnic data (Chinese, Malay, and Indian ethnicities) from 561 pregnancies in Singapore's most deeply phenotyped mother-offspring cohort study-Growing Up in Singapore Towards healthy Outcomes-were used for predictive modeling. The feature variables included were demographics, medical or obstetric history, physical measures, lifestyle information, and GDM diagnosis. Shapley values were combined with CatBoost tree ensembles to perform feature selection. Our game theoretical approach for predictive analytics enables population subtyping and pattern discovery for data-driven precision care. The predictive models were trained using 4 machine learning algorithms: logistic regression, support vector machine, CatBoost gradient boosting, and artificial neural network. We used 5-fold stratified cross-validation to preserve the same proportion of T2D cases in each fold. Grid search pipelines were built to evaluate the best performing hyperparameters. RESULTS: A high performance prediction model for postpartum T2D comprising of 2 midgestation features-midpregnancy BMI after gestational weight gain and diagnosis of GDM-was developed (BMI_GDM CatBoost model: AUC=0.86, 95% CI 0.72-0.99). Prepregnancy BMI alone was inadequate in predicting postpartum T2D risk (ppBMI CatBoost model: AUC=0.62, 95% CI 0.39-0.86). A 2-hour postprandial glucose test (BMI_2hour CatBoost model: AUC=0.86, 95% CI 0.76-0.96) showed a stronger postpartum T2D risk prediction effect compared to fasting glucose test (BMI_Fasting CatBoost model: AUC=0.76, 95% CI 0.61-0.91). The BMI_GDM model was also robust when using a modified 2-point International Association of the Diabetes and Pregnancy Study Groups (IADPSG) 2018 criteria for GDM diagnosis (BMI_GDM2 CatBoost model: AUC=0.84, 95% CI 0.72-0.97). Total gestational weight gain was inversely associated with postpartum T2D outcome, independent of prepregnancy BMI and diagnosis of GDM (P=.02; OR 0.88, 95% CI 0.79-0.98). CONCLUSIONS: Midgestation weight gain effects, combined with the metabolic derangements underlying GDM during pregnancy, signal future T2D risk in Singaporean women. Further studies will be required to examine the influence of metabolic adaptations in pregnancy on postpartum maternal metabolic health outcomes. The state-of-the-art machine learning model can be leveraged as a rapid risk stratification tool during prenatal care. TRIAL REGISTRATION: ClinicalTrials.gov NCT01174875; https://clinicaltrials.gov/ct2/show/NCT01174875.

Allergic sensitization trajectories to age 8 years in the Singapore GUSTO cohort.

Background: Allergic sensitization is linked to allergy development, with early sensitization often associated with worse outcomes. We aimed to identify if distinct allergic sensitization trajectories existed within a diverse and multi-ethnic Asian cohort. Methods: We administered modified ISAAC questionnaires in the first 8 years and conducted skin prick testing at ages 18 months, 3, 5 and 8 years in the Growing Up in Singapore Towards healthy Outcomes (GUSTO) cohort. We used latent class analysis to derive allergic sensitization trajectories, and adjusted odds ratios (AOR) to evaluate predictive risk factors and associations with allergic comorbidities. Results: Among 997 children, three trajectories were identified: early food and mite sensitization (16.2%), late mite sensitization (24.2%) and no/low sensitization (59.6%). Early food and mite sensitization was associated with early eczema by 6 months [AOR (95%CI) 4.67 (1.78-12.28)], increased risk of wheeze by 3-8 years (ARR 1.72-1.99) and eczema in the first 8 years of life (ARR 1.87-2.41). Late mite sensitization was associated with female sex [AOR 0.58 (0.35-0.96)], cesarean section [AOR 0.54 (0.30-0.98)], early eczema by 6 months [AOR 3.40 (1.38-8.42)], and increased risk of eczema by 18 months [ARR 1.47 (1.03-2.08)] and 8 years [ARR 1.35 (1.05-1.73)]. Conclusion: Early onset of eczema and early allergic sensitization were strongly associated. Early sensitization, especially to house dust mites, was associated with increased risks of developing wheeze and eczema, pointing to the importance of developing preventive perinatal interventions and effective therapeutics for sensitized toddlers.

Population-based plasma lipidomics reveals developmental changes in metabolism and signatures of obesity risk: a mother-offspring cohort study.

BACKGROUND: Lipids play a vital role in health and disease, but changes to their circulating levels and the link with obesity remain poorly characterized in expecting mothers and their offspring in early childhood. METHODS: LC-MS/MS-based quantitation of 480 lipid species was performed on 2491 plasma samples collected at 4 time points in the mother-offspring Asian cohort GUSTO (Growing Up in Singapore Towards healthy Outcomes). These 4 time points constituted samples collected from mothers at 26-28 weeks of gestation (n=752) and 4-5 years postpartum (n=650), and their offspring at birth (n=751) and 6 years of age (n=338). Linear regression models were used to identify the pregnancy and developmental age-specific variations in the plasma lipidomic profiles, and their association with obesity risk. An independent birth cohort (n=1935), the Barwon Infant Study (BIS), comprising mother-offspring dyads of Caucasian origin was used for validation. RESULTS: Levels of 36% of the profiled lipids were significantly higher (absolute fold change > 1.5 and Padj < 0.05) in antenatal maternal circulation as compared to the postnatal phase, with phosphatidylethanolamine levels changing the most. Compared to antenatal maternal lipids, cord blood showed lower concentrations of most lipid species (79%) except lysophospholipids and acylcarnitines. Changes in lipid concentrations from birth to 6 years of age were much higher in magnitude (log2FC=-2.10 to 6.25) than the changes observed between a 6-year-old child and an adult (postnatal mother) (log2FC=-0.68 to 1.18). Associations of cord blood lipidomic profiles with birth weight displayed distinct trends compared to the lipidomic profiles associated with child BMI at 6 years. Comparison of the results between the child and adult BMI identified similarities in association with consistent trends (R2=0.75). However, large number of lipids were associated with BMI in adults (67%) compared to the children (29%). Pre-pregnancy BMI was specifically associated with decrease in the levels of phospholipids, sphingomyelin, and several triacylglycerol species in pregnancy. CONCLUSIONS: In summary, our study provides a detailed landscape of the in utero lipid environment provided by the gestating mother to the growing fetus, and the magnitude of changes in plasma lipidomic profiles from birth to early childhood. We identified the effects of adiposity on the circulating lipid levels in pregnant and non-pregnant women as well as offspring at birth and at 6 years of age. Additionally, the pediatric vs maternal overlap of the circulating lipid phenotype of obesity risk provides intergenerational insights and early opportunities to track and intervene the onset of metabolic adversities. CLINICAL TRIAL REGISTRATION: This birth cohort is a prospective observational study, which was registered on 1 July 2010 under the identifier NCT01174875 .

Automated Machine Learning (AutoML)-Derived Preconception Predictive Risk Model to Guide Early Intervention for Gestational Diabetes Mellitus.

The increasing prevalence of gestational diabetes mellitus (GDM) is contributing to the rising global burden of type 2 diabetes (T2D) and intergenerational cycle of chronic metabolic disorders. Primary lifestyle interventions to manage GDM, including second trimester dietary and exercise guidance, have met with limited success due to late implementation, poor adherence and generic guidelines. In this study, we aimed to build a preconception-based GDM predictor to enable early intervention. We also assessed the associations of top predictors with GDM and adverse birth outcomes. Our evolutionary algorithm-based automated machine learning (AutoML) model was implemented with data from 222 Asian multi-ethnic women in a preconception cohort study, Singapore Preconception Study of Long-Term Maternal and Child Outcomes (S-PRESTO). A stacked ensemble model with a gradient boosting classifier and linear support vector machine classifier (stochastic gradient descent training) was derived using genetic programming, achieving an excellent AUC of 0.93 based on four features (glycated hemoglobin A1c (HbA1c), mean arterial blood pressure, fasting insulin, triglycerides/HDL ratio). The results of multivariate logistic regression model showed that each 1 mmol/mol increase in preconception HbA1c was positively associated with increased risks of GDM (p = 0.001, odds ratio (95% CI) 1.34 (1.13-1.60)) and preterm birth (p = 0.011, odds ratio 1.63 (1.12-2.38)). Optimal control of preconception HbA1c may aid in preventing GDM and reducing the incidence of preterm birth. Our trained predictor has been deployed as a web application that can be easily employed in GDM intervention programs, prior to conception.