Supervised Modelling of Longitudinal Human Milk and Infant Gut Microbiome Reveal Maternal Pre-Pregnancy BMI and Early Life Growth Interactions

Supervised Modelling of Longitudinal Human Milk and Infant Gut Microbiome Reveal Maternal Pre-Pregnancy BMI and Early Life Growth Interactions | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Supervised Modelling of Longitudinal Human Milk and Infant Gut Microbiome Reveal Maternal Pre-Pregnancy BMI and Early Life Growth Interactions Rasmus Jakobsen, Geert Roelof van der Ploeg, Ulrik Sundekilde, and 7 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6244750/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Maternal obesity is a key risk factor for excessive foetal growth and childhood obesity, yet its influence on human milk (HM) composition and the infant gut microbiome development remains unclear. This study examined 169 mother-infant dyads analyzing 570 HM metabolome, 495 HM microbiome, and 348 infant faecal microbiome samples over three months of exclusive breastfeeding, alongside infant anthropometric data through three years postpartum. While BMI was not directly correlated with infant growth (weight-for-length/height z-score), N-way Partial Least Squares modelling revealed microbial and metabolite signatures linked to maternal ppBMI and infant growth. High maternal ppBMI and infant growth were associated with altered HM oligosaccharides and enrichment of Bifidobacterium spp. in the infant gut. In contrast, elevated HM simple sugars, amino acid derivatives, and gut Klebsiella and Escherichia spp. relative abundance linked to slower growth. These findings highlight maternal-infant nutritional dynamics, informing targeted strategies to support infant growth. Biological sciences/Microbiology/Microbial communities/Microbiome Biological sciences/Systems biology/Time series Biological sciences/Computational biology and bioinformatics/Data integration Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction The first months of life represent a critical window for growth and development, during which excessive weight gain is associated with later obesity risk 1 . Maternal obesity is a significant risk factor for excessive fetal growth 2 , infant overweight, and childhood obesity 3 , 4 . Although the mechanisms underlying the intergenerational transmission of obesity are not fully understood, early-life nutrition is thought to play a critical role. Among its numerous health benefits, breastfeeding is linked with a decreased risk of childhood obesity and slower growth rates compared to formula feeding 5 . However, maternal obesity is associated with reduced initiation and shorter durations of breastfeeding, potentially diminishing these protective effects 1 , 6 – 8 . Human milk (HM) provides multiple bioactive compounds that influence infant growth and the establishment of the infant gut microbiome 3 . During infancy, the infant gut microbiome undergoes rapid compositional changes, with its maturation by 1–2 years strongly linked to long-term health outcomes 9 – 11 . Emerging evidence suggests that maternal pre-pregnancy BMI (ppBMI) can affect HM composition and the initial microbial colonization of the infant gut, potentially influencing both infant growth trajectories and infant gut microbiome establishment 3 . For example, high maternal ppBMI has been associated with increased HM content of monosaccharides and sugar alcohols, and decreased HM content of oligosaccharides 12 , 13 . However, studies investigating the relationship between HM composition, infant growth and maternal ppBMI are scarce and largely limited to cross-sectional analyses 14 – 16 . Several mechanisms may explain how HM composition influences growth and reduces obesity risk relative to formula feeding 17 (Fig. 1 A). Bioactive HM components, including human milk oligosaccharides (HMOs) and lipids, shape the infant gut microbiome and promote a profile associated with metabolic health 18 . For example, HMOs are selectively utilized by Bifidobacterium spp., which in turn have been associated with lower obesity risk 19 , 20 . Additionally, microbial transfer from HM to the infant gut is also hypothesized to influence early microbiota assembly 21 . These mechanisms may act synergistically with the direct provision of HM nutrients to modulate growth patterns. Maternal ppBMI is also associated with altered HM microbiome and metabolome composition 12 , 13 , 22 , which might also be a key pathway contributing to the intergenerational transmission of obesity. The disentanglement of ppBMI associated changes in HM and their interactions with the infant gut microbiome is needed to uncover the mechanisms of inter-generational obesity. To investigate these, we recruited and longitudinally sampled a cohort of 169 mother-infant dyads at 3–4 time points across the first three months of exclusive breastfeeding, followed by growth anthropometrics at 6 months through 3 years 23 . Infant anthropometrics were recorded from birth through three years postpartum. An exploratory analysis of these datasets using principal component analysis (PCA) and principal coordinates analysis (PCoA) reported that maternal pre-pregnancy BMI and infant growth significantly correlated with HM metabolite profile and infant gut microbial diversity and composition with significant associations between HM oligosaccharide clusters and infant gut bifidobacteria 24 . However, PCA and PCoA only capture the dominant source of variation and do not consider longitudinal dynamics. To address these limitations, we applied N-way Partial Least Squares (NPLS) modelling to investigate microbial and metabolite signatures associated with maternal ppBMI and infant growth. NPLS is well-suited for this type of analysis since it enables simultaneous analysis of longitudinal microbiome and metabolomics data to uncover biologically relevant associations. By integrating temporal dynamics across the HM metabolome, microbiome and infant gut microbiome, NPLS captures relationships between features within these datasets and maternal and infant phenotypes. By elucidating these relationships, our findings may inform targeted nutritional interventions, such as supplementing key components to breast milk or using tailored infant formula, to mitigate obesity risk and promote optimal infant health. Results Microbiome and Metabolome Profiles Differ Between Maternal ppBMI Groups and Shift Over Time This study included samples from 169 healthy Danish mother-infant dyads from the MAINHEALTH cohort 23 , consisting of human milk (HM) samples collected at days 3, 30, 60 and 90 postpartum and infant faecal samples collected at days 30, 60 and 90 postpartum, along with extensive anthropometric and clinical data ( Figure 1B ). The offspring were exclusively breastfed throughout this period. The average pre-pregnancy BMI (ppBMI) was 27±5.4 kg/m 2 , and the mean gestational age at delivery was 283±2.5 days (mean ± sd, Table 1 ). Infants (86 male, 80 female) had average birth weights of 3704±450 g. The maternal cohort was stratified into three ppBMI groups: normal weight (NW; BMI < 25 kg/m 2 ), overweight (OW; BMI 25-30 kg/m 2 ) and obesity (OB; BMI ≥ 30 kg/m 2 ) ( Supplementary Figure 1 ). Infant anthropometrics were recorded at birth and 1, 2, 3, 6, 12, 24 and 36 months after birth ( Supplementary Figure 2 ). HM metabolites were characterised using 1 H nuclear magnetic resonance spectroscopy (NMR) and HM and infant faecal bacterial communities by V1-V9 16S rRNA gene Nanopore long-read amplicon sequencing. After quality control, 570 HM metabolite samples, 495 HM microbiome samples, and 348 infant faecal microbiome samples were included for analysis. Examining the HM and infant faecal microbiomes, both showed a difference in composition between maternal ppBMI groups and over time, consistent with previous reports 25–27 . At the genus level, infant faecal microbiomes from overweight-or-obese mothers were depleted of Klebsiella (p=0.008, BH-corrected Kruskall-Wallis test) relative to infants normal ppBMI mothers ( Figure 1D ). HM microbiomes from overweight-or-obese mothers were enriched in Gemella (p=0.008) but depleted in Escherichia-Shigella ( p=0.01 ), Lactobacillus (p<0.001), Bifidobacterium (p=0.01), Acinetobacter (p=0.002) relative to infants of normal ppBMI mothers. Examining longitudinal dynamics (day 30 to 90), the infant faecal microbiomes saw an increase in the abundances of Lacticaseibacillus (p=0.03, BH-corrected Friedman test), and a decrease in Streptococcus (p<0.001), while the HM microbiome samples showed a decrease in Staphylococcus (p=0.07), Gemella (p=0.04) and an increase in Lactobacillus (p=0.05) ( Figure 1E ). Similarly, multiple HM metabolites also differed significantly between maternal ppBMIs and over time. HM samples from normal weight mothers were enriched in myo-inositol (p=0.004 BH-corrected Kruskall-Wallis test), glucose (p=0.01), and lactose (p=0.05), while five unclassified human milk oligosaccharides (HMOs) were increased in overweight-or-obese mothers (p<0.02) ( Figure 1F, Supplementary table 1 ). Temporal changes in HM composition included increases in lactose (p<0.001, BH-corrected Friedman test), myo-inositol (p<0.001), 3-fucosyllactose (3-FL) (p<0.001), glucose (p<0.001) and glutamate (p<0.001), and decreases in 2’-fucosyllactose (2’-FL) (p<0.001) and two unclassified HMOs (p<0.001, p<0.001) ( Figure 1G, Supplementary table 1 ). Table 1. Maternal and infant characteristics. Continuous data are presented as means ± standard deviation. Categorical data are presented as numbers included in each category. Statistical tests: ANOVA for numeric variables and Chi-square test for categorical. BMI: body mass index, BMIz: BMI-for-age, WHZ: weight-for-length/height z-score, C-section: Caesarean section, n: number of subjects. Variable n Mean n Mean n Mean n Mean Test Days after birth 3 30 60 90 BMI (kg/m²) 126 27±5.6 145 27±5.2 150 27±5.3 148 27±5.3 F=0.239 BMI group 126 145 150 148 χ²=1.60 Normal 50 40% 57 39% 66 44% 65 44% Overweight 41 33% 50 34% 47 31% 49 33% Obese 35 28% 38 26% 37 25% 34 23% Maternal age (years) 126 31±4.1 145 31±4 150 31±4 148 31±4.1 F=0.124 Gestational diabetes 125 144 150 147 χ²=0.374 No 118 94% 138 96% 142 95% 139 95% Yes 7 6% 6 4% 8 5% 8 5% Siblings 126 145 150 148 χ²=0.614 No 52 41% 63 43% 66 44% 68 46% Yes 74 59% 82 57% 84 56% 80 54% Birth mode 126 145 150 148 χ²=2.20 C-section 6 5% 13 9% 13 9% 13 9% Vaginal 120 95% 132 91% 137 91% 135 91% Infant sex 124 143 149 146 χ²=0.401 Female 67 54% 72 50% 79 53% 77 53% Male 57 46% 71 50% 70 47% 69 47% Gestational age after birth (Days) 123 282±7.9 142 283±7.7 147 283±7.2 144 283±7.5 F=0.214 Birth weight (kg) 122 3.7±0.48 140 3.7±0.45 146 3.7±0.46 143 3.7±0.46 F=0.081 Secretor status 126 145 150 148 χ²=0.62 Non-secretor 34 27% 34 23% 35 23% 36 24% Secretor 92 73% 111 77% 115 77% 112 76% Lewis status 126 145 150 148 χ²=0.847 Lewis negative 6 5% 6 4% 6 4% 4 3% Lewis positive 120 95% 139 96% 144 96% 144 97% Antibiotics mother 122 145 149 146 χ²=81.1*** No 121 99% 94 65% 134 90% 136 93% Yes 1 1% 51 35% 15 10% 10 7% Antibiotics infant 0 142 147 145 χ²=2.82 No 0 128 90% 134 91% 138 95% Yes 0 14 10% 13 9% 7 5% Breast infection 125 145 149 146 χ²=25.1*** No 87 70% 103 71% 130 87% 128 88% Yes 38 30% 42 29% 19 13% 18 12% Breastfeeding issues 125 145 149 146 χ²=16.3** No 89 71% 99 68% 118 79% 123 84% Yes 35 28% 45 31% 30 20% 20 14% Δ BMI 3 months 115 -0.47±1.2 133 -0.57±1.2 142 -0.61±1.3 146 -0.61±1.2 F=0.327 Δ HAZ 99 0.62±1.3 113 0.58±1.3 123 0.61±1.2 124 0.66±1.2 F=0.087 Δ WAZ 100 0.69±1.1 114 0.61±1.2 124 0.6±1.1 125 0.61±1.1 F=0.132 Δ WHZ 106 0.66±1.4 121 0.65±1.4 132 0.61±1.4 130 0.57±1.4 F=0.115 Maternal ppBMI Does Not Correlate with Infant Growth The relationship between maternal ppBMI and infant anthropometrics, including weight-for-length z-scores (WHZ) at 6 months and BMI-for-age z-scores (BMIz) at one, two and three years, was assessed ( Figure 1C, Supplementary Figure 2 ). Infant 6-month WHZ was chosen to measure the accumulated effect of exclusive breastfeeding on infant growth trajectories. To represent early childhood weight outcomes, we included 1-year WHZ and BMI-for-age z-scores at 2 and 3 years. Surprisingly, we found no significant correlations between maternal ppBMI and infant growth outcomes (p>0.05; Pearson correlation test; Table 2 ). In contrast, all infant growth metrics from 6 months to 3 years showed significant correlations with each other (p≤0.05). While all correlations between growth metrics were significant, they consistently weakened as the time interval between measurements increased. Table 2. Pearson correlation test results between pre-pregnancy BMI and infant growth outcomes. Values correspond to the correlation value, while stars indicate significance (*: p≤0.05, **: p≤0.01, ***: p≤0.001). ppBMI: pre-pregnancy BMI, BMIz: BMI-for-age, WHZ: weight-for-length/height z-score. -- maternal ppBMI Infant 6-month WHZ Infant 1-year WHZ Infant 2-year BMIz score Infant 3-year BMIz score maternal ppBMI 1 0.07 0.08 -0.03 0.11 Infant 6-month WHZ 1 0.63*** 0.42*** 0.41*** Infant 1-year WHZ 1 0.59*** 0.57*** Infant 2-year BMIz score 1 0.57*** Infant 3-year BMIz score 1 NPLS Models Describe Longitudinal Differences in Microbiome and Metabolome Profiles While maternal ppBMI did not appear to directly influence infant growth ( Table 2 ), variations in microbiome and metabolome profiles highlighted potential indirect effects mediated through HM composition and infant gut microbiome. To address this, we applied N-way Partial Least Squares (NPLS) 28 to focus the analysis of the longitudinal infant faecal microbiome, HM microbiome and HM metabolome on variation that is associated with either maternal pre-pregnancy BMI or infant growth ( Figure 2 ). Regressing the HM metabolome, HM microbiome and infant faecal microbiome data onto maternal ppBMI and infant growth allows for a direct examination of their specific contributions, as exploratory methods might not capture variation related to the complex hypothesized interactions in such datasets. NPLS was chosen for its ability to uncover mediating effects linking maternal ppBMI, HM composition, the infant gut microbiome and to infant growth by incorporating a dependent variable into the analysis. This was done by converting each dataset into a tensor of bacterial or metabolite profiles across the time points 29 and subsequently regressing them onto either maternal ppBMI or infant 6-month WHZ, resulting in six NPLS models total ( Figure 2A-B, Supplementary Figures 3-4 ). A one-component model was selected for each NPLS model, as cross-validation metrics indicated that selecting additional components did not sufficiently improve explanatory power or predictive performance ( Supplementary Figures 5-6 ). The NPLS models explained 2-6% of variation in their respective independent datasets (HM microbiome, HM metabolome, or infant faecal microbiome) and 9-31% of variation in their dependent variable (ppBMI or 6-month WHZ) ( Table 3, Supplementary Figure 5 ). The predictive performance of each NPLS model was assessed by correlating predicted and measured values of the dependent variable. This approach revealed strong predictive performance in the NPLS models of the infant faecal microbiome, HM microbiome and HM metabolome regressed onto maternal ppBMI (R=0.35, R=0.34, R=0.56 respectively; p=5.7e-6, p=8.4e-6, p=9.2e-15, respectively; Pearson correlation test; Table 3 ). Similarly, we found strong predictive performance in the NPLS models of the infant faecal microbiome, HM microbiome and HM metabolome regressed onto infant 6-month WHZ (R=0.32, R=0.31, R=0.37, respectively; p=1.7e-4, p=2.8e-4, p=1.2e-5, respectively). To explore the influence of maternal ppBMI, we examined the associations between microbial and metabolite profiles across the HM metabolome, HM microbiome and infant gut microbiome ( Figure 2C-E, Supplementary Figures 7-9 ). The NPLS models describing the HM metabolomics revealed that compounds associated with energy metabolism such as fumarate and lactate alongside several unclassified HMOs were associated with high ppBMI. In contrast, low ppBMI was linked to amino acid derivatives such as 2-aminobutyrate, glutamine and taurine along with simple sugars such as glucose, myo-inositol and acetone, suggesting differences in HM metabolic energy allocation ( Figure 2C, Supplementary Figure 9 ). Differences were also observed in the NPLS models describing the HM microbiome, where Staphylococcus spp. , Gemella spp., and Streptococcus spp. were associated with high ppBMI while several unidentified Bacteroides spp., Clostridium spp . along with Staphylococcus aureus , Bifidobacterium longum and Bifidobacterium pseudocatenulatum were associated with low ppBMI ( Figure 2D, Supplementary Figure 8 ). For the infant gut microbiome, the NPLS model revealed that several Bifidobacterium breve and Bifidobacterium bifidum zOTUs (zero-radius operational taxonomic units) were associated with high maternal ppBMI alongside Klebsiella spp., Flavonifractor spp., Escherichia spp. and a diverse set of Firmicutes members. In contrast, low maternal ppBMI was linked to Libanicoccus spp . , Actinomyces spp., Rothia mucilaginosa and Bifidobacterium adolescentis , possibly reflecting downstream effects of maternal health on infant gut colonization ( Figure 2E, Supplementary Figure 7 ). We also examined the association between microbial and metabolic profiles and infant 6-month WHZ ( Figure 2F-H, Supplementary Figures 10-12 ). In the NPLS model describing the HM metabolomics, multiple HMOs alongside succinate, lysine and hydroxybutyrate were associated with high infant 6-month WHZ, suggesting the importance of these compounds for infant growth optimization. In contrast, low infant 6-month WHZ was associated with the amino acids phenylalanine and methionine, alongside simple sugars such as glucose, lactose and galactose together with butyrate and sn -glycero-3-phosphocholine, indicating a different nutrient subset associated with slower growth trajectories ( Figure 2F, Supplementary Figure 12 ). Growth-related bacterial associations were also observed in the NPLS model describing the HM microbiome, where high infant 6-month WHZ was associated with Rothia spp. along with Streptococcus spp., Enterococcus spp. and Staphylococcus spp., while low infant 6-month WHZ was associated with Acinetobacter spp., Pseudomonas spp. and Staphylococcus spp., Lactobacillus spp . , and several Bifidobacterium spp., suggesting microbiome-mediated growth-modulatory effects ( Figure 2G, Supplementary Figure 11 ). NPLS model describing the infant faecal microbiome showed some overlap with the ppBMI-regressed model, with high infant 6-month WHZ associating with Actinobacteriota members, including Actinomyces spp., and multiple Bifidobacterium spp. along with a diverse set of Firmicutes zOTUs. In contrast, multiple Erysipelatoclostridium spp., Clostridium spp., Bacteroides spp., Klebsiella spp. and Escherichia-Shigella spp. were associated with low infant 6-month WHZ, suggesting a dysbiotic microbial profile potentially linked to slower infant growth ( Figure 2H, Supplementary Figure 10 ). To investigate the potential long-term effects of weight-related microbial taxa and metabolites on infant growth, we analyzed the correlations between subject loadings derived from NPLS models for maternal ppBMI and 6-month WHZ with infant growth outcomes measured at 1 to 3 years. Across independent datasets, no significant correlations were observed between the infant growth outcomes and subject loadings in NPLS models that were regressed onto maternal ppBMI (p>0.05, BH-corrected Pearson correlation test, Supplementary Figure 13 , Supplementary Table 2 ). These findings are consistent with the observed lack of direct correlation between maternal ppBMI and infant growth. However, the NPLS models did show good predictive performance by capturing time-resolved variation in the HM microbiome, HM metabolome and infant gut microbiome associated with either maternal ppBMI or infant 6-month WHZ ( Table 3 ). This demonstrates the utility of NPLS modeling in identifying distinct patterns in microbiome and metabolome profiles over time, even when a direct correlation between maternal ppBMI and infant growth outcomes is not apparent. Table 3 . Pearson correlation test results of the NPLS model prediction with the used dependent variable. R = correlation value, RMSE = root mean squared error. ppBMI: pre-pregnancy BMI, WHZ: weight-for-length/height z-score. σ: the standard deviation of the dependent variable. varExpX: variation explained by the model in the independent dataset. varExpY: variation explained by the model in the dependent variable. See Supplementary Figure 14 for the scatter plots underlying these correlation values. Independent dataset Dependent variable R p-value RMSE σ varExpX (%) varExpY(%) Infant Faecal Bacteria Maternal ppBMI 0.35 5.7e-6 4.90 5.26 5.77 12.30 Infant 6-month WHZ 0.32 1.7e-4 1.06 1.03 2.61 10.34 Milk Bacteria Maternal ppBMI 0.34 8.4e-6 4.87 5.26 5.31 11.22 Infant 6-month WHZ 0.31 2.8e-4 1.08 1.03 1.89 9.43 Milk Metabolites Maternal ppBMI 0.56 9.2e-15 4.32 5.26 3.06 31.04 Infant 6-month WHZ 0.37 1.2e-5 1.05 1.03 2.97 13.40 NPLS Models Describe Microbial Subcommunities with Similar Dynamics Per Maternal ppBMI Group To evaluate whether the zOTUs strongly associated with maternal ppBMI in the NPLS models constituted significant differences in overall microbial composition , we examined their combined relative abundances in the HM and infant faecal microbiomes. Bacterial zOTUs strongly associated with either low maternal ppBMI or high maternal ppBMI were clustered based on their fitted responses ( Figure 2D-E , Supplementary Table 3 ). The summed relative abundances of these clusters were then compared between maternal ppBMI groups across all time points ( Figure 3 ). In the NPLS model describing the infant faecal microbiome ( Figure 3A ), the cluster associated with high maternal ppBMI (containing Bifidobacterium bifidum, Bifidobacterium breve, Clostridium perfringens, Escherichia coli, Klebsiella variicola, and Staphylococcus aureus ) was enriched in samples from infants with overweight-or-obese mothers compared to samples from infants with normal weight mothers at all time points (p=0.016, p<1e-3, p=0.0015, respectively; BH-corrected permutation test of mean difference, n=999; Supplementary Table 4 ). The cluster associated with low maternal ppBMI (containing Bacteroides fragilis, Bifidobacterium adolescentis, Citrobacter freundii, Enterococcus faecalis, Lactobacillus paracasei, Parabacteroides distasonis, and Rothia mucilaginosa ) was enriched in samples from infants with normal weight mothers compared to samples from infants with overweight-or-obese mothers at days 60 and 90 (p=0.0015, p<1e-3, respectively). Similarly, in the NPLS model describing the HM microbiome ( Figure 3B ), the cluster associated with high maternal ppBMI (containing Rothia mucilaginosa, Staphylococcus epidermidis, and Streptococcus oralis ) was enriched in samples from overweight-or-obese mothers compared to samples from normal weight mothers at all time points (p<1e-3, p=0.0053, p<1e-3, p<1e-3, respectively; Supplementary Table 4 ). Conversely, the cluster associated with low maternal ppBMI (containing Bifidobacterium longum, Bifidobacterium pseudocatenulatum, Cutibacterium acnes, Lactobacillus gasseri, Ligilactobacillus murinus, Staphylococcus aureus, and Stenotrophomonas maltophilia ) was enriched in samples from normal weight mothers compared to those from overweight-or-obese mothers at days 3 and 90 (p=0.0032, p=0.0032, respectively). Comparisons of Microbiome Loadings Reveal Shared Bacterial Signatures Between Human Milk and Infant Gut Although maternal ppBMI and infant 6-month WHZ were not correlated, many bacteria and metabolites were associated with both metrics in the NPLS models ( Figure 2 ). To explore these relationships further, we performed pairwise comparisons of the loadings for each bacterial zOTU and metabolite between NPLS models regressed onto maternal ppBMI and infant 6-month WHZ ( Figure 4A ). The NPLS model describing the HM metabolome regressed onto maternal ppBMI was compared to the model of the same data regressed onto infant 6-month WHZ. A large proportion of metabolites had similar associations to both dependent variables ( Figure 4B ). HMOs were consistently associated with high maternal ppBMI and high infant 6-month WHZ along with energy metabolism-related compounds and fucose as the sole sugar. Conversely, compounds associated with low maternal ppBMI and low infant 6-month WHZ included sugars, fatty acids, amino acids, and their derivatives. Several metabolites were uniquely associated with one metric but showed neutral or modest associations with the other. Most notably, high infant 6-month WHZ was associated with 2-hydroxybutyrate, fucose, lysine, fucose, LNFPH I, and LNDFH I, while phenylalanine, butyrate, galactose and lactose were associated with with low infant 6-month WHZ. None of these metabolites were strongly associated with ppBMI. Conversely, 2-amminobutyrate, ethanolamine, taurine, myo-inositol, and threonine were associated with low ppBMI, but had neutral associations with infant 6-month WHZ. A few compounds displayed inconsistent associations; for example, caffeine was associated with high ppBMI but also with low infant 6-month WHZ while acetone showed opposite associations. No metabolites were uniquely associated with high ppBMI. In the NPLS models describing the HM microbiome, most zOTUs had similar associations for both maternal ppBMI and infant 6-month WHZ ( Figure 4C ). Streptococcus, Veillonella, and Gemella zOTUs were associated with high maternal ppBMI and high infant 6-month WHZ, while Acinetobacter , Staphylococcus and Pseudomonas zOTUs were associated with both low maternal ppBMI and low infant 6-month WHZ. Notably, some zOTUs had opposite associations, with a Clostridium sp. and Lactobacillus spp. being associated with low ppBMI but high infant 6-month WHZ, while Bifidobacterium animalis , Bacteroides sp., and Acidovorax sp. were associated with high ppBMI but low infant 6-month WHZ. Several zOTUs had distinct associations, with Rothia sp. and Echerichia-Shigella sp. associated to high ppBMI but were neutral in their association to infant 6-month WHZ. Conversely, Enterococcus faecalis , Staphylococcus aureus and Streptococcus sp. were associated to high infant 6-month WHZ, but had neutral loadings in the ppBMI model. For the infant faecal microbiome, our approach revealed multiple Bifidobacterium spp . and an Enterococcus sp. that were associated with high maternal ppBMI as well as a high infant 6-month WHZ ( Figure 4D ). Several Bifidobacterium adolescentis zOTUs contrasted by associating with low maternal ppBMI and a low infant 6-month WHZ, along with Bacteriodes spp., Klebiella spp., and Libanicoccus sp. While most bifidobacteria, other than B. adoliscentis , were associated with high WHZ, several had neutral-to-moderate associations to ppBMI. zOTUs associated with both low maternal ppBMI and low infant 6-month WHZ included Bacteroides fragilis, and Klebsiella spp ., while no zOTUs had strong associations to low ppBMI and high 6-month WHZ. Notable exclusive associations for infant 6-month WHZ were Actinomyces associated to high WHZ, and Klebsiella variicola , Erysipelatoclostridium spp . , and Parabacteroides to low WHZ. Low ppBMI had distinct associations for Streptococcus spp. and high ppBMI for Klebsiella variicola . To identify the impact of microbial transfer from HM to infant gut, bacterial associations in the NPLS models, for the HM and infant faecal microbiomes regressed onto maternal ppBMI were compared ( Supplementary Figure 15 ). B. longum and B. pseudocatenulatum zOTUs were associated with low ppBMI in both models, while E. coli was associated to high ppBMI only in the NPLS model describing the infant faecal microbiome. Veillonella spp. and some Streptococcus spp. were moderately associated with high ppBMI in both compartments. One Streptococcus sp. and Rothia mucilaginosa were associated with high ppBMI in the NPLS model describing the HM microbiome, but were associated with low ppBMI in the model describing the infant faecal microbiome, while a Blautia sp. showed the opposite. Comparing 6-month infant WHZ-regressed NPLS models of the infant faecal microbiome and the HM microbiome ( Supplementary figure 15 ), L. gasseri was associated with high infant 6-month WHZ in both compartments. S. aureus and E. coli were linked to low infant 6-month WHZ, while S. epidermidis showed the opposite. A Veillonella sp. was associated with high infant 6-month WHZ only in the HM microbiome. In summary this shows that while the metabolite and microbiome associations for ppBMI and 6-month WHZ are shared, many bacteria have strong associations to only one of the two, or have contrasting associations. Discussion This study reveals that maternal pre-pregnancy BMI (ppBMI) influences human milk (HM) composition and infant gut microbiome development, although ppBMI does not directly correlate with infant growth metrics. Using N-way Partial Least Squares (NPLS) models to focus the analysis of the longitudinal variation in the infant faecal microbiome, HM microbiome, and HM metabolome, we identified microbial taxa and metabolite signatures linked to maternal ppBMI and early infant growth. By comparing the associations of the modelled microbiota and metabolites with maternal ppBMI and infant 6-month WHZ in their respective models, we found that some features were associated with both metrics while others were associated with only one. These findings shed light on the complex interactions between maternal health, HM composition, and infant gut microbiome development during breastfeeding and their links to infant growth. The lack of a clear link between maternal ppBMI and 6-month infant WHZ was unexpected, given the shared feature loadings between the NPLS models (Fig. 4). A possible explanation is that the effect of maternal ppBMI is neutralized by the other compartments of the system, such as the infant gut microbiome or the infant immune system. Additionally, features with strong associations with only one metric indicate independent effects within the system. Unmeasured variables, such as unrecorded maternal diet, breastfeeding frequency and rate of transition to solid food, intrauterine fetal growth trajectories, genetics, environmental factors, or unmeasured dataset properties (e.g., HM lipids, proteins, or species-level or genetic microbiome differences), may also confound the relationship between maternal overweight and infant growth. This was also reflected by the bacterial subcommunities that were associated with high and low ppBMI, consisting of a mixture of beneficial and dysbiosis-related zOTUs. The homogeneity of our cohort in terms of all exclusively breastfeeding should also be considered, as exclusive breastfeeding has been shown to mitigate the transmission of maternal obesity 14,30 . It is also possible that a high maternal ppBMI associated infant faecal microbiome signal is masked during the exclusive breastfeeding period but could reappear later in life. The HMO profile of the HM could also be more directly influenced by maternal dietary uptake, rather than by the overweight phenotype 31 . The association of multiple HMOs and Bifidobacterium spp. with high maternal ppBMI is an unexpected result, as both are considered to be highly beneficial to early-life gut function 20,32–36 ,while high maternal ppBMI have been associated with the infant also being in risk of excessive weight gain (REFS). Their enrichment may reflect metabolic trade-offs driven by altered HMOs that promote Bifidobacterium colonization. This could mean that high ppBMI-associated HMOs and Bifidobacterium spp. counteract the effects of other detrimental microbiota members or compounds linked to maternal obesity, thereby counteracting the inter-generational transmission of the phenotype. In contrast, E. coli and Klebsiella spp. were associated with low infant 6-month WHZ, likely reflecting dysbiosis or inflammation that impairs growth 37–40 . Evidence also links low maternal ppBMI and low infant 6-month WHZ to typically pathogenic species 41,42 , including Cutibacterium acnes , Pseudomonas spp. and Staphylococcus aureus , as well as Acinetobacter spp., which has been associated with postpartum stress 43 . These findings highlight potential pathways through which maternal obesity might influence the infant microbiome development, both via changes in HM composition and via indirect effects mediated by microbiota shifts. For example, the enrichment of Bifidobacterium spp. may represent a protective mechanism against adverse effects of maternal obesity, while the depletion of Klebsiella spp. and E. coli may indicate disrupted microbial dynamics. Future studies integrating functional assays are needed to disentangle these interactions and implications for targeted nutritional interventions. We have observed that the NPLS models describe 2–6% of the variation in the independent dataset and 9–31% of the variation in the dependent variable. While this may seem limited, they demonstrate that biological variation of interest is often hidden underneath other sources of variability. Additionally, variation of interest in microbiome data tends to be very low in general, and there is very little variation in the milk metabolomics data due to its composition being the same across time. Despite these issues, we have shown that these models are predictive, and that the features we identified are known to play a role in early infant nutrition and gut function. We have shown that highly focused data analysis methods are needed to find weak signals in such datasets. In this work we have analysed each independent dataset separately. Other multi-way data integration approaches like Advanced Coupled Matrix and Tensor Factorization (ACMTF) could also have been used to better identify shared and distinct sources of variation 44–46 . However, ACMTF solutions cannot be steered towards a specific solution by maximising covariation using a dependent variable. Future work should focus on adapting the ACMTF algorithm to include a joint regression step using a dependent variable of interest. In conclusion, we identified microbial subcommunities linked to maternal ppBMI and infant growth, unexpectedly finding that multiple HMOs and bifidobacteria associated with high maternal ppBMI and high infant 6 month WHZ scores. This highlights pathways through which maternal obesity could affect infant microbiome development, by a combination of changes in HM composition and secondary effects mediated by microbiota alterations. While no direct correlation between maternal ppBMI and infant growth was observed within the first three years of life, the application of NPLS modeling proved to be a key methodological approach to identifying shared and distinct microbial and metabolite signatures associated with each metric. These findings not only advance our understanding of maternal-infant nutritional dynamics but also underscore the utility of NPLS as a method for exploring longitudinal relationships in a multi-omics context. Limitations While this study focuses on a relatively homogenous cohort of healthy, vaginally delivered, exclusively breastfed infants, the findings may not generalize to populations with differing delivery modes, feeding practices or maternal health profiles. For example, caesarean delivery, formula feeding and antibiotics use are known to influence infant gut microbiome development and may interact with maternal ppBMI in different ways 47,48 . Future research should include diverse cohorts to evaluate whether the observed microbiota and metabolite associations are consistent across these contexts. Unmeasured variables such as maternal diet, breastfeeding duration, and environmental exposures may have influenced the observed relationships. For instance, dietary intake is known to affect HMO profiles, potentially confounding associations attributed to maternal ppBMI. Single-point weight measurements also have a degree of uncertainty, due to the coMaternal obesity is a key risk factor for excessive fetal growth and childhood obesity, yet its influence on human milk (HM) composition and the infant gut microbiome development remains unclear. This considerable variation in day-to-day weight of infants, combined with intermittent periods of weight-gain and growth-spurts 49,50 . Furthermore, offspring that is large for gestational age at delivery may display a growth pattern with diminished weight accretion in the first weeks or months of life, and newborns of small size at delivery may have an accelerated weight gain 51 . Such “catch-up” or “slow-down” effects could obscure the associations investigated in the present study, even though the study population intendedly had a homogeneous construction. Follow-up studies measuring growth in later life may shed light on such issues. Methods Participants and Sample Collection Infants and their mothers who provided samples for this study were recruited as part of the MaInHealth (Maternal Infant Health) cohort established in Aarhus, Denmark, to investigate the natural human milk (HM) variation and its possible effects on offspring metabolism and gut microbiota 23 . The cohort has been approved by the Central Denmark Regional Committees on Health Research Ethics (ethical approval reference: 1-10-72-296-18) and registered on ClinicalTrials. gov (identification number: NCT05111990). Pregnant women were recruited from Aarhus University Hospital, Aarhus, Denmark, from 2019 to 2021. Women included in the study were healthy, non-smokers, expecting to give birth vaginally, and planning to breastfeed for the first four to six months. Pre-pregnancy weight and height were self-reported by the mothers upon recruitment at gestational week 18–20. Infants included were healthy, with birth weights of 2500-5000g and were born full-term, i.e. gestational week 37 + 0 or later. See the study protocol for a detailed project description and recruitment and exclusion criteria 23 . Briefly, the mothers collected milk during the first week postpartum and at one, two, and three months after birth. Approximately 20 mL of foremilk was taken during each collection, avoiding the initial drops. Fecal samples (~ 2 g) were collected at the same time points from the first faeces passed after HM collection. These samples were stored in the participants' freezers at -20°C for up to two weeks, then transported on dry ice to Aarhus University, where they were stored at -80°C until analysis. HM samples were thawed, mixed, divided into 1 mL aliquots, and returned to -80°C storage. One aliquot was used for metabolomics and another for microbiome characterization. Faecal samples were thawed, and 250 mg was mixed with PBS buffer at a 1:5 ratio (w/v), vortexed, and centrifuged at 10,000 × g for 10 minutes at 4°C. The pellet was frozen at -80°C for DNA extraction, while the supernatant was saved for metabolomics analysis. BMI was calculated as (weight in kg/(height in m) 2 ). Maternal pre-pregnancy BMI was stratified into three groups: normal weight (NW; BMI 25 or lower), overweight (OW; BMI 25–30) and obesity (OB; BMI 30 or higher) ( Supplementary Fig. 1 ). Infant anthropometric z-scores indicate standard deviations from the mean height-for-age, weight-for-age, weight-for-height, and BMI-for-age referenced to the 2006 World Health Organization child growth standards, calculated in R v4.2.1 52 using the “zscorer” package v0.3.1 53 . Infant anthropometrics were recorded at birth and 1, 2, 3, 6, 12, 24 and 36 months after birth ( Supplementary Fig. 2 ). 1 H Nuclear Magnetic Resonance Spectroscopy Metabolomics Analysis of Milk HM samples for 1 H nuclear magnetic resonance spectroscopy (NMR)-based metabolomics were processed following a standard protocol for milk-based metabolomics as described previously 54 . Thawed samples were centrifuged to remove the fat layer, filtered, and placed in NMR tubes with deuterated water (D₂O) and 3-(trimethylsilyl) propionic acid (TSP) for referencing. 1 H NMR spectra acquisition was performed using a Bruker NEO 600 spectrometer at 300 K and a 1 H frequency of 600.03 MHz. Spectra were referenced to the TSP signal at 0 ppm and processed using Topspin 4.09 software for phase and baseline correction. Metabolites were identified using the Chenomx NMR suite 10.1 (Chenomx Inc., Edmonton, AB, Canada) and normalised to lactose concentration. A weighted normalisation approach was applied to account for lactose level variation, combining lactose and total metabolite factors. Metabolite concentrations were centred for analysis but not scaled. Metabolite identification was performed using the Chenomx NMR Suite 10.1 and compared with the Chenomx standard metabolite library, supplemented with an in-house HMO library. Lactose constitutes > 80% of the total concentration of the HM metabolites detected by NMR-based metabolomics and significantly varies between stages of lactation 55 . Unidentified or partially identified compounds were included in the analysis and normalisation, as their area under-curve still allowed quantification. A weighted normalisation with lactose weight 3 and total metabolites weight 1 was therefore used: a lactose factor normalising to the average concentration in HM of 215 mM (L-factor) and a total metabolite (excluding lactose) across all samples (T-factor) were combined as the factor (L-factor×3 + T-factor)/4. Metabolite concentrations were centred but not scaled before analysis. Sequencing of bacterial communities in mother’s milk and infant faecal samples DNA Extraction and preparation for sequencing Sample processing and preparation for sequencing were performed as described previously 24 . Briefly, HM samples were thawed at 4°C and centrifuged at 12,000 × g for 20 minutes at 4°C to remove the fat layer with a sterile cotton swab. DNA was extracted from the remaining pellet using the Bead-Beat Micro AX Gravity Kit (A&A Biotechnology, Gdynia, Poland) according to the manufacturer’s protocol. Negative controls with sterile MilliQ water were used during extraction, PCR, and sequencing. A 16S rRNA gene amplicon library was constructed by amplifying the 16S rRNA gene using primers targeting the 16S V1-V9 regions, as described previously 56 . In brief, amplicons were purified with SpeedBeads™ magnetic carboxylate (Sigma-Aldrich) and verified (~ 1500 bp) via agarose gel electrophoresis. Sequencing libraries were prepared by pooling barcoded PCR products from up to 196 samples and sequenced using Oxford Nanopore Technologies’ GridION X5. Preprocessing of raw reads Raw reads were processed with the Long Amplicon Consensus Analysis (LACA) pipeline for de-novo clustering and taxonomic classification using the SILVA v138.1 database 57 . zOTU (zero-radius operational taxonomic units) consensus sequences were obtained using a multiple de-novo clustering approach. zOTUs present in fewer than 5% of samples and with mean relative abundances below 0.05% were filtered out, retaining 98% of the total reads. Processing of microbiome data The zOTU data and taxonomic information of the infant faecal microbiome and mother milk microbiome were processed in MATLAB (version 2023a). For the infant faecal microbiome data, features were kept if they had ≤ 75% sparsity across the dataset. This step resulted in 93 out of 565 features being selected. For the mother milk microbiome data, features were kept if they had ≤ 85% sparsity across the dataset. This step resulted in 115 out of 707 features being selected. Sparsity filtering was purposefully done across all samples, as opposed to groupwise, to keep the processing consistent between the maternal ppBMI and infant 6-month WHZ-based models. We performed a centred log-ratio transformation with a pseudo-count of 1 to correct for compositionality 58,59 . Subsequently the data was converted to a three-way array, keeping missing measurements as a row of NAs. In the infant faecal microbiome data cube, subject 332 was removed due to being an outlier, resulting in a cube of size 132 subjects x 93 microbial abundances x 3 time points. This was not necessary for the mother milk microbiome data, yielding a three-way array of size 169 subjects x 115 microbial abundances x 4 time points. The data was then centred across the subject mode to make the samples comparable per time point and scaled within the feature mode to make all microbial abundances equally important for the modelling procedure 29,60 . Processing of metabolomics data The mother milk metabolomics data were processed using MATLAB (version 2023a). Values below the detection limit were imputed with a random value between 0 and the detection limit per metabolite to preserve their distribution. Next, the dataset was (natural) log transformed to stabilise the variance. The dataset was then converted to a three-way array of size 164 subjects x 80 metabolites x 4 time points. Subsequently the data was then centred across the subject mode to make the samples comparable per time point and scaled within the feature mode to make all metabolites equally important for the modelling procedure 60 . Notation and Definitions We briefly define the mathematical notation that will be used throughout this paper. Scalars are indicated by lower-case italics such as a . Vectors are indicated by bold lower-case characters such as b . Two-way matrices are indicated with bold capitalised characters such as X . Underlined bold capitalised characters are used for three-way arrays such as . The letters I , J , and K are reserved to indicate the dimension of the subject, microbial or metabolite abundance, and time mode, respectively. Hence the element for subject i , microbial abundance j at time point k of a three-way array is called \(\:{x}_{ijk}\). We create the data cubes in this study such that the subjects are in the first mode, the measured microbial abundances or metabolites are in the second mode, and time points are in the third mode. We will not distinguish between the terms ‘factor’ and ‘component’, nor between the terms ‘way’ and ‘mode’. Generally, we will use the words ‘component’ and ‘mode’ throughout this paper. N-way Partial Least Squares (NPLS) Details on the creation of NPLS models for various types of data have been described elsewhere 28 . The aim of NPLS is to find a model of the input data $$\:{x}_{ijk}={t}_{i}{{w}^{J}}_{j}{{w}^{K}}_{k}$$ Where \(\:{x}_{ijk}\) is the element in corresponding to the i -th subject, j -th microbial or metabolite abundance and k -th timepoint, t contains the subject scores, \(\:{w}^{J}\) contains the feature loadings and \(\:{w}^{K}\) contains the time loadings such that the covariance of t and a centred output variable y is maximised and the sum of squares of the residuals is minimised. The NPLS implementation from the N-way toolbox (version 1.8.0.0, https://nl.mathworks.com/matlabcentral/fileexchange/1088-the-n-way-toolbox) in MATLAB (version 2023a) was used to create NPLS models for all datasets with maternal ppBMI or infant 6-month WHZ as dependent variables 61 . Regressing to changes in WHZ (deltaWHZ) and BMIz score (deltaBMIz) at 6 months, i.e., whether the infant is growing faster or slower than the median growth curve, did not produce biologically interpretable models. Similar to Parallel Factor Analysis 28,29 , the correct number of components of the NPLS model needed to be determined to create an optimal model. This was done by inspecting the variation explained in the input data as well as the root mean-squared error of prediction (RMSEP) of the dependent variable ( Supplementary Figs. 5–6 ). For all models shown in this paper, a 1-component model was sufficient. While the RMSEP metric suggests a two-component model for the HM metabolome data regressed onto maternal ppBMI, we found that the extra component did not describe any relevant biological variation. All plots and subsequent analyses were created with the ggplot2 62 package in R 63 (ggplot2 v3.5.1, R v4.4.1). All generated NPLS models reported in the main text are shown in Supplementary Figs. 3 and 4 . The predictive power of all NPLS models described in the main text is reported in Table 3. Correlations of the subject scores of the NPLS models with various metadata of interest are reported in Supplementary Tables 2 and Supplementary Fig. 13 . Correlations of the fitted features with the dependent variable were used to determine associations with maternal ppBMI and infant 6-month WHZ. Post-hoc clustering of NPLS loadings Clustering of the feature loadings can help identify microbial subcommunities that are connected to a particular time profile in the NPLS model. We performed the clustering procedure of the microbiota loadings as follows. Microbiota were not considered if the variation explained by the model for their entire time trajectory was lower than that of the model itself. The remaining microbiota were then clustered based on their modelled time trajectories using the K-medoids algorithm from the cluster R package 64 (v2.1.6) with 50 random starts to be robust against outliers. The number of clusters was determined using the within-cluster sum of squares, silhouette width and gap statistic metrics as reported by the factoextra R package 65 (v1.0.7; Supplementary Figs. 16–19 ). Relative abundances per cluster were summed per subject, after which the mean and standard error of the mean were calculated per maternal ppBMI group (Fig. 3). Relative abundance sum permutation test results are reported in Supplementary Table 4 . Cluster members are reported in Supplementary Tables 3 . A similar analysis was performed to identify microbial subcommunities connected to infant 6-month WHZ groups are reported in Supplementary Fig. 20 and Supplementary Table 3 . Pairwise comparisons of NPLS feature loadings For each independent dataset (HM microbiome, HM metabolome, and infant faecal microbiome), feature loadings from the NPLS model regressed onto maternal ppBMI were compared with those from the model regressed onto infant 6-month WHZ. Scatterplots were generated with feature loadings from the ppBMI-associated model on the x-axis and those from the WHZ-associated model on the y-axis, enabling identification of shared and distinct associations. Visualizations and annotations were created using the ggplot2 66 package in R 63 (ggplot2 v3.5.1, R v4.4.1). A full overview of feature loadings is provided in Supplementary Figs. 7–12 , with detailed associations listed in Supplementary Tables 5–7 . Declarations Data Availability Raw sequencing data is available at the European Nucleotide Archive (ENA) at (https://www.ebi.ac.uk/ena/browser/view/PRJEB82744). Code Availability The underlying code for this study is available on GitHub and can be accessed via https://doi.org/10.5281/zenodo.14899658. Ethics statement Informed consent was obtained from both parents per the Declaration of Helsinki II. Ethical approval for this study was granted by The Central Denmark Regional Committee on Health Research Ethics (journal number 1-10-72-296-18v6). The study is registered at ClinicalTrials.gov, with the identifier NCT05111990. Author Contributions R.R. Jakobsen: investigation, methodology, writing - original draft; G.R. van der Ploeg: formal analysis, methodology, writing - original draft; U.K. Sundekilde: conceptualisation,funding,sample collection,laboratory analysis; J. Astono: sample collection,laboratory analysis; K.O. Poulsen: sample collection,laboratory analysis; J.A. Westerhuis: conceptualization, methodology, supervision, formal analysis, writing - original draft; J.Fuglsang: participant recruitment, clinical follow-up, A. Heintz-Buschart: conceptualization, formal analysis, supervision, writing - original draft; A.K. Smilde: conceptualization, methodology, supervision, funding acquisition, writing - review & editing; D.S. Nielsen: conceptualization, supervision, funding, writing - review & editing. All authors reviewed and approved the manuscript. Competing interests The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. K.O. Poulsen currently holds a position at Arla Foods, but was employed at AU Food when contributing to the project. Acknowledgements G.R. van der Ploeg was funded by a grant from the University of Amsterdam, Research Priority Area on Personal Microbiome Health. Funding for R.R. Jakobsen and the MAINHEALTH study was provided from an Arla Food for Health strategic research grant. Funding for K.O. Poulsen was provided by the Sino-Danish Center for Research. Data was generated by accessing research infrastructure at Copenhagen University and Aarhus University, including equipment financed by FOODHAY (Food and Health Open Innovation Laboratory, Danish Roadmap for Research Infrastructure). References Monteiro, P. O. A. & Victora, C. G. Rapid growth in infancy and childhood and obesity in later life – a systematic review. Obes. Rev. 6 , 143–154 (2005). Gaudet, L., Ferraro, Z. M., Wen, S. W. & Walker, M. Maternal obesity and occurrence of fetal macrosomia: A systematic review and meta-analysis. BioMed Res. Int. 2014 , (2014). Voerman, E. et al. Maternal body mass index, gestational weight gain, and the risk of overweight and obesity across childhood: An individual participant data meta-analysis. PLOS Med. 16 , e1002744 (2019). Weng, S. F., Redsell, S. A., Swift, J. A., Yang, M. & Glazebrook, C. P. Systematic review and meta-analyses of risk factors for childhood overweight identifiable during infancy. Arch. Dis. Child. 97 , 1019–1026 (2012). Horta, B. L., Rollins, N., Dias, M. S., Garcez, V. & Pérez-Escamilla, R. Systematic review and meta-analysis of breastfeeding and later overweight or obesity expands on previous study for World Health Organization. Acta Paediatr. 112 , 34–41 (2023). Amir, L. H. & Donath, S. A systematic review of maternal obesity and breastfeeding intention, initiation and duration. BMC Pregnancy Childbirth 7 , 9 (2007). Ong, K. & Loos, R. Rapid infancy weight gain and subsequent obesity: Systematic reviews and hopeful suggestions. Acta Pædiatrica 95 , 904–908 (2006). Lind, M. V., Larnkjær, A., Mølgaard, C. & Michaelsen, K. F. Breastfeeding, Breast Milk Composition, and Growth Outcomes. (2018) doi:10.1159/000486493. Ahearn-Ford, S., Berrington, J. E. & Stewart, C. J. Development of the gut microbiome in early life. Exp. Physiol. 107 , 415 (2022). Sarkar, A., Yoo, J. Y., Dutra, S. V. O., Morgan, K. H. & Groer, M. The Association between Early-Life Gut Microbiota and Long-Term Health and Diseases. J. Clin. Med. 2021 Vol 10 Page 459 10 , 459 (2021). Stokholm, J. et al. Maturation of the gut microbiome and risk of asthma in childhood. Nat. Commun. 2018 91 9 , 1–10 (2018). Isganaitis, E. et al. Maternal obesity and the human milk metabolome: associations with infant body composition and postnatal weight gain. Am. J. Clin. Nutr. 110 , 111–120 (2019). Saben, J. L., Sims, C. R., Piccolo, B. D. & Andres, A. Maternal adiposity alters the human milk metabolome: associations between nonglucose monosaccharides and infant adiposity. Am. J. Clin. Nutr. 112 , 1228–1239 (2020). Beyerlein, A. & Von Kries, R. Breastfeeding and body composition in children: will there ever be conclusive empirical evidence for a protective effect against overweight? Am. J. Clin. Nutr. 94 , S1772–S1775 (2011). Horta, B. L., Rollins, N., Dias, M. S., Garcez, V. & Pérez-Escamilla, R. Systematic review and meta-analysis of breastfeeding and later overweight or obesity expands on previous study for World Health Organization. Acta Paediatr. 112 , 34–41 (2023). Vieira Queiroz De Paula, M., Grant, M., Lanigan, J. & Singhal, A. Does human milk composition predict later risk of obesity? A systematic review. BMC Nutr. 9 , 1–10 (2023). Petraroli, M., Castellone, E., Patianna, V. & Esposito, S. Gut Microbiota and Obesity in Adults and Children: The State of the Art. Front. Pediatr. 9 , 657020 (2021). Cheng, L., Akkerman, R., Kong, C., Walvoort, M. T. C. & de Vos, P. More than sugar in the milk: human milk oligosaccharides as essential bioactive molecules in breast milk and current insight in beneficial effects. Crit. Rev. Food Sci. Nutr. 61 , 1184–1200 (2021). Thomson, P., Medina, D. A. & Garrido, D. Human milk oligosaccharides and infant gut bifidobacteria: Molecular strategies for their utilization. Food Microbiol. 75 , 37–46 (2018). Korpela, K. et al. Childhood BMI in relation to microbiota in infancy and lifetime antibiotic use. Microbiome 5 , 26 (2017). Demmelmair, H., Jiménez, E., Collado, M. C., Salminen, S. & McGuire, M. K. Maternal and Perinatal Factors Associated with the Human Milk Microbiome. Curr. Dev. Nutr. 4 , nzaa027 (2020). Daiy, K., Harries, V., Nyhan, K. & Marcinkowska, U. M. Maternal weight status and the composition of the human milk microbiome: A scoping review. PLoS ONE 17 , e0274950 (2022). Poulsen, K. O. et al. Influence of maternal body mass index on human milk composition and associations to infant metabolism and gut colonisation: MAINHEALTH - a study protocol for an observational birth cohort. BMJ Open 12 , e059552 (2022). Jakobsen, R. et al. Maternal pre-pregnancy BMI influences breast milk composition, infant gut microbiome development, and early-life growth of term infants. Preprint at https://doi.org/10.21203/rs.3.rs-6075035/v1 (2025). Ajeeb, T. T., Gonzalez, E., Solomons, N. W., Vossenaar, M. & Koski, K. G. Human milk microbiome: associations with maternal diet and infant growth. Front. Nutr. 11 , 1341777 (2024). Baumgartel, K. et al. The Human Milk Metabolome: A Scoping Literature Review. J. Hum. Lact. 39 , 255–277 (2023). Stinson, L. F. et al. The human milk microbiome: who, what, when, where, why, and how? Nutr. Rev. 79 , 529–543 (2021). Bro, R. Multiway calibration. Multilinear PLS. J. Chemom. 10 , 47–61 (1996). van der Ploeg, G. R., Westerhuis, J. A., Heintz-Buschart, A. & Smilde, A. K. parafac4microbiome: Exploratory analysis of longitudinal microbiome data using Parallel Factor Analysis. Preprint at https://doi.org/10.1101/2024.05.02.592191 (2024). Arenz, S., Rückerl, R., Koletzko, B. & Von Kries, R. Breast-feeding and childhood obesity--a systematic review. Int. J. Obes. Relat. Metab. Disord. J. Int. Assoc. Study Obes. 28 , 1247–1256 (2004). Fan, Y., Vinjamuri, A., Tu, D., Lebrilla, C. B. & Donovan, S. M. Determinants of human milk oligosaccharides profiles of participants in the STRONG kids 2 cohort. Front. Nutr. 10 , (2023). Duranti, S. et al. Bifidobacterium bifidum and the infant gut microbiota: an intriguing case of microbe-host co-evolution. Environ. Microbiol. 21 , 3683–3695 (2019). Walker, W. A. & Iyengar, R. S. Breast milk, microbiota, and intestinal immune homeostasis. Pediatr. Res. 77 , 220–228 (2015). Stuivenberg, G. A., Burton, J. P., Bron, P. A. & Reid, G. Why Are Bifidobacteria Important for Infants? Microorg. 2022 Vol 10 Page 278 10 , 278 (2022). Vu, K. et al. From Birth to Overweight and Atopic Disease: Multiple and Common Pathways of the Infant Gut Microbiome. Gastroenterology 160 , 128-144.e10 (2021). Full article: Functional effects of human milk oligosaccharides (HMOs). https://www.tandfonline.com/doi/full/10.1080/19490976.2023.2186115. Winter, S. E. & Bäumler, A. J. Gut dysbiosis: Ecological causes and causative effects on human disease. Proc. Natl. Acad. Sci. 120 , e2316579120 (2023). Bäckhed, F. et al. Dynamics and Stabilization of the Human Gut Microbiome during the First Year of Life. Cell Host Microbe 17 , 690–703 (2015). Ju, T. et al. The Gut Commensal Escherichia coli Aggravates High-Fat-Diet-Induced Obesity and Insulin Resistance in Mice. Appl. Environ. Microbiol. 89 , e01628-22 (2023). Fåk, F., Karlsson, C. L. J., Ahrné, S., Molin, G. & Weström, B. Effects of a high-fat diet during pregnancy and lactation are modulated by E. coli in rat offspring. Int. J. Obes. 36 , 744–751 (2012). Azad, M. B., Brockway, M. M. & Reyes, S. M. Human milk composition and infant anthropometrics: overview of a systematic review with clinical and research implications. Int. Breastfeed. J. 19 , 45 (2024). Leghi, G. E. et al. The Impact of Maternal Obesity on Human Milk Macronutrient Composition: A Systematic Review and Meta-Analysis. Nutrients 12 , 934 (2020). Juncker, H. G. et al. Maternal stress in the early postpartum period is associated with alterations in human milk microbiome composition. Brain. Behav. Immun. 124 , 74–84 (2025). Acar, E., Bro, R. & Smilde, A. Data Fusion in Metabolomics Using Coupled Matrix and Tensor Factorizations. Proc. IEEE 103 , 1602 (2015). Acar, E. et al. Structure-revealing data fusion. BMC Bioinformatics 15 , 239 (2014). Acar, E., Kolda, T. G. & Dunlavy, D. M. All-at-once Optimization for Coupled Matrix and Tensor Factorizations. Preprint at http://arxiv.org/abs/1105.3422 (2011). Tun, H. M. et al. Roles of Birth Mode and Infant Gut Microbiota in Intergenerational Transmission of Overweight and Obesity From Mother to Offspring. JAMA Pediatr. 172 , 368–377 (2018). Mueller, N. T. et al. Does vaginal delivery mitigate or strengthen the intergenerational association of overweight and obesity? Findings from the Boston Birth Cohort. Int. J. Obes. 41 , 497–501 (2017). Cooke, R. et al. Catch-Up Growth in Infants and Young Children With Faltering Growth: Expert Opinion to Guide General Clinicians. J. Pediatr. Gastroenterol. Nutr. 77 , 7 (2023). Lampl, M., Thompson, A. L. & Frongillo, E. A. Sex Differences in the Relationships Among Weight Gain, Subcutaneous Skinfold Tissue and Saltatory Length Growth Spurts in Infancy. Pediatr. Res. 58 , 1238–1242 (2005). De Wit, C. C., Sas, T. C. J., Wit, J. M. & Cutfield, W. S. Patterns of Catch-Up Growth. J. Pediatr. 162 , 415–420 (2013). Team, R. C. R: A Language and Environment for Statistical Title. R Foundation for Statistical Computing. Organization, W. H. WHO Child Growth Standards : Length/Height-for-Age, Weight-for-Age, Weight-for-Length, Weight -for-Height and Body Mass Index-for-Age : Methods and Development . (World Health Organization, 2006). Sundekilde, U. K. et al. The Effect of Gestational and Lactational Age on the Human Milk Metabolome. Nutrients 8 , (2016). Andreas, N. J. et al. Multiplatform characterization of dynamic changes in breast milk during lactation. Electrophoresis 36 , 2269–2285 (2015). Andersen-Civil, A. I. S. et al. Dietary proanthocyanidins promote localized antioxidant responses in porcine pulmonary and gastrointestinal tissues during Ascaris suum-induced type 2 inflammation. FASEB J. 36 , e22256 (2022). Hui, Y., Nielsen, D. S. & Krych, L. De novo clustering of long-read amplicons improves phylogenetic insight into microbiome data. bioRxiv 2023.11.26.568539 (2023) doi:10.1101/2023.11.26.568539. Gloor, G. B., Macklaim, J. M., Pawlowsky-Glahn, V. & Egozcue, J. J. Microbiome Datasets Are Compositional: And This Is Not Optional. Front. Microbiol. 8 , 2224 (2017). Aitchison, J. The Statistical Analysis of Compositional Data. J. R. Stat. Soc. Ser. B Methodol. 44 , 139–160 (1982). Bro, R. & Smilde, A. K. Centering and scaling in component analysis. J. Chemom. 17 , 16–33 (2003). Bro, R. The N-way Toolbox. The N-way Toolbox https://nl.mathworks.com/matlabcentral/fileexchange/1088-the-n-way-toolbox (2023). Wickham, H. Data Analysis. in ggplot2 189–201 (Springer International Publishing, Cham, 2016). doi:10.1007/978-3-319-24277-4_9. R Core Team, R. R: A language and environment for statistical computing. (2013). Maechler, M. Finding groups in data: Cluster analysis extended Rousseeuw et al. R Package Version 2 , 242–248 (2019). Kassambara, A. & Mundt, F. Package ‘factoextra’. Extr. Vis. Results Multivar. Data Anal. 76 , (2017). Wickham, H., Chang, W. & Wickham, M. H. Package ‘ggplot2’. Create Elegant Data Vis. Using Gramm. Graph. Version 2 , 1–189 (2016). Additional Declarations Yes there is potential Competing Interest. The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. K.O. Poulsen currently holds a position at Arla Foods, but was employed at AU Food when contributing to the project. Supplementary Files JakobsenPloeg2025supplementaryfigurestables.docx Supplementary figures and tables Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6244750","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":433686752,"identity":"a3158af6-c4b6-47b1-b606-37d23350882f","order_by":0,"name":"Rasmus Jakobsen","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAuUlEQVRIiWNgGAWjYPACG34gYcDYwMDAOINILWmSDaRqOUyCFvP240838/w5L8HP3ryBcWYbg+zMBgJaZM7kmN3mbbstIdlzrIBxYxuD8WxCtkgw5LDd5m24XWdwI8eA8WEbQ+I8glr4nz+7zfPnnAQJWiQSzG7zsB2AaAE6LJGwwyTemN2c25YM9svBGeckjAl6X4I//dmNN3/sQCG28WFPmY3sjAOErEEGB0DBMQpGwSgYBaOACgAA5aZA7PqAnOoAAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0001-7416-1358","institution":"University of Copenhagen","correspondingAuthor":true,"prefix":"","firstName":"Rasmus","middleName":"","lastName":"Jakobsen","suffix":""},{"id":433686753,"identity":"7a6ee98b-da91-4285-9e32-af3bcbbe59ec","order_by":1,"name":"Geert Roelof van der Ploeg","email":"","orcid":"","institution":"University of Amsterdam","correspondingAuthor":false,"prefix":"","firstName":"Geert","middleName":"Roelof","lastName":"van der Ploeg","suffix":""},{"id":433686754,"identity":"836fb644-e16e-4013-bbe6-714d029225da","order_by":2,"name":"Ulrik Sundekilde","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Ulrik","middleName":"","lastName":"Sundekilde","suffix":""},{"id":433686755,"identity":"ef864ed8-b913-414f-a5dd-8b28d8bb9b28","order_by":3,"name":"Julie Astono","email":"","orcid":"","institution":"Aarhus University","correspondingAuthor":false,"prefix":"","firstName":"Julie","middleName":"","lastName":"Astono","suffix":""},{"id":433686756,"identity":"56c227d6-6afd-4c5d-8228-0133fde35d1b","order_by":4,"name":"Katrine Poulsen","email":"","orcid":"","institution":"Aarhus University","correspondingAuthor":false,"prefix":"","firstName":"Katrine","middleName":"","lastName":"Poulsen","suffix":""},{"id":433686757,"identity":"c9833d38-837e-4404-a37f-ce275579c26c","order_by":5,"name":"Jens Fuglsang","email":"","orcid":"","institution":"Aarhus University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Jens","middleName":"","lastName":"Fuglsang","suffix":""},{"id":433686758,"identity":"a15c0113-496b-45e8-82a7-29b8cb25d3d5","order_by":6,"name":"Johan Westerhuis","email":"","orcid":"https://orcid.org/0000-0002-6747-9779","institution":"University of Amsterdam","correspondingAuthor":false,"prefix":"","firstName":"Johan","middleName":"","lastName":"Westerhuis","suffix":""},{"id":433686759,"identity":"a7b1dee8-d4d6-4208-8386-33834c277eed","order_by":7,"name":"Anna Heintz-Buschart","email":"","orcid":"https://orcid.org/0000-0002-9780-1933","institution":"University of Amsterdam – SILS","correspondingAuthor":false,"prefix":"","firstName":"Anna","middleName":"","lastName":"Heintz-Buschart","suffix":""},{"id":433686760,"identity":"5875d4d3-e582-47f8-b742-a0d8ff965121","order_by":8,"name":"Age Smilde","email":"","orcid":"","institution":"University of Amsterdam","correspondingAuthor":false,"prefix":"","firstName":"Age","middleName":"","lastName":"Smilde","suffix":""},{"id":433686761,"identity":"db47bf8e-3a13-434f-87c4-d0ea5042d40c","order_by":9,"name":"Dennis Nielsen","email":"","orcid":"https://orcid.org/0000-0001-8121-1114","institution":"Department of Food Science, University of Copenhagen","correspondingAuthor":false,"prefix":"","firstName":"Dennis","middleName":"","lastName":"Nielsen","suffix":""}],"badges":[],"createdAt":"2025-03-17 12:50:29","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6244750/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6244750/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":79579952,"identity":"84c20ea0-eb2b-41af-a886-bc8cdea52829","added_by":"auto","created_at":"2025-03-31 11:43:59","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":141937,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eStudy overview and sample compositions. \u003c/strong\u003e\u003cem\u003e(A) Concept figure showing the hypothesised connection between maternal overweight and infant growth: maternal overweight affects the HM metabolite profile, which during the exclusive breastfeeding directly affects growth while also influencing both the HM microbiome and the infant faecal microbiome, resulting in additional indirect effects on infant growth. (B) Overview of sample and metadata collection and timeline. (C) Infant weight-for-length/height z-score at 6 months per maternal ppBMI group over time. No significant differences in WHZ were observed between maternal ppBMI groups. Bacterial relative abundance bar plots summarising genus-level composition of HM and infant faecal samples by (D) maternal ppBMI, and (E) days after birth. Metabolite concentrations of the 15 most abundant HM components, grouped by (F) maternal ppBMI and (G) days after birth; the ‘other metabolites’ category is the summed concentration of all other detected compounds, BH-corrected Wilcoxon rank-sum test (*: p≤0.05, **: p≤0.01, ***: p≤0.001), BMIz: BMI-for-age, WHZ: weight-for-length/height z-score.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6244750/v1/edb1a40d484439a4f7ed429e.jpg"},{"id":79580845,"identity":"b59e2a92-dd1e-41fd-8700-c8f462d719c1","added_by":"auto","created_at":"2025-03-31 11:51:59","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":157291,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eStatistical analysis.\u003c/strong\u003e\u003cem\u003e\u003cstrong\u003e \u003c/strong\u003e\u003c/em\u003e\u003cem\u003e(A) Visual description of N-way Partial Least Squares (NPLS). A decomposition of X into triads of subject loadings t, feature loadings W\u003c/em\u003e\u003csub\u003e\u003cem\u003ej\u003c/em\u003e\u003c/sub\u003e\u003csub\u003e\u003cem\u003e \u003c/em\u003e\u003c/sub\u003e\u003cem\u003eand time loadings w\u003c/em\u003e\u003csub\u003e\u003cem\u003ek\u003c/em\u003e\u003c/sub\u003e\u003cem\u003e is sought such that the covariance of t and y is maximised while the residual sum of squares of the model of X is minimised. (B) An NPLS model was created for each of the HM metabolome, HM microbiome and infant faecal microbiome datasets, regressing on either maternal ppBMI or infant 6-month WHZ. Letters refer to the other panels in this figure where the feature loadings are examined. (C) HM metabolite associations to high and low ppBMI. (D) HM microbiota associations to high and low ppBMI. (E) Infant faecal microbiota associations to high and low ppBMI. (F) HM metabolite associations to high and low infant 6-month WHZ. (G) HM microbiota associations to high and low infant 6-month WHZ. (H) Infant faecal microbiota associations to high and low infant 6-month WHZ. Feature loadings in C-F were filtered to an absolute value \u003c/em\u003e≥\u003cem\u003e 0.1 for visualisation purposes, showing only the strongest associations. The unfiltered version of these plots are shown in \u003c/em\u003e\u003cem\u003e\u003cstrong\u003eSupplementary Figures 7-12\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6244750/v1/002f5d5ee4e0718579fcab67.jpg"},{"id":79579951,"identity":"9e60fbb2-0621-4fa4-9ac2-dcdc623172bf","added_by":"auto","created_at":"2025-03-31 11:43:58","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":68952,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eRelative abundance sum per identified zOTU cluster for each maternal ppBMI group.\u003c/strong\u003e The total relative abundance (out of 1) per identified zOTU cluster using the fitted responses from the NPLS models regressed onto ppBMI of (A) the infant faecal microbiome and (B) the HM microbiome. \u003cem\u003eError bars correspond to the standard error of the mean (SEM) across all samples per maternal ppBMI group. The mean difference between the normal and overweight-or-obese group per time point was tested using a Benjamini-Hochberg corrected permutation test of 999 iterations (*: p≤0.05, **: p≤0.01, ***: p≤0.001). See \u003c/em\u003e\u003cem\u003e\u003cstrong\u003eSupplementary Table 3\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e for the list of identified species per zOTU cluster.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6244750/v1/8ebdb8c9783a4ad5e36ab0f5.jpg"},{"id":79579959,"identity":"25a4fddb-72b4-4ec7-aa6c-4d12037ef2c7","added_by":"auto","created_at":"2025-03-31 11:43:59","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":157111,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eWithin-dataset comparisons of bacterial and metabolite loadings. \u003c/strong\u003e(A) \u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cem\u003eAn NPLS model was created for each of the HM metabolome, HM microbiome and infant faecal microbiome datasets, regressing on either maternal ppBMI or infant 6-month WHZ. Letters refer to the other panels in this figure where the feature loadings are compared. (B) Feature loading comparison for the NPLS models describing the HM metabolome regressed onto maternal ppBMI (x-axis) or infant 6-month WHZ (y-axis). (C) Feature loading comparison for the NPLS models describing the HM microbiome regressed onto maternal ppBMI (x-axis) or infant 6-month WHZ (y-axis). (D) Feature loading comparison for the NPLS models describing the infant faecal microbiome regressed onto maternal ppBMI (x-axis) or infant 6-month WHZ (y-axis). A full overview of the loadings per metabolite and microbiota is given in \u003c/em\u003e\u003cem\u003e\u003cstrong\u003eSupplementary Figures 7-12\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e. An overview of the associations per metabolite and microbial taxa is given in \u003c/em\u003e\u003cem\u003e\u003cstrong\u003eSupplementary Tables 5-7\u003c/strong\u003e\u003c/em\u003e\u003cem\u003e.\u003c/em\u003e\u003c/p\u003e","description":"","filename":"4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-6244750/v1/24519a5873339022bbded8d3.jpg"},{"id":80256133,"identity":"24484ea8-4cd8-4df0-ad06-fc153d80ed3f","added_by":"auto","created_at":"2025-04-09 19:06:14","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2484332,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6244750/v1/2be604ab-471d-4b0d-8a02-15d9d74b0095.pdf"},{"id":79580852,"identity":"c0819924-f532-4bff-afc3-1e61d7d75785","added_by":"auto","created_at":"2025-03-31 11:51:59","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":5724787,"visible":true,"origin":"","legend":"Supplementary figures and tables","description":"","filename":"JakobsenPloeg2025supplementaryfigurestables.docx","url":"https://assets-eu.researchsquare.com/files/rs-6244750/v1/464944cd31612bf530849f2f.docx"}],"financialInterests":"\u003cb\u003eYes\u003c/b\u003e there is potential Competing Interest.\nThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. K.O. Poulsen currently holds a position at Arla Foods, but was employed at AU Food when contributing to the project.","formattedTitle":"Supervised Modelling of Longitudinal Human Milk and Infant Gut Microbiome Reveal Maternal Pre-Pregnancy BMI and Early Life Growth Interactions","fulltext":[{"header":"Introduction","content":"\u003cp\u003eThe first months of life represent a critical window for growth and development, during which excessive weight gain is associated with later obesity risk\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e. Maternal obesity is a significant risk factor for excessive fetal growth\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e, infant overweight, and childhood obesity\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e,\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e. Although the mechanisms underlying the intergenerational transmission of obesity are not fully understood, early-life nutrition is thought to play a critical role. Among its numerous health benefits, breastfeeding is linked with a decreased risk of childhood obesity and slower growth rates compared to formula feeding\u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e. However, maternal obesity is associated with reduced initiation and shorter durations of breastfeeding, potentially diminishing these protective effects\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e,\u003cspan additionalcitationids=\"CR7\" citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eHuman milk (HM) provides multiple bioactive compounds that influence infant growth and the establishment of the infant gut microbiome\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e. During infancy, the infant gut microbiome undergoes rapid compositional changes, with its maturation by 1\u0026ndash;2 years strongly linked to long-term health outcomes\u003csup\u003e\u003cspan additionalcitationids=\"CR10\" citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e. Emerging evidence suggests that maternal pre-pregnancy BMI (ppBMI) can affect HM composition and the initial microbial colonization of the infant gut, potentially influencing both infant growth trajectories and infant gut microbiome establishment\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e. For example, high maternal ppBMI has been associated with increased HM content of monosaccharides and sugar alcohols, and decreased HM content of oligosaccharides\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e,\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e. However, studies investigating the relationship between HM composition, infant growth and maternal ppBMI are scarce and largely limited to cross-sectional analyses\u003csup\u003e\u003cspan additionalcitationids=\"CR15\" citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eSeveral mechanisms may explain how HM composition influences growth and reduces obesity risk relative to formula feeding\u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA). Bioactive HM components, including human milk oligosaccharides (HMOs) and lipids, shape the infant gut microbiome and promote a profile associated with metabolic health\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e. For example, HMOs are selectively utilized by \u003cem\u003eBifidobacterium\u003c/em\u003e spp., which in turn have been associated with lower obesity risk\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e,\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e. Additionally, microbial transfer from HM to the infant gut is also hypothesized to influence early microbiota assembly\u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e. These mechanisms may act synergistically with the direct provision of HM nutrients to modulate growth patterns. Maternal ppBMI is also associated with altered HM microbiome and metabolome composition\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e,\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e,\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e, which might also be a key pathway contributing to the intergenerational transmission of obesity.\u003c/p\u003e \u003cp\u003eThe disentanglement of ppBMI associated changes in HM and their interactions with the infant gut microbiome is needed to uncover the mechanisms of inter-generational obesity. To investigate these, we recruited and longitudinally sampled a cohort of 169 mother-infant dyads at 3\u0026ndash;4 time points across the first three months of exclusive breastfeeding, followed by growth anthropometrics at 6 months through 3 years\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e. Infant anthropometrics were recorded from birth through three years postpartum. An exploratory analysis of these datasets using principal component analysis (PCA) and principal coordinates analysis (PCoA) reported that maternal pre-pregnancy BMI and infant growth significantly correlated with HM metabolite profile and infant gut microbial diversity and composition with significant associations between HM oligosaccharide clusters and infant gut bifidobacteria\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e. However, PCA and PCoA only capture the dominant source of variation and do not consider longitudinal dynamics. To address these limitations, we applied N-way Partial Least Squares (NPLS) modelling to investigate microbial and metabolite signatures associated with maternal ppBMI and infant growth. NPLS is well-suited for this type of analysis since it enables simultaneous analysis of longitudinal microbiome and metabolomics data to uncover biologically relevant associations. By integrating temporal dynamics across the HM metabolome, microbiome and infant gut microbiome, NPLS captures relationships between features within these datasets and maternal and infant phenotypes. By elucidating these relationships, our findings may inform targeted nutritional interventions, such as supplementing key components to breast milk or using tailored infant formula, to mitigate obesity risk and promote optimal infant health.\u003c/p\u003e"},{"header":"Results","content":"\u003ch2\u003eMicrobiome and Metabolome Profiles Differ Between Maternal ppBMI Groups and Shift Over Time\u003c/h2\u003e\n\u003cp\u003eThis study included samples from 169 healthy Danish mother-infant dyads from the MAINHEALTH cohort\u003csup\u003e23\u003c/sup\u003e, consisting of human milk (HM) samples collected at days 3, 30, 60 and 90 postpartum and infant faecal samples collected at days 30, 60 and 90 postpartum, along with extensive anthropometric and clinical data (\u003cstrong\u003eFigure 1B\u003c/strong\u003e). The offspring were exclusively breastfed throughout this period. The average pre-pregnancy BMI (ppBMI) was 27\u0026plusmn;5.4 kg/m\u003csup\u003e2\u003c/sup\u003e, and the mean gestational age at delivery was 283\u0026plusmn;2.5 days (mean \u0026plusmn; sd, \u003cstrong\u003eTable 1\u003c/strong\u003e). Infants (86 male, 80 female) had average birth weights of 3704\u0026plusmn;450 g. The maternal cohort was stratified into three ppBMI groups: normal weight (NW; BMI \u003cem\u003e\u0026lt;\u003c/em\u003e 25 kg/m\u003csup\u003e2\u003c/sup\u003e), overweight (OW; BMI 25-30 kg/m\u003csup\u003e2\u003c/sup\u003e) and obesity (OB; BMI \u0026ge; 30 kg/m\u003csup\u003e2\u003c/sup\u003e) (\u003cstrong\u003eSupplementary Figure 1\u003c/strong\u003e). Infant anthropometrics were recorded at birth and 1, 2, 3, 6, 12, 24 and 36 months after birth (\u003cstrong\u003eSupplementary Figure 2\u003c/strong\u003e). HM metabolites were characterised using \u003csup\u003e1\u003c/sup\u003eH nuclear magnetic resonance spectroscopy (NMR) and HM and infant faecal bacterial communities by V1-V9 16S rRNA gene Nanopore long-read amplicon sequencing. After quality control, 570 HM metabolite samples, 495 HM microbiome samples, and 348 infant faecal microbiome samples were included for analysis.\u003c/p\u003e\n\u003cp\u003eExamining the HM and infant faecal microbiomes, both showed a difference in composition between maternal ppBMI groups and over time, consistent with previous reports\u003csup\u003e25\u0026ndash;27\u003c/sup\u003e. At the genus level, infant faecal microbiomes from overweight-or-obese mothers were depleted of \u003cem\u003eKlebsiella\u0026nbsp;\u003c/em\u003e(p=0.008, BH-corrected Kruskall-Wallis test) relative to infants normal ppBMI mothers (\u003cstrong\u003eFigure 1D\u003c/strong\u003e). HM microbiomes from overweight-or-obese mothers were enriched in \u003cem\u003eGemella\u003c/em\u003e (p=0.008) but depleted in \u003cem\u003eEscherichia-Shigella (\u003c/em\u003ep=0.01\u003cem\u003e), Lactobacillus\u0026nbsp;\u003c/em\u003e(p\u0026lt;0.001), \u003cem\u003eBifidobacterium\u0026nbsp;\u003c/em\u003e(p=0.01), \u003cem\u003eAcinetobacter\u0026nbsp;\u003c/em\u003e(p=0.002) relative to infants of normal ppBMI mothers. Examining longitudinal dynamics (day 30 to 90), the infant faecal microbiomes saw an increase in the abundances of \u003cem\u003eLacticaseibacillus\u0026nbsp;\u003c/em\u003e(p=0.03, BH-corrected Friedman test), and a decrease in \u003cem\u003eStreptococcus\u0026nbsp;\u003c/em\u003e(p\u0026lt;0.001), while the HM microbiome samples showed a decrease in \u003cem\u003eStaphylococcus\u003c/em\u003e (p=0.07), \u003cem\u003eGemella\u0026nbsp;\u003c/em\u003e(p=0.04) and an increase in \u003cem\u003eLactobacillus\u003c/em\u003e (p=0.05) (\u003cstrong\u003eFigure 1E\u003c/strong\u003e). Similarly, multiple HM metabolites also differed significantly between maternal ppBMIs and over time. HM samples from normal weight mothers were enriched in myo-inositol (p=0.004 BH-corrected Kruskall-Wallis test), glucose (p=0.01), and lactose (p=0.05), while five unclassified human milk oligosaccharides (HMOs) were increased in overweight-or-obese mothers (p\u0026lt;0.02) (\u003cstrong\u003eFigure 1F, Supplementary table 1\u003c/strong\u003e). Temporal changes in HM composition included increases in lactose (p\u0026lt;0.001, BH-corrected Friedman test), myo-inositol (p\u0026lt;0.001), 3-fucosyllactose (3-FL) (p\u0026lt;0.001), glucose (p\u0026lt;0.001) and glutamate (p\u0026lt;0.001), and decreases in 2\u0026rsquo;-fucosyllactose (2\u0026rsquo;-FL) (p\u0026lt;0.001) and two unclassified HMOs (p\u0026lt;0.001, p\u0026lt;0.001) (\u003cstrong\u003eFigure 1G, Supplementary table 1\u003c/strong\u003e).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 1. Maternal and infant characteristics.\u0026nbsp;\u003c/strong\u003e\u003cem\u003eContinuous data are presented as means \u0026plusmn; standard deviation. Categorical data are presented as numbers included in each category. Statistical tests: ANOVA for numeric variables and Chi-square test for categorical. BMI: body mass index, BMIz: BMI-for-age, WHZ: weight-for-length/height z-score, C-section: Caesarean section, n: number of subjects.\u003c/em\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eVariable\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003en\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eMean\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003en\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eMean\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003en\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eMean\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003en\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eMean\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eTest\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eDays after birth\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e60\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e90\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eBMI (kg/m\u0026sup2;)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e126\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e27\u0026plusmn;5.6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e145\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e27\u0026plusmn;5.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e27\u0026plusmn;5.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e148\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e27\u0026plusmn;5.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eF=0.239\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eBMI group\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e126\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e145\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e148\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026chi;\u0026sup2;=1.60\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026nbsp;Normal\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e40%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e57\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e39%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e66\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e44%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e65\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e44%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026nbsp;Overweight\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e41\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e33%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e34%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e31%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e49\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e33%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026nbsp;Obese\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e35\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e28%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e38\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e26%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e37\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e25%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e34\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e23%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eMaternal age (years)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e126\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e31\u0026plusmn;4.1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e145\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e31\u0026plusmn;4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e31\u0026plusmn;4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e148\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e31\u0026plusmn;4.1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eF=0.124\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eGestational diabetes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e125\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e144\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e147\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026chi;\u0026sup2;=0.374\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026nbsp;No\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e118\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e94%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e138\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e96%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e142\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e95%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e139\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e95%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026nbsp;Yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e6%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e4%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e5%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e5%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eSiblings\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e126\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e145\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e148\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026chi;\u0026sup2;=0.614\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026nbsp;No\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e52\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e41%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e63\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e43%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e66\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e44%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e68\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e46%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026nbsp;Yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e74\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e59%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e82\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e57%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e84\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e56%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e80\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e54%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eBirth mode\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e126\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e145\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e148\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026chi;\u0026sup2;=2.20\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026nbsp;C-section\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e5%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e9%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e9%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e9%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026nbsp;Vaginal\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e120\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e95%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e132\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e91%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e137\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e91%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e135\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e91%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eInfant sex\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e124\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e143\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e149\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e146\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026chi;\u0026sup2;=0.401\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026nbsp;Female\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e67\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e54%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e72\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e50%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e79\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e53%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e77\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e53%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026nbsp;Male\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e57\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e46%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e71\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e50%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e70\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e47%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e69\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e47%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eGestational age after birth (Days)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e123\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e282\u0026plusmn;7.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e142\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e283\u0026plusmn;7.7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e147\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e283\u0026plusmn;7.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e144\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e283\u0026plusmn;7.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eF=0.214\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eBirth weight (kg)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e122\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e3.7\u0026plusmn;0.48\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e140\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e3.7\u0026plusmn;0.45\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e146\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e3.7\u0026plusmn;0.46\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e143\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e3.7\u0026plusmn;0.46\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eF=0.081\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eSecretor status\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e126\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e145\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e148\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026chi;\u0026sup2;=0.62\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026nbsp;Non-secretor\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e34\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e27%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e34\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e23%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e35\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e23%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e24%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026nbsp;Secretor\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e92\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e73%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e111\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e77%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e115\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e77%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e112\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e76%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eLewis status\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e126\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e145\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e150\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e148\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026chi;\u0026sup2;=0.847\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026nbsp;Lewis negative\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e5%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e4%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e4%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e3%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026nbsp;Lewis positive\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e120\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e95%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e139\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e96%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e144\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e96%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e144\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e97%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eAntibiotics mother\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e122\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e145\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e149\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e146\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026chi;\u0026sup2;=81.1***\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026nbsp;No\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e121\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e99%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e94\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e65%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e134\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e90%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e136\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e93%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026nbsp;Yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e1%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e51\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e35%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e10%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e7%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eAntibiotics infant\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e142\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e147\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e145\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026chi;\u0026sup2;=2.82\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026nbsp;No\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e128\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e90%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e134\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e91%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e138\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e95%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026nbsp;Yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e14\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e10%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e9%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e5%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eBreast infection\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e125\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e145\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e149\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e146\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026chi;\u0026sup2;=25.1***\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026nbsp;No\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e87\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e70%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e103\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e71%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e130\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e87%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e128\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e88%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026nbsp;Yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e38\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e30%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e42\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e29%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e13%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e18\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e12%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eBreastfeeding issues\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e125\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e145\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e149\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e146\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026chi;\u0026sup2;=16.3**\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026nbsp;No\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e89\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e71%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e99\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e68%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e118\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e79%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e123\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e84%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026nbsp;Yes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e35\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e28%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e45\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e31%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e20%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e14%\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026Delta; BMI 3 months\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e115\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e-0.47\u0026plusmn;1.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e133\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e-0.57\u0026plusmn;1.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e142\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e-0.61\u0026plusmn;1.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e146\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e-0.61\u0026plusmn;1.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eF=0.327\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026Delta; HAZ\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e99\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0.62\u0026plusmn;1.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e113\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0.58\u0026plusmn;1.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e123\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0.61\u0026plusmn;1.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e124\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0.66\u0026plusmn;1.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eF=0.087\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026Delta; WAZ\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e100\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0.69\u0026plusmn;1.1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e114\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0.61\u0026plusmn;1.2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e124\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0.6\u0026plusmn;1.1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e125\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0.61\u0026plusmn;1.1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eF=0.132\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e\u0026Delta; WHZ\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e106\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0.66\u0026plusmn;1.4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e121\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0.65\u0026plusmn;1.4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e132\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0.61\u0026plusmn;1.4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e130\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003e0.57\u0026plusmn;1.4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"bottom\"\u003e\n \u003cp\u003eF=0.115\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003ch2\u003eMaternal ppBMI Does Not Correlate with Infant Growth\u003c/h2\u003e\n\u003cp\u003eThe relationship between maternal ppBMI and infant anthropometrics, including weight-for-length z-scores (WHZ) at 6 months and BMI-for-age z-scores (BMIz) at one, two and three years, was assessed (\u003cstrong\u003eFigure 1C, Supplementary Figure 2\u003c/strong\u003e). Infant 6-month WHZ was chosen to measure the accumulated effect of exclusive breastfeeding on infant growth trajectories. To represent early childhood weight outcomes, we included 1-year WHZ and BMI-for-age z-scores at 2 and 3 years. Surprisingly, we found no significant correlations between maternal ppBMI and infant growth outcomes (p\u0026gt;0.05; Pearson correlation test; \u003cstrong\u003eTable 2\u003c/strong\u003e). In contrast, all infant growth metrics from 6 months to 3 years showed significant correlations with each other (p\u0026le;0.05). While all correlations between growth metrics were significant, they consistently weakened as the time interval between measurements increased.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2. Pearson correlation test results between pre-pregnancy BMI and infant growth outcomes.\u0026nbsp;\u003c/strong\u003e\u003cem\u003eValues correspond to the correlation value, while stars indicate significance (*: p\u0026le;0.05, **: p\u0026le;0.01, ***: p\u0026le;0.001). ppBMI: pre-pregnancy BMI, BMIz: BMI-for-age, WHZ: weight-for-length/height z-score.\u003c/em\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e--\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ematernal ppBMI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eInfant 6-month WHZ\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eInfant 1-year WHZ\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eInfant 2-year BMIz score\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eInfant 3-year BMIz score\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ematernal ppBMI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.07\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.08\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e-0.03\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.11\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eInfant 6-month WHZ\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.63***\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.42***\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.41***\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eInfant 1-year WHZ\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.59***\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.57***\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eInfant 2-year BMIz score\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e0.57***\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eInfant 3-year BMIz score\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003ch2\u003eNPLS Models Describe Longitudinal Differences in Microbiome and Metabolome Profiles\u003c/h2\u003e\n\u003cp\u003eWhile maternal ppBMI did not appear to directly influence infant growth (\u003cstrong\u003eTable 2\u003c/strong\u003e), variations in microbiome and metabolome profiles highlighted potential indirect effects mediated through HM composition and infant gut microbiome. To address this, we applied N-way Partial Least Squares (NPLS)\u003csup\u003e28\u003c/sup\u003e to focus the analysis of the longitudinal infant faecal microbiome, HM microbiome and HM metabolome on variation that is associated with either maternal pre-pregnancy BMI or infant growth (\u003cstrong\u003eFigure 2\u003c/strong\u003e). Regressing the HM metabolome, HM microbiome and infant faecal microbiome data onto maternal ppBMI and infant growth allows for a direct examination of their specific contributions, as exploratory methods might not capture variation related to the complex hypothesized interactions in such datasets. NPLS was chosen for its ability to uncover mediating effects linking maternal ppBMI, HM composition, the infant gut microbiome and to infant growth by incorporating a dependent variable into the analysis. This was done by converting each dataset into a tensor of bacterial or metabolite profiles across the time points\u003csup\u003e29\u003c/sup\u003e and subsequently regressing them onto either maternal ppBMI or infant 6-month WHZ, resulting in six NPLS models total (\u003cstrong\u003eFigure 2A-B, Supplementary Figures 3-4\u003c/strong\u003e). A one-component model was selected for each NPLS model, as cross-validation metrics indicated that selecting additional components did not sufficiently improve explanatory power or predictive performance (\u003cstrong\u003eSupplementary Figures 5-6\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eThe NPLS models explained 2-6% of variation in their respective independent datasets (HM microbiome, HM metabolome, or infant faecal microbiome) and 9-31% of variation in their dependent variable (ppBMI or 6-month WHZ) (\u003cstrong\u003eTable 3, Supplementary Figure 5\u003c/strong\u003e). The predictive performance of each NPLS model was assessed by correlating predicted and measured values of the dependent variable. This approach revealed strong predictive performance in the NPLS models of the infant faecal microbiome, HM microbiome and HM metabolome regressed onto maternal ppBMI (R=0.35, R=0.34, R=0.56 respectively; p=5.7e-6, p=8.4e-6, p=9.2e-15, respectively; Pearson correlation test; \u003cstrong\u003eTable 3\u003c/strong\u003e). Similarly, we found strong predictive performance in the NPLS models of the infant faecal microbiome, HM microbiome and HM metabolome regressed onto infant 6-month WHZ (R=0.32, R=0.31, R=0.37, respectively; p=1.7e-4, p=2.8e-4, p=1.2e-5, respectively).\u003c/p\u003e\n\u003cp\u003eTo explore the influence of maternal ppBMI, we examined the associations between microbial and metabolite profiles across the HM metabolome, HM microbiome and infant gut microbiome (\u003cstrong\u003eFigure 2C-E, Supplementary Figures 7-9\u003c/strong\u003e). The NPLS models describing the HM metabolomics revealed that compounds associated with energy metabolism such as fumarate and lactate alongside several unclassified HMOs were associated with high ppBMI. In contrast, low ppBMI was linked to amino acid derivatives such as 2-aminobutyrate, glutamine and taurine along with simple sugars such as glucose, myo-inositol and acetone, suggesting differences in HM metabolic energy allocation (\u003cstrong\u003eFigure 2C, Supplementary Figure 9\u003c/strong\u003e). Differences \u0026nbsp;were also observed in the NPLS models describing the HM microbiome, where \u003cem\u003eStaphylococcus\u0026nbsp;\u003c/em\u003espp.\u003cem\u003e, Gemella\u003c/em\u003e spp., and \u003cem\u003eStreptococcus\u0026nbsp;\u003c/em\u003espp. were associated with high ppBMI while several unidentified \u003cem\u003eBacteroides\u0026nbsp;\u003c/em\u003espp., \u003cem\u003eClostridium\u0026nbsp;\u003c/em\u003espp\u003cem\u003e.\u003c/em\u003e along with \u003cem\u003eStaphylococcus aureus\u003c/em\u003e, \u003cem\u003eBifidobacterium longum\u003c/em\u003e and \u003cem\u003eBifidobacterium pseudocatenulatum\u003c/em\u003e were associated with low ppBMI (\u003cstrong\u003eFigure 2D, Supplementary Figure 8\u003c/strong\u003e). For the infant gut microbiome, the NPLS model revealed that several \u003cem\u003eBifidobacterium breve\u003c/em\u003e and \u003cem\u003eBifidobacterium bifidum\u003c/em\u003e zOTUs (zero-radius operational taxonomic units) were associated with high maternal ppBMI alongside \u003cem\u003eKlebsiella\u0026nbsp;\u003c/em\u003espp., \u003cem\u003eFlavonifractor\u0026nbsp;\u003c/em\u003espp., \u003cem\u003eEscherichia\u0026nbsp;\u003c/em\u003espp. and a diverse set of \u003cem\u003eFirmicutes\u003c/em\u003e members. In contrast, low maternal ppBMI was linked to \u003cem\u003eLibanicoccus\u0026nbsp;\u003c/em\u003espp\u003cem\u003e.\u003c/em\u003e, \u003cem\u003eActinomyces\u0026nbsp;\u003c/em\u003espp., \u003cem\u003eRothia mucilaginosa\u003c/em\u003e and \u003cem\u003eBifidobacterium adolescentis\u003c/em\u003e, possibly reflecting downstream effects of maternal health on infant gut colonization (\u003cstrong\u003eFigure 2E, Supplementary Figure 7\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eWe also examined the association between microbial and metabolic profiles and infant 6-month WHZ (\u003cstrong\u003eFigure 2F-H, Supplementary Figures 10-12\u003c/strong\u003e). In the NPLS model describing the HM metabolomics, multiple HMOs alongside succinate, lysine and hydroxybutyrate were associated with high infant 6-month WHZ, suggesting the importance of these compounds for infant growth optimization. In contrast, low infant 6-month WHZ was associated with the amino acids phenylalanine and methionine, alongside simple sugars such as glucose, lactose and galactose together with butyrate and \u003cem\u003esn\u003c/em\u003e-glycero-3-phosphocholine, indicating a different nutrient subset associated with slower growth trajectories (\u003cstrong\u003eFigure 2F, Supplementary Figure 12\u003c/strong\u003e). Growth-related bacterial associations were also observed in the NPLS model describing the HM microbiome, where high infant 6-month WHZ was associated with \u003cem\u003eRothia\u003c/em\u003e spp. along with \u003cem\u003eStreptococcus\u0026nbsp;\u003c/em\u003espp., \u003cem\u003eEnterococcus\u0026nbsp;\u003c/em\u003espp. and \u003cem\u003eStaphylococcus\u0026nbsp;\u003c/em\u003espp., while low infant 6-month WHZ was associated with \u003cem\u003eAcinetobacter\u0026nbsp;\u003c/em\u003espp., \u003cem\u003ePseudomonas\u0026nbsp;\u003c/em\u003espp. and \u003cem\u003eStaphylococcus\u0026nbsp;\u003c/em\u003espp., \u003cem\u003eLactobacillus\u0026nbsp;\u003c/em\u003espp\u003cem\u003e.\u003c/em\u003e, and several \u003cem\u003eBifidobacterium\u0026nbsp;\u003c/em\u003espp., suggesting microbiome-mediated growth-modulatory effects \u0026nbsp;(\u003cstrong\u003eFigure 2G, Supplementary Figure 11\u003c/strong\u003e). NPLS model describing the infant faecal microbiome showed some overlap with the ppBMI-regressed model, with high infant 6-month WHZ associating with \u003cem\u003eActinobacteriota\u003c/em\u003e members, including \u003cem\u003eActinomyces\u0026nbsp;\u003c/em\u003espp., and multiple \u003cem\u003eBifidobacterium\u0026nbsp;\u003c/em\u003espp. along with a diverse set of \u003cem\u003eFirmicutes\u003c/em\u003e zOTUs. In contrast, multiple \u003cem\u003eErysipelatoclostridium\u0026nbsp;\u003c/em\u003espp., \u003cem\u003eClostridium\u0026nbsp;\u003c/em\u003espp., \u003cem\u003eBacteroides\u0026nbsp;\u003c/em\u003espp., \u003cem\u003eKlebsiella\u0026nbsp;\u003c/em\u003espp. and \u003cem\u003eEscherichia-Shigella\u0026nbsp;\u003c/em\u003espp. were associated with low infant 6-month WHZ, suggesting a dysbiotic microbial profile potentially linked to slower infant growth (\u003cstrong\u003eFigure 2H, Supplementary Figure 10\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eTo investigate the potential long-term effects of weight-related microbial taxa and metabolites on infant growth, we analyzed the correlations between subject loadings derived from NPLS models for maternal ppBMI and 6-month WHZ with infant growth outcomes measured at 1 to 3 years. Across independent datasets, no significant correlations were observed between the infant growth outcomes and subject loadings in NPLS models that were regressed onto maternal ppBMI (p\u0026gt;0.05, BH-corrected Pearson correlation test, \u003cstrong\u003eSupplementary Figure 13\u003c/strong\u003e, \u003cstrong\u003eSupplementary Table 2\u003c/strong\u003e). These findings are consistent with the observed lack of direct correlation between maternal ppBMI and infant growth. However, the NPLS models did show good predictive performance by capturing time-resolved variation in the HM microbiome, HM metabolome and infant gut microbiome associated with either maternal ppBMI or infant 6-month WHZ (\u003cstrong\u003eTable 3\u003c/strong\u003e). This demonstrates the utility of NPLS modeling in identifying distinct patterns in microbiome and metabolome profiles over time, even when a direct correlation between maternal ppBMI and infant growth outcomes is not apparent.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 3\u003c/strong\u003e. \u003cstrong\u003ePearson correlation test results of the NPLS model prediction with the used dependent variable.\u003c/strong\u003e \u003cem\u003eR = correlation value, RMSE = root mean squared error. ppBMI: pre-pregnancy BMI, WHZ: weight-for-length/height z-score.\u0026nbsp;\u003c/em\u003e\u0026sigma;: the standard deviation of the dependent variable.\u003cem\u003e\u0026nbsp;varExpX: variation explained by the model in the independent dataset. varExpY: variation explained by the model in the dependent variable. See \u003cstrong\u003eSupplementary Figure 14\u003c/strong\u003e for the scatter plots underlying these correlation values.\u003c/em\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eIndependent dataset\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eDependent variable\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003ep-value\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eRMSE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e\u0026sigma;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003evarExpX (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003evarExpY(%)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003eInfant\u003c/p\u003e\n \u003cp\u003eFaecal\u003c/p\u003e\n \u003cp\u003eBacteria\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMaternal ppBMI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.35\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e5.7e-6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e4.90\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e5.26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e5.77\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e12.30\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eInfant 6-month WHZ\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.7e-4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.06\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.03\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2.61\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e10.34\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003eMilk\u003c/p\u003e\n \u003cp\u003eBacteria\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMaternal ppBMI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.34\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e8.4e-6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e4.87\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e5.26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e5.31\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e11.22\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eInfant 6-month WHZ\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.31\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2.8e-4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.08\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.03\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.89\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e9.43\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" valign=\"top\"\u003e\n \u003cp\u003eMilk\u003c/p\u003e\n \u003cp\u003eMetabolites\u003c/p\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eMaternal ppBMI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.56\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e9.2e-15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e4.32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e5.26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e3.06\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e31.04\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003eInfant 6-month WHZ\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e0.37\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.2e-5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.05\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e1.03\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e2.97\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\"\u003e\n \u003cp\u003e13.40\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003ch2\u003eNPLS Models Describe Microbial Subcommunities with Similar Dynamics Per Maternal ppBMI Group\u003c/h2\u003e\n\u003cp\u003eTo evaluate whether the zOTUs strongly associated with maternal ppBMI in the NPLS models constituted significant differences in overall microbial composition , we examined their combined relative abundances in the HM and infant faecal microbiomes. Bacterial zOTUs strongly associated with either low maternal ppBMI or high maternal ppBMI were clustered based on their fitted responses (\u003cstrong\u003eFigure 2D-E\u003c/strong\u003e, \u003cstrong\u003eSupplementary Table 3\u003c/strong\u003e). The summed relative abundances of these clusters were then compared between maternal ppBMI groups across all time points (\u003cstrong\u003eFigure 3\u003c/strong\u003e). In the NPLS model describing the infant faecal microbiome (\u003cstrong\u003eFigure 3A\u003c/strong\u003e), the cluster associated with high maternal ppBMI (containing \u003cem\u003eBifidobacterium bifidum, Bifidobacterium breve, Clostridium perfringens, Escherichia coli, Klebsiella variicola,\u0026nbsp;\u003c/em\u003eand\u003cem\u003e\u0026nbsp;Staphylococcus aureus\u003c/em\u003e) was enriched in samples from infants with overweight-or-obese mothers compared to samples from infants with normal weight mothers at all time points (p=0.016, p\u0026lt;1e-3, p=0.0015, respectively; BH-corrected permutation test of mean difference, n=999; \u003cstrong\u003eSupplementary Table 4\u003c/strong\u003e). The cluster associated with low maternal ppBMI (containing \u003cem\u003eBacteroides fragilis, Bifidobacterium adolescentis, Citrobacter freundii, Enterococcus faecalis, Lactobacillus paracasei, Parabacteroides distasonis,\u0026nbsp;\u003c/em\u003eand \u003cem\u003eRothia mucilaginosa\u003c/em\u003e) was enriched in samples from infants with normal weight mothers compared to samples from infants with overweight-or-obese mothers at days 60 and 90 (p=0.0015, p\u0026lt;1e-3, respectively).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSimilarly, in the NPLS model describing the HM microbiome (\u003cstrong\u003eFigure 3B\u003c/strong\u003e), the cluster associated with high maternal ppBMI (containing \u003cem\u003eRothia mucilaginosa, Staphylococcus epidermidis,\u0026nbsp;\u003c/em\u003eand \u003cem\u003eStreptococcus oralis\u003c/em\u003e) was enriched in samples from overweight-or-obese mothers compared to samples from normal weight mothers at all time points (p\u0026lt;1e-3, p=0.0053, p\u0026lt;1e-3, p\u0026lt;1e-3, respectively; \u003cstrong\u003eSupplementary Table 4\u003c/strong\u003e). Conversely, the cluster associated with low maternal ppBMI (containing \u003cem\u003eBifidobacterium longum, Bifidobacterium pseudocatenulatum, Cutibacterium acnes, Lactobacillus gasseri, Ligilactobacillus murinus, Staphylococcus aureus,\u0026nbsp;\u003c/em\u003eand \u003cem\u003eStenotrophomonas maltophilia\u003c/em\u003e) was enriched in samples from normal weight mothers compared to those from overweight-or-obese mothers at days 3 and 90 (p=0.0032, p=0.0032, respectively).\u003c/p\u003e\n\u003ch2\u003eComparisons of Microbiome Loadings Reveal Shared Bacterial Signatures Between Human Milk and Infant Gut\u003c/h2\u003e\n\u003cp\u003eAlthough maternal ppBMI and infant 6-month WHZ were not correlated, many bacteria and metabolites were associated with both metrics in the NPLS models (\u003cstrong\u003eFigure 2\u003c/strong\u003e). To explore these relationships further, we performed pairwise comparisons of the loadings for each bacterial zOTU and metabolite between NPLS models regressed onto maternal ppBMI and infant 6-month WHZ (\u003cstrong\u003eFigure 4A\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003eThe NPLS model describing the HM metabolome regressed onto maternal ppBMI was compared to the model of the same data regressed onto infant 6-month WHZ. A large proportion of metabolites had similar associations to both dependent variables (\u003cstrong\u003eFigure 4B\u003c/strong\u003e). HMOs were consistently associated with high maternal ppBMI and high infant 6-month WHZ along with energy metabolism-related compounds and fucose as the sole sugar. Conversely, compounds associated with low maternal ppBMI and low infant 6-month WHZ included sugars, fatty acids, amino acids, and their derivatives. Several metabolites were uniquely associated with one metric but showed neutral or modest associations with the other. Most notably, high infant 6-month WHZ was associated with \u0026nbsp;2-hydroxybutyrate, fucose, lysine, fucose, LNFPH I, and LNDFH I, while phenylalanine, butyrate, galactose and lactose were associated with with low infant 6-month WHZ. None of these metabolites were strongly associated with ppBMI. Conversely, 2-amminobutyrate, ethanolamine, taurine, myo-inositol, and threonine were associated with low ppBMI, but had neutral associations with infant 6-month WHZ. A few compounds displayed inconsistent associations; for example, caffeine was associated with high ppBMI but also with low infant 6-month WHZ while acetone showed opposite associations. No metabolites were uniquely associated with high ppBMI.\u003c/p\u003e\n\u003cp\u003eIn the NPLS models describing the HM microbiome, most zOTUs had similar associations for both maternal ppBMI and infant 6-month WHZ (\u003cstrong\u003eFigure 4C\u003c/strong\u003e). \u003cem\u003eStreptococcus, Veillonella,\u0026nbsp;\u003c/em\u003eand \u003cem\u003eGemella\u003c/em\u003e zOTUs were associated with high maternal ppBMI and high infant 6-month WHZ, while \u003cem\u003eAcinetobacter\u003c/em\u003e, \u003cem\u003eStaphylococcus\u003c/em\u003e and \u003cem\u003ePseudomonas\u003c/em\u003e zOTUs were associated with both low maternal ppBMI and low infant 6-month WHZ. Notably, some zOTUs had opposite associations, with a \u003cem\u003eClostridium\u0026nbsp;\u003c/em\u003esp. and \u003cem\u003eLactobacillus\u0026nbsp;\u003c/em\u003espp. being associated with low ppBMI but high infant 6-month WHZ, while \u003cem\u003eBifidobacterium animalis\u003c/em\u003e, \u003cem\u003eBacteroides\u0026nbsp;\u003c/em\u003esp., and \u003cem\u003eAcidovorax\u0026nbsp;\u003c/em\u003esp. were associated with high ppBMI but low infant 6-month WHZ. Several zOTUs had distinct associations, with \u003cem\u003eRothia\u0026nbsp;\u003c/em\u003esp. and \u003cem\u003eEcherichia-Shigella\u003c/em\u003e sp. associated to high ppBMI but were neutral in their association to infant 6-month WHZ. Conversely, \u003cem\u003eEnterococcus faecalis\u003c/em\u003e, \u003cem\u003eStaphylococcus aureus\u003c/em\u003e and \u003cem\u003eStreptococcus\u0026nbsp;\u003c/em\u003esp. were associated to high infant 6-month WHZ, but had neutral loadings in the ppBMI model.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFor the infant faecal microbiome, our approach revealed multiple \u003cem\u003eBifidobacterium\u0026nbsp;\u003c/em\u003espp\u003cem\u003e.\u003c/em\u003e and an \u003cem\u003eEnterococcus\u0026nbsp;\u003c/em\u003esp. that were associated with high maternal ppBMI as well as a high infant 6-month WHZ (\u003cstrong\u003eFigure 4D\u003c/strong\u003e). Several \u003cem\u003eBifidobacterium adolescentis\u003c/em\u003e zOTUs contrasted by associating with low maternal ppBMI and a low infant 6-month WHZ, along with \u003cem\u003eBacteriodes\u0026nbsp;\u003c/em\u003espp., \u003cem\u003eKlebiella\u0026nbsp;\u003c/em\u003espp., and \u003cem\u003eLibanicoccus\u0026nbsp;\u003c/em\u003esp. While most bifidobacteria, other than \u003cem\u003eB. adoliscentis\u003c/em\u003e, were associated with high WHZ, several had neutral-to-moderate associations to ppBMI. zOTUs associated with both low maternal ppBMI and low infant 6-month WHZ included \u003cem\u003eBacteroides fragilis,\u0026nbsp;\u003c/em\u003eand \u003cem\u003eKlebsiella\u0026nbsp;\u003c/em\u003espp\u003cem\u003e.,\u0026nbsp;\u003c/em\u003ewhile no zOTUs had strong associations to low ppBMI and high 6-month WHZ. Notable exclusive associations for infant 6-month WHZ were \u003cem\u003eActinomyces\u003c/em\u003e associated to high WHZ, and \u003cem\u003eKlebsiella\u003c/em\u003e \u003cem\u003evariicola\u003c/em\u003e, \u003cem\u003eErysipelatoclostridium\u0026nbsp;\u003c/em\u003espp\u003cem\u003e.\u003c/em\u003e, and \u003cem\u003eParabacteroides\u0026nbsp;\u003c/em\u003eto low WHZ. Low ppBMI had distinct associations for \u003cem\u003eStreptococcus\u0026nbsp;\u003c/em\u003espp. and high ppBMI for \u003cem\u003eKlebsiella variicola\u003c/em\u003e.\u003c/p\u003e\n\u003cp\u003eTo identify the impact of microbial transfer from HM to infant gut, bacterial associations in the NPLS models, for the HM and infant faecal microbiomes regressed onto maternal ppBMI were compared (\u003cstrong\u003eSupplementary Figure 15\u003c/strong\u003e). \u003cem\u003eB. longum\u003c/em\u003e and \u003cem\u003eB. pseudocatenulatum\u003c/em\u003e zOTUs were associated with low ppBMI in both models, while \u003cem\u003eE. coli\u003c/em\u003e was associated to high ppBMI only in the NPLS model describing the infant faecal microbiome. \u003cem\u003eVeillonella\u003c/em\u003e spp. and some \u003cem\u003eStreptococcus\u003c/em\u003e spp. were moderately associated with high ppBMI in both compartments. One \u003cem\u003eStreptococcus\u003c/em\u003e sp. and \u003cem\u003eRothia mucilaginosa\u003c/em\u003e were associated with high ppBMI in the NPLS model describing the HM microbiome, but were associated with low ppBMI \u0026nbsp;in the model describing the infant faecal microbiome, while a \u003cem\u003eBlautia\u003c/em\u003e sp. showed the opposite. Comparing 6-month infant WHZ-regressed NPLS models of the infant faecal microbiome and the HM microbiome (\u003cstrong\u003eSupplementary figure 15\u003c/strong\u003e), \u003cem\u003eL. gasseri\u003c/em\u003e was associated with high infant 6-month WHZ in both compartments. \u003cem\u003eS. aureus\u003c/em\u003e and \u003cem\u003eE. coli\u003c/em\u003e were linked to low infant 6-month WHZ, while \u003cem\u003eS. epidermidis\u003c/em\u003e showed the opposite. A \u003cem\u003eVeillonella\u003c/em\u003e sp. was associated with high infant 6-month WHZ only in the HM microbiome. In summary this shows that while the metabolite and microbiome associations for ppBMI and 6-month WHZ are shared, many bacteria have strong associations to only one of the two, or have contrasting associations.\u0026nbsp;\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis study reveals that maternal pre-pregnancy BMI (ppBMI) influences human milk (HM) composition and infant gut microbiome development, although ppBMI does not directly correlate with infant growth metrics. Using N-way Partial Least Squares (NPLS) models to focus the analysis of the longitudinal variation in the infant faecal microbiome, HM microbiome, and HM metabolome, we identified microbial taxa and metabolite signatures linked to maternal ppBMI and early infant growth. By comparing the associations of the modelled microbiota and metabolites with maternal ppBMI and infant 6-month WHZ in their respective models, we found that some features were associated with both metrics while others were associated with only one. These findings shed light on the complex interactions between maternal health, HM composition, and infant gut microbiome development during breastfeeding and their links to infant growth.\u003c/p\u003e\n\u003cp\u003eThe lack of a clear link between maternal ppBMI and 6-month infant WHZ was unexpected, given the shared feature loadings between the NPLS models (Fig.\u0026nbsp;4). A possible explanation is that the effect of maternal ppBMI is neutralized by the other compartments of the system, such as the infant gut microbiome or the infant immune system. Additionally, features with strong associations with only one metric indicate independent effects within the system. Unmeasured variables, such as unrecorded maternal diet, breastfeeding frequency and rate of transition to solid food, intrauterine fetal growth trajectories, genetics, environmental factors, or unmeasured dataset properties (e.g., HM lipids, proteins, or species-level or genetic microbiome differences), may also confound the relationship between maternal overweight and infant growth. This was also reflected by the bacterial subcommunities that were associated with high and low ppBMI, consisting of a mixture of beneficial and dysbiosis-related zOTUs. The homogeneity of our cohort in terms of all exclusively breastfeeding should also be considered, as exclusive breastfeeding has been shown to mitigate the transmission of maternal obesity\u003csup\u003e14,30\u003c/sup\u003e. It is also possible that a high maternal ppBMI associated infant faecal microbiome signal is masked during the exclusive breastfeeding period but could reappear later in life. The HMO profile of the HM could also be more directly influenced by maternal dietary uptake, rather than by the overweight phenotype\u003csup\u003e31\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eThe association of multiple HMOs and \u003cem\u003eBifidobacterium\u003c/em\u003e spp. with high maternal ppBMI is an unexpected result, as both are considered to be highly beneficial to early-life gut function\u003csup\u003e20,32\u0026ndash;36\u003c/sup\u003e,while high maternal ppBMI have been associated with the infant also being in risk of excessive weight gain (REFS). Their enrichment may reflect metabolic trade-offs driven by altered HMOs that promote \u003cem\u003eBifidobacterium\u003c/em\u003e colonization. This could mean that high ppBMI-associated HMOs and \u003cem\u003eBifidobacterium\u003c/em\u003e spp. counteract the effects of other detrimental microbiota members or compounds linked to maternal obesity, thereby counteracting the inter-generational transmission of the phenotype. In contrast, \u003cem\u003eE. coli\u003c/em\u003e and \u003cem\u003eKlebsiella\u003c/em\u003e spp. were associated with low infant 6-month WHZ, likely reflecting dysbiosis or inflammation that impairs growth\u003csup\u003e37\u0026ndash;40\u003c/sup\u003e. Evidence also links low maternal ppBMI and low infant 6-month WHZ to typically pathogenic species\u003csup\u003e41,42\u003c/sup\u003e, including \u003cem\u003eCutibacterium acnes\u003c/em\u003e, \u003cem\u003ePseudomonas\u003c/em\u003e spp. and \u003cem\u003eStaphylococcus aureus\u003c/em\u003e, as well as \u003cem\u003eAcinetobacter\u003c/em\u003e spp., which has been associated with postpartum stress\u003csup\u003e43\u003c/sup\u003e. These findings highlight potential pathways through which maternal obesity might influence the infant microbiome development, both via changes in HM composition and via indirect effects mediated by microbiota shifts. For example, the enrichment of \u003cem\u003eBifidobacterium\u003c/em\u003e spp. may represent a protective mechanism against adverse effects of maternal obesity, while the depletion of \u003cem\u003eKlebsiella\u003c/em\u003e spp. and \u003cem\u003eE. coli\u003c/em\u003e may indicate disrupted microbial dynamics. Future studies integrating functional assays are needed to disentangle these interactions and implications for targeted nutritional interventions.\u003c/p\u003e\n\u003cp\u003eWe have observed that the NPLS models describe 2\u0026ndash;6% of the variation in the independent dataset and 9\u0026ndash;31% of the variation in the dependent variable. While this may seem limited, they demonstrate that biological variation of interest is often hidden underneath other sources of variability. Additionally, variation of interest in microbiome data tends to be very low in general, and there is very little variation in the milk metabolomics data due to its composition being the same across time. Despite these issues, we have shown that these models are predictive, and that the features we identified are known to play a role in early infant nutrition and gut function. We have shown that highly focused data analysis methods are needed to find weak signals in such datasets. In this work we have analysed each independent dataset separately. Other multi-way data integration approaches like Advanced Coupled Matrix and Tensor Factorization (ACMTF) could also have been used to better identify shared and distinct sources of variation\u003csup\u003e44\u0026ndash;46\u003c/sup\u003e. However, ACMTF solutions cannot be steered towards a specific solution by maximising covariation using a dependent variable. Future work should focus on adapting the ACMTF algorithm to include a joint regression step using a dependent variable of interest.\u003c/p\u003e\n\u003cp\u003eIn conclusion, we identified microbial subcommunities linked to maternal ppBMI and infant growth, unexpectedly finding that multiple HMOs and bifidobacteria associated with high maternal ppBMI and high infant 6 month WHZ scores. This highlights pathways through which maternal obesity could affect infant microbiome development, by a combination of changes in HM composition and secondary effects mediated by microbiota alterations. While no direct correlation between maternal ppBMI and infant growth was observed within the first three years of life, the application of NPLS modeling proved to be a key methodological approach to identifying shared and distinct microbial and metabolite signatures associated with each metric. These findings not only advance our understanding of maternal-infant nutritional dynamics but also underscore the utility of NPLS as a method for exploring longitudinal relationships in a multi-omics context.\u003c/p\u003e\n\u003cp\u003eLimitations\u003c/p\u003e\n\u003cp\u003eWhile this study focuses on a relatively homogenous cohort of healthy, vaginally delivered, exclusively breastfed infants, the findings may not generalize to populations with differing delivery modes, feeding practices or maternal health profiles. For example, caesarean delivery, formula feeding and antibiotics use are known to influence infant gut microbiome development and may interact with maternal ppBMI in different ways\u003csup\u003e47,48\u003c/sup\u003e. Future research should include diverse cohorts to evaluate whether the observed microbiota and metabolite associations are consistent across these contexts. Unmeasured variables such as maternal diet, breastfeeding duration, and environmental exposures may have influenced the observed relationships. For instance, dietary intake is known to affect HMO profiles, potentially confounding associations attributed to maternal ppBMI. Single-point weight measurements also have a degree of uncertainty, due to the coMaternal obesity is a key risk factor for excessive fetal growth and childhood obesity, yet its influence on human milk (HM) composition and the infant gut microbiome development remains unclear. This considerable variation in day-to-day weight of infants, combined with intermittent periods of weight-gain and growth-spurts\u003csup\u003e49,50\u003c/sup\u003e. Furthermore, offspring that is large for gestational age at delivery may display a growth pattern with diminished weight accretion in the first weeks or months of life, and newborns of small size at delivery may have an accelerated weight gain\u003csup\u003e51\u003c/sup\u003e. Such \u0026ldquo;catch-up\u0026rdquo; or \u0026ldquo;slow-down\u0026rdquo; effects could obscure the associations investigated in the present study, even though the study population intendedly had a homogeneous construction. Follow-up studies measuring growth in later life may shed light on such issues.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003eParticipants and Sample Collection\u003c/p\u003e\n\u003cp\u003eInfants and their mothers who provided samples for this study were recruited as part of the MaInHealth (Maternal Infant Health) cohort established in Aarhus, Denmark, to investigate the natural human milk (HM) variation and its possible effects on offspring metabolism and gut microbiota\u003csup\u003e23\u003c/sup\u003e. The cohort has been approved by the Central Denmark Regional Committees on Health Research Ethics (ethical approval reference: 1-10-72-296-18) and registered on ClinicalTrials. gov (identification number: NCT05111990). Pregnant women were recruited from Aarhus University Hospital, Aarhus, Denmark, from 2019 to 2021. Women included in the study were healthy, non-smokers, expecting to give birth vaginally, and planning to breastfeed for the first four to six months. Pre-pregnancy weight and height were self-reported by the mothers upon recruitment at gestational week 18\u0026ndash;20. Infants included were healthy, with birth weights of 2500-5000g and were born full-term, i.e. gestational week 37\u003csup\u003e+\u0026thinsp;0\u003c/sup\u003e or later. See the study protocol for a detailed project description and recruitment and exclusion criteria\u003csup\u003e23\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eBriefly, the mothers collected milk during the first week postpartum and at one, two, and three months after birth. Approximately 20 mL of foremilk was taken during each collection, avoiding the initial drops. Fecal samples (~\u0026thinsp;2 g) were collected at the same time points from the first faeces passed after HM collection. These samples were stored in the participants\u0026apos; freezers at -20\u0026deg;C for up to two weeks, then transported on dry ice to Aarhus University, where they were stored at -80\u0026deg;C until analysis. HM samples were thawed, mixed, divided into 1 mL aliquots, and returned to -80\u0026deg;C storage. One aliquot was used for metabolomics and another for microbiome characterization. Faecal samples were thawed, and 250 mg was mixed with PBS buffer at a 1:5 ratio (w/v), vortexed, and centrifuged at 10,000 \u0026times; g for 10 minutes at 4\u0026deg;C. The pellet was frozen at -80\u0026deg;C for DNA extraction, while the supernatant was saved for metabolomics analysis.\u003c/p\u003e\n\u003cp\u003eBMI was calculated as (weight in kg/(height in m)\u003csup\u003e2\u003c/sup\u003e). Maternal pre-pregnancy BMI was stratified into three groups: normal weight (NW; BMI 25 or lower), overweight (OW; BMI 25\u0026ndash;30) and obesity (OB; BMI 30 or higher) (\u003cstrong\u003eSupplementary Fig.\u0026nbsp;1\u003c/strong\u003e). Infant anthropometric z-scores indicate standard deviations from the mean height-for-age, weight-for-age, weight-for-height, and BMI-for-age referenced to the 2006 World Health Organization child growth standards, calculated in R v4.2.1\u003csup\u003e52\u003c/sup\u003e using the \u0026ldquo;zscorer\u0026rdquo; package v0.3.1\u003csup\u003e53\u003c/sup\u003e. Infant anthropometrics were recorded at birth and 1, 2, 3, 6, 12, 24 and 36 months after birth (\u003cstrong\u003eSupplementary Fig.\u0026nbsp;2\u003c/strong\u003e).\u003c/p\u003e\n\u003cp\u003e\u003csup\u003e\u0026nbsp;1\u0026nbsp;\u003c/sup\u003eH Nuclear Magnetic Resonance Spectroscopy Metabolomics Analysis of Milk\u003c/p\u003e\n\u003cp\u003eHM samples for \u003csup\u003e1\u003c/sup\u003eH nuclear magnetic resonance spectroscopy (NMR)-based metabolomics were processed following a standard protocol for milk-based metabolomics as described previously\u003csup\u003e54\u003c/sup\u003e. Thawed samples were centrifuged to remove the fat layer, filtered, and placed in NMR tubes with deuterated water (D₂O) and 3-(trimethylsilyl) propionic acid (TSP) for referencing. \u003csup\u003e1\u003c/sup\u003eH NMR spectra acquisition was performed using a Bruker NEO 600 spectrometer at 300 K and a \u003csup\u003e1\u003c/sup\u003eH frequency of 600.03 MHz. Spectra were referenced to the TSP signal at 0 ppm and processed using Topspin 4.09 software for phase and baseline correction. Metabolites were identified using the Chenomx NMR suite 10.1 (Chenomx Inc., Edmonton, AB, Canada) and normalised to lactose concentration. A weighted normalisation approach was applied to account for lactose level variation, combining lactose and total metabolite factors. Metabolite concentrations were centred for analysis but not scaled.\u003c/p\u003e\n\u003cp\u003eMetabolite identification was performed using the Chenomx NMR Suite 10.1 and compared with the Chenomx standard metabolite library, supplemented with an in-house HMO library. Lactose constitutes\u0026thinsp;\u0026gt;\u0026thinsp;80% of the total concentration of the HM metabolites detected by NMR-based metabolomics and significantly varies between stages of lactation\u003csup\u003e55\u003c/sup\u003e. Unidentified or partially identified compounds were included in the analysis and normalisation, as their area under-curve still allowed quantification. A weighted normalisation with lactose weight 3 and total metabolites weight 1 was therefore used: a lactose factor normalising to the average concentration in HM of 215 mM (L-factor) and a total metabolite (excluding lactose) across all samples (T-factor) were combined as the factor (L-factor\u0026times;3\u0026thinsp;+\u0026thinsp;T-factor)/4. Metabolite concentrations were centred but not scaled before analysis.\u003c/p\u003e\n\u003cp\u003eSequencing of bacterial communities in mother\u0026rsquo;s milk and infant faecal samples\u003c/p\u003e\n\u003ch3\u003eDNA Extraction and preparation for sequencing\u003c/h3\u003e\n\u003cp\u003eSample processing and preparation for sequencing were performed as described previously\u003csup\u003e24\u003c/sup\u003e. Briefly, HM samples were thawed at 4\u0026deg;C and centrifuged at 12,000 \u0026times; g for 20 minutes at 4\u0026deg;C to remove the fat layer with a sterile cotton swab. DNA was extracted from the remaining pellet using the Bead-Beat Micro AX Gravity Kit (A\u0026amp;A Biotechnology, Gdynia, Poland) according to the manufacturer\u0026rsquo;s protocol. Negative controls with sterile MilliQ water were used during extraction, PCR, and sequencing. A 16S rRNA gene amplicon library was constructed by amplifying the 16S rRNA gene using primers targeting the 16S V1-V9 regions, as described previously\u003csup\u003e56\u003c/sup\u003e. In brief, amplicons were purified with SpeedBeads\u0026trade; magnetic carboxylate (Sigma-Aldrich) and verified (~\u0026thinsp;1500 bp) via agarose gel electrophoresis. Sequencing libraries were prepared by pooling barcoded PCR products from up to 196 samples and sequenced using Oxford Nanopore Technologies\u0026rsquo; GridION X5.\u003c/p\u003e\n\u003ch3\u003ePreprocessing of raw reads\u003c/h3\u003e\n\u003cp\u003eRaw reads were processed with the Long Amplicon Consensus Analysis (LACA) pipeline for de-novo clustering and taxonomic classification using the SILVA v138.1 database\u003csup\u003e57\u003c/sup\u003e. zOTU (zero-radius operational taxonomic units) consensus sequences were obtained using a multiple de-novo clustering approach. zOTUs present in fewer than 5% of samples and with mean relative abundances below 0.05% were filtered out, retaining 98% of the total reads.\u003c/p\u003e\n\u003cp\u003eProcessing of microbiome data\u003c/p\u003e\n\u003cp\u003eThe zOTU data and taxonomic information of the infant faecal microbiome and mother milk microbiome were processed in MATLAB (version 2023a). For the infant faecal microbiome data, features were kept if they had\u0026thinsp;\u0026le;\u0026thinsp;75% sparsity across the dataset. This step resulted in 93 out of 565 features being selected. For the mother milk microbiome data, features were kept if they had\u0026thinsp;\u0026le;\u0026thinsp;85% sparsity across the dataset. This step resulted in 115 out of 707 features being selected. Sparsity filtering was purposefully done across all samples, as opposed to groupwise, to keep the processing consistent between the maternal ppBMI and infant 6-month WHZ-based models.\u003c/p\u003e\n\u003cp\u003eWe performed a centred log-ratio transformation with a pseudo-count of 1 to correct for compositionality\u003csup\u003e58,59\u003c/sup\u003e. Subsequently the data was converted to a three-way array, keeping missing measurements as a row of NAs. In the infant faecal microbiome data cube, subject 332 was removed due to being an outlier, resulting in a cube of size 132 subjects x 93 microbial abundances x 3 time points. This was not necessary for the mother milk microbiome data, yielding a three-way array of size 169 subjects x 115 microbial abundances x 4 time points. The data was then centred across the subject mode to make the samples comparable per time point and scaled within the feature mode to make all microbial abundances equally important for the modelling procedure\u003csup\u003e29,60\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eProcessing of metabolomics data\u003c/p\u003e\n\u003cp\u003eThe mother milk metabolomics data were processed using MATLAB (version 2023a). Values below the detection limit were imputed with a random value between 0 and the detection limit per metabolite to preserve their distribution. Next, the dataset was (natural) log transformed to stabilise the variance. The dataset was then converted to a three-way array of size 164 subjects x 80 metabolites x 4 time points. Subsequently the data was then centred across the subject mode to make the samples comparable per time point and scaled within the feature mode to make all metabolites equally important for the modelling procedure\u003csup\u003e60\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003eNotation and Definitions\u003c/p\u003e\n\u003cp\u003eWe briefly define the mathematical notation that will be used throughout this paper. Scalars are indicated by lower-case italics such as \u003cem\u003ea\u003c/em\u003e. Vectors are indicated by bold lower-case characters such as \u003cstrong\u003eb\u003c/strong\u003e. Two-way matrices are indicated with bold capitalised characters such as \u003cstrong\u003eX\u003c/strong\u003e. Underlined bold capitalised characters are used for three-way arrays such as \u003cimg src=\"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAwAAAANCAYAAACdKY9CAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsMAAA7DAcdvqGQAAACQSURBVChTjVDBDQMxCHNmYRhmyShMcoNkDmZxH1cik/RRSzwwMsaAX8w5CYDuXtSGmRHAWzoo8nmezbn75jOzCyKCAGhmpLgC4FqLJLuA4qKb1fESlEtVRLT5JcjM5nLiEuj2ytLm2ujdv+6nCvQjmbn70wU8gtZGzdK+pGecH1HXwuAb9G8MkhxjnPyF2vsBRXl8AVkOLjUAAAAASUVORK5CYII=\" height=\"13\" width=\"12\"\u003e. The letters \u003cem\u003eI\u003c/em\u003e, \u003cem\u003eJ\u003c/em\u003e, and \u003cem\u003eK\u003c/em\u003e are reserved to indicate the dimension of the subject, microbial or metabolite abundance, and time mode, respectively. Hence the element for subject \u003cem\u003ei\u003c/em\u003e, microbial abundance \u003cem\u003ej\u003c/em\u003e at time point \u003cem\u003ek\u003c/em\u003e of a three-way array \u003cimg src=\"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAwAAAANCAYAAACdKY9CAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsMAAA7DAcdvqGQAAACQSURBVChTjVDBDQMxCHNmYRhmyShMcoNkDmZxH1cik/RRSzwwMsaAX8w5CYDuXtSGmRHAWzoo8nmezbn75jOzCyKCAGhmpLgC4FqLJLuA4qKb1fESlEtVRLT5JcjM5nLiEuj2ytLm2ujdv+6nCvQjmbn70wU8gtZGzdK+pGecH1HXwuAb9G8MkhxjnPyF2vsBRXl8AVkOLjUAAAAASUVORK5CYII=\" height=\"13\" width=\"12\"\u003e\u0026nbsp;is called \\(\\:{x}_{ijk}\\). We create the data cubes in this study such that the subjects are in the first mode, the measured microbial abundances or metabolites are in the second mode, and time points are in the third mode. We will not distinguish between the terms \u0026lsquo;factor\u0026rsquo; and \u0026lsquo;component\u0026rsquo;, nor between the terms \u0026lsquo;way\u0026rsquo; and \u0026lsquo;mode\u0026rsquo;. Generally, we will use the words \u0026lsquo;component\u0026rsquo; and \u0026lsquo;mode\u0026rsquo; throughout this paper.\u003c/p\u003e\n\u003cp\u003eN-way Partial Least Squares (NPLS)\u003c/p\u003e\n\u003cp\u003eDetails on the creation of NPLS models for various types of data have been described elsewhere\u003csup\u003e28\u003c/sup\u003e. The aim of NPLS is to find a model of the input data \u003cimg src=\"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAwAAAANCAYAAACdKY9CAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsMAAA7DAcdvqGQAAACQSURBVChTjVDBDQMxCHNmYRhmyShMcoNkDmZxH1cik/RRSzwwMsaAX8w5CYDuXtSGmRHAWzoo8nmezbn75jOzCyKCAGhmpLgC4FqLJLuA4qKb1fESlEtVRLT5JcjM5nLiEuj2ytLm2ujdv+6nCvQjmbn70wU8gtZGzdK+pGecH1HXwuAb9G8MkhxjnPyF2vsBRXl8AVkOLjUAAAAASUVORK5CYII=\" height=\"13\" width=\"12\"\u003e\u003c/p\u003e\n\u003cdiv id=\"Equa\"\u003e\n \u003cdiv id=\"FileID_Equa\" name=\"EquationSource\"\u003e$$\\:{x}_{ijk}={t}_{i}{{w}^{J}}_{j}{{w}^{K}}_{k}$$\u003c/div\u003e\n\u003c/div\u003e\n\u003cp\u003eWhere \\(\\:{x}_{ijk}\\) is the element in \u003cimg src=\"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAwAAAANCAYAAACdKY9CAAAAAXNSR0IArs4c6QAAAARnQU1BAACxjwv8YQUAAAAJcEhZcwAADsMAAA7DAcdvqGQAAACQSURBVChTjVDBDQMxCHNmYRhmyShMcoNkDmZxH1cik/RRSzwwMsaAX8w5CYDuXtSGmRHAWzoo8nmezbn75jOzCyKCAGhmpLgC4FqLJLuA4qKb1fESlEtVRLT5JcjM5nLiEuj2ytLm2ujdv+6nCvQjmbn70wU8gtZGzdK+pGecH1HXwuAb9G8MkhxjnPyF2vsBRXl8AVkOLjUAAAAASUVORK5CYII=\" height=\"13\" width=\"12\"\u003e corresponding to the \u003cem\u003ei\u003c/em\u003e-th subject, \u003cem\u003ej\u003c/em\u003e-th microbial or metabolite abundance and \u003cem\u003ek\u003c/em\u003e-th timepoint, \u003cem\u003et\u003c/em\u003e contains the subject scores, \\(\\:{w}^{J}\\) contains the feature loadings and \\(\\:{w}^{K}\\) contains the time loadings such that the covariance of \u003cem\u003et\u003c/em\u003e and a centred output variable \u003cstrong\u003ey\u003c/strong\u003e is maximised and the sum of squares of the residuals is minimised. The NPLS implementation from the N-way toolbox (version 1.8.0.0, https://nl.mathworks.com/matlabcentral/fileexchange/1088-the-n-way-toolbox) in MATLAB (version 2023a) was used to create NPLS models for all datasets with maternal ppBMI or infant 6-month WHZ as dependent variables\u003csup\u003e61\u003c/sup\u003e. Regressing to changes in WHZ (deltaWHZ) and BMIz score (deltaBMIz) at 6 months, i.e., whether the infant is growing faster or slower than the median growth curve, did not produce biologically interpretable models. Similar to Parallel Factor Analysis\u003csup\u003e28,29\u003c/sup\u003e, the correct number of components of the NPLS model needed to be determined to create an optimal model. This was done by inspecting the variation explained in the input data as well as the root mean-squared error of prediction (RMSEP) of the dependent variable (\u003cstrong\u003eSupplementary Figs.\u0026nbsp;5\u0026ndash;6\u003c/strong\u003e). For all models shown in this paper, a 1-component model was sufficient. While the RMSEP metric suggests a two-component model for the HM metabolome data regressed onto maternal ppBMI, we found that the extra component did not describe any relevant biological variation. All plots and subsequent analyses were created with the ggplot2\u003csup\u003e62\u003c/sup\u003e package in R\u003csup\u003e63\u003c/sup\u003e (ggplot2 v3.5.1, R v4.4.1). All generated NPLS models reported in the main text are shown in \u003cstrong\u003eSupplementary Figs.\u0026nbsp;3 and 4\u003c/strong\u003e. The predictive power of all NPLS models described in the main text is reported in Table 3. Correlations of the subject scores of the NPLS models with various metadata of interest are reported in \u003cstrong\u003eSupplementary Tables\u0026nbsp;2\u003c/strong\u003e and \u003cstrong\u003eSupplementary Fig.\u0026nbsp;13\u003c/strong\u003e. Correlations of the fitted features with the dependent variable were used to determine associations with maternal ppBMI and infant 6-month WHZ.\u003c/p\u003e\n\u003cp\u003ePost-hoc clustering of NPLS loadings\u003c/p\u003e\n\u003cp\u003eClustering of the feature loadings can help identify microbial subcommunities that are connected to a particular time profile in the NPLS model. We performed the clustering procedure of the microbiota loadings as follows. Microbiota were not considered if the variation explained by the model for their entire time trajectory was lower than that of the model itself. The remaining microbiota were then clustered based on their modelled time trajectories using the K-medoids algorithm from the cluster R package\u003csup\u003e64\u003c/sup\u003e (v2.1.6) with 50 random starts to be robust against outliers. The number of clusters was determined using the within-cluster sum of squares, silhouette width and gap statistic metrics as reported by the factoextra R package\u003csup\u003e65\u003c/sup\u003e (v1.0.7; \u003cstrong\u003eSupplementary Figs.\u0026nbsp;16\u0026ndash;19\u003c/strong\u003e). Relative abundances per cluster were summed per subject, after which the mean and standard error of the mean were calculated per maternal ppBMI group (Fig. 3). Relative abundance sum permutation test results are reported in \u003cstrong\u003eSupplementary Table\u0026nbsp;4\u003c/strong\u003e. Cluster members are reported in \u003cstrong\u003eSupplementary Tables\u0026nbsp;3\u003c/strong\u003e. A similar analysis was performed to identify microbial subcommunities connected to infant 6-month WHZ groups are reported in \u003cstrong\u003eSupplementary Fig.\u0026nbsp;20\u003c/strong\u003e and \u003cstrong\u003eSupplementary Table\u0026nbsp;3\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003ePairwise comparisons of NPLS feature loadings\u003c/p\u003e\n\u003cp\u003eFor each independent dataset (HM microbiome, HM metabolome, and infant faecal microbiome), feature loadings from the NPLS model regressed onto maternal ppBMI were compared with those from the model regressed onto infant 6-month WHZ. Scatterplots were generated with feature loadings from the ppBMI-associated model on the x-axis and those from the WHZ-associated model on the y-axis, enabling identification of shared and distinct associations. Visualizations and annotations were created using the ggplot2\u003csup\u003e66\u003c/sup\u003e package in R\u003csup\u003e63\u003c/sup\u003e (ggplot2 v3.5.1, R v4.4.1). A full overview of feature loadings is provided in \u003cstrong\u003eSupplementary Figs.\u0026nbsp;7\u0026ndash;12\u003c/strong\u003e, with detailed associations listed in \u003cstrong\u003eSupplementary Tables\u0026nbsp;5\u0026ndash;7\u003c/strong\u003e.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eData Availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eRaw sequencing data is available at the European Nucleotide Archive (ENA) at (https://www.ebi.ac.uk/ena/browser/view/PRJEB82744).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCode Availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe underlying code for this study is available on GitHub and can be accessed via https://doi.org/10.5281/zenodo.14899658.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics statement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eInformed consent was obtained from both parents per the Declaration of Helsinki II. Ethical approval for this study was granted by The Central Denmark Regional Committee on Health Research Ethics (journal number 1-10-72-296-18v6). The study is registered at ClinicalTrials.gov, with the identifier NCT05111990.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eR.R. Jakobsen: investigation, methodology, writing - original draft; G.R. van der Ploeg: formal analysis, methodology, writing - original draft; U.K. Sundekilde: conceptualisation,funding,sample collection,laboratory analysis; J. Astono: sample collection,laboratory analysis; K.O. Poulsen: sample collection,laboratory analysis; J.A. Westerhuis: conceptualization, methodology, supervision, formal analysis, writing - original draft; J.Fuglsang: participant recruitment, clinical follow-up, A. Heintz-Buschart: conceptualization, formal analysis, supervision, writing - original draft; A.K. Smilde: conceptualization, methodology, supervision, funding acquisition, writing - review \u0026amp; editing; D.S. Nielsen: conceptualization, supervision, funding, writing - review \u0026amp; editing. All authors reviewed and approved the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. K.O. Poulsen currently holds a position at Arla Foods, but was employed at AU Food when contributing to the project.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eG.R. van der Ploeg was funded by a grant from the University of Amsterdam, Research Priority Area on Personal Microbiome Health.\u003c/p\u003e\n\u003cp\u003eFunding for R.R. Jakobsen and the MAINHEALTH study was provided from an Arla Food for Health strategic research grant.\u003c/p\u003e\n\u003cp\u003eFunding for K.O. Poulsen was provided by the Sino-Danish Center for Research.\u003c/p\u003e\n\u003cp\u003eData was generated by accessing research infrastructure at Copenhagen University and Aarhus University, including equipment financed by FOODHAY (Food and Health Open Innovation Laboratory, Danish Roadmap for Research Infrastructure).\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eMonteiro, P. O. A. \u0026amp; Victora, C. G. Rapid growth in infancy and childhood and obesity in later life \u0026ndash; a systematic review. \u003cem\u003eObes. Rev. \u003c/em\u003e\u003cstrong\u003e6\u003c/strong\u003e, 143\u0026ndash;154 (2005).\u003c/li\u003e\n\u003cli\u003eGaudet, L., Ferraro, Z. M., Wen, S. W. \u0026amp; Walker, M. Maternal obesity and occurrence of fetal macrosomia: A systematic review and meta-analysis. \u003cem\u003eBioMed Res. Int. \u003c/em\u003e\u003cstrong\u003e2014\u003c/strong\u003e, (2014).\u003c/li\u003e\n\u003cli\u003eVoerman, E. \u003cem\u003eet al.\u003c/em\u003e Maternal body mass index, gestational weight gain, and the risk of overweight and obesity across childhood: An individual participant data meta-analysis. \u003cem\u003ePLOS Med. \u003c/em\u003e\u003cstrong\u003e16\u003c/strong\u003e, e1002744 (2019).\u003c/li\u003e\n\u003cli\u003eWeng, S. F., Redsell, S. A., Swift, J. A., Yang, M. \u0026amp; Glazebrook, C. P. Systematic review and meta-analyses of risk factors for childhood overweight identifiable during infancy. \u003cem\u003eArch. Dis. Child. \u003c/em\u003e\u003cstrong\u003e97\u003c/strong\u003e, 1019\u0026ndash;1026 (2012).\u003c/li\u003e\n\u003cli\u003eHorta, B. L., Rollins, N., Dias, M. S., Garcez, V. \u0026amp; P\u0026eacute;rez-Escamilla, R. Systematic review and meta-analysis of breastfeeding and later overweight or obesity expands on previous study for World Health Organization. \u003cem\u003eActa Paediatr. \u003c/em\u003e\u003cstrong\u003e112\u003c/strong\u003e, 34\u0026ndash;41 (2023).\u003c/li\u003e\n\u003cli\u003eAmir, L. H. \u0026amp; Donath, S. A systematic review of maternal obesity and breastfeeding intention, initiation and duration. \u003cem\u003eBMC Pregnancy Childbirth \u003c/em\u003e\u003cstrong\u003e7\u003c/strong\u003e, 9 (2007).\u003c/li\u003e\n\u003cli\u003eOng, K. \u0026amp; Loos, R. Rapid infancy weight gain and subsequent obesity: Systematic reviews and hopeful suggestions. \u003cem\u003eActa P\u0026aelig;diatrica \u003c/em\u003e\u003cstrong\u003e95\u003c/strong\u003e, 904\u0026ndash;908 (2006).\u003c/li\u003e\n\u003cli\u003eLind, M. V., Larnkj\u0026aelig;r, A., M\u0026oslash;lgaard, C. \u0026amp; Michaelsen, K. F. Breastfeeding, Breast Milk Composition, and Growth Outcomes. (2018) doi:10.1159/000486493.\u003c/li\u003e\n\u003cli\u003eAhearn-Ford, S., Berrington, J. E. \u0026amp; Stewart, C. J. Development of the gut microbiome in early life. \u003cem\u003eExp. Physiol. \u003c/em\u003e\u003cstrong\u003e107\u003c/strong\u003e, 415 (2022).\u003c/li\u003e\n\u003cli\u003eSarkar, A., Yoo, J. Y., Dutra, S. V. O., Morgan, K. H. \u0026amp; Groer, M. The Association between Early-Life Gut Microbiota and Long-Term Health and Diseases. \u003cem\u003eJ. Clin. Med. 2021 Vol 10 Page 459 \u003c/em\u003e\u003cstrong\u003e10\u003c/strong\u003e, 459 (2021).\u003c/li\u003e\n\u003cli\u003eStokholm, J. \u003cem\u003eet al.\u003c/em\u003e Maturation of the gut microbiome and risk of asthma in childhood. \u003cem\u003eNat. Commun. 2018 91 \u003c/em\u003e\u003cstrong\u003e9\u003c/strong\u003e, 1\u0026ndash;10 (2018).\u003c/li\u003e\n\u003cli\u003eIsganaitis, E. \u003cem\u003eet al.\u003c/em\u003e Maternal obesity and the human milk metabolome: associations with infant body composition and postnatal weight gain. \u003cem\u003eAm. J. Clin. Nutr. \u003c/em\u003e\u003cstrong\u003e110\u003c/strong\u003e, 111\u0026ndash;120 (2019).\u003c/li\u003e\n\u003cli\u003eSaben, J. L., Sims, C. R., Piccolo, B. D. \u0026amp; Andres, A. Maternal adiposity alters the human milk metabolome: associations between nonglucose monosaccharides and infant adiposity. \u003cem\u003eAm. J. Clin. Nutr. \u003c/em\u003e\u003cstrong\u003e112\u003c/strong\u003e, 1228\u0026ndash;1239 (2020).\u003c/li\u003e\n\u003cli\u003eBeyerlein, A. \u0026amp; Von Kries, R. Breastfeeding and body composition in children: will there ever be conclusive empirical evidence for a protective effect against overweight? \u003cem\u003eAm. J. Clin. Nutr. \u003c/em\u003e\u003cstrong\u003e94\u003c/strong\u003e, S1772\u0026ndash;S1775 (2011).\u003c/li\u003e\n\u003cli\u003eHorta, B. L., Rollins, N., Dias, M. S., Garcez, V. \u0026amp; P\u0026eacute;rez-Escamilla, R. Systematic review and meta-analysis of breastfeeding and later overweight or obesity expands on previous study for World Health Organization. \u003cem\u003eActa Paediatr. \u003c/em\u003e\u003cstrong\u003e112\u003c/strong\u003e, 34\u0026ndash;41 (2023).\u003c/li\u003e\n\u003cli\u003eVieira Queiroz De Paula, M., Grant, M., Lanigan, J. \u0026amp; Singhal, A. Does human milk composition predict later risk of obesity? A systematic review. \u003cem\u003eBMC Nutr. \u003c/em\u003e\u003cstrong\u003e9\u003c/strong\u003e, 1\u0026ndash;10 (2023).\u003c/li\u003e\n\u003cli\u003ePetraroli, M., Castellone, E., Patianna, V. \u0026amp; Esposito, S. Gut Microbiota and Obesity in Adults and Children: The State of the Art. \u003cem\u003eFront. Pediatr. \u003c/em\u003e\u003cstrong\u003e9\u003c/strong\u003e, 657020 (2021).\u003c/li\u003e\n\u003cli\u003eCheng, L., Akkerman, R., Kong, C., Walvoort, M. T. C. \u0026amp; de Vos, P. More than sugar in the milk: human milk oligosaccharides as essential bioactive molecules in breast milk and current insight in beneficial effects. \u003cem\u003eCrit. Rev. Food Sci. Nutr. \u003c/em\u003e\u003cstrong\u003e61\u003c/strong\u003e, 1184\u0026ndash;1200 (2021).\u003c/li\u003e\n\u003cli\u003eThomson, P., Medina, D. A. \u0026amp; Garrido, D. Human milk oligosaccharides and infant gut bifidobacteria: Molecular strategies for their utilization. \u003cem\u003eFood Microbiol. \u003c/em\u003e\u003cstrong\u003e75\u003c/strong\u003e, 37\u0026ndash;46 (2018).\u003c/li\u003e\n\u003cli\u003eKorpela, K. \u003cem\u003eet al.\u003c/em\u003e Childhood BMI in relation to microbiota in infancy and lifetime antibiotic use. \u003cem\u003eMicrobiome \u003c/em\u003e\u003cstrong\u003e5\u003c/strong\u003e, 26 (2017).\u003c/li\u003e\n\u003cli\u003eDemmelmair, H., Jim\u0026eacute;nez, E., Collado, M. C., Salminen, S. \u0026amp; McGuire, M. K. Maternal and Perinatal Factors Associated with the Human Milk Microbiome. \u003cem\u003eCurr. Dev. Nutr. \u003c/em\u003e\u003cstrong\u003e4\u003c/strong\u003e, nzaa027 (2020).\u003c/li\u003e\n\u003cli\u003eDaiy, K., Harries, V., Nyhan, K. \u0026amp; Marcinkowska, U. M. Maternal weight status and the composition of the human milk microbiome: A scoping review. \u003cem\u003ePLoS ONE \u003c/em\u003e\u003cstrong\u003e17\u003c/strong\u003e, e0274950 (2022).\u003c/li\u003e\n\u003cli\u003ePoulsen, K. O. \u003cem\u003eet al.\u003c/em\u003e Influence of maternal body mass index on human milk composition and associations to infant metabolism and gut colonisation: MAINHEALTH - a study protocol for an observational birth cohort. \u003cem\u003eBMJ Open \u003c/em\u003e\u003cstrong\u003e12\u003c/strong\u003e, e059552 (2022).\u003c/li\u003e\n\u003cli\u003eJakobsen, R. \u003cem\u003eet al.\u003c/em\u003e Maternal pre-pregnancy BMI influences breast milk composition, infant gut microbiome development, and early-life growth of term infants. Preprint at https://doi.org/10.21203/rs.3.rs-6075035/v1 (2025).\u003c/li\u003e\n\u003cli\u003eAjeeb, T. T., Gonzalez, E., Solomons, N. W., Vossenaar, M. \u0026amp; Koski, K. G. Human milk microbiome: associations with maternal diet and infant growth. \u003cem\u003eFront. Nutr. \u003c/em\u003e\u003cstrong\u003e11\u003c/strong\u003e, 1341777 (2024).\u003c/li\u003e\n\u003cli\u003eBaumgartel, K. \u003cem\u003eet al.\u003c/em\u003e The Human Milk Metabolome: A Scoping Literature Review. \u003cem\u003eJ. Hum. Lact. \u003c/em\u003e\u003cstrong\u003e39\u003c/strong\u003e, 255\u0026ndash;277 (2023).\u003c/li\u003e\n\u003cli\u003eStinson, L. F. \u003cem\u003eet al.\u003c/em\u003e The human milk microbiome: who, what, when, where, why, and how? \u003cem\u003eNutr. Rev. \u003c/em\u003e\u003cstrong\u003e79\u003c/strong\u003e, 529\u0026ndash;543 (2021).\u003c/li\u003e\n\u003cli\u003eBro, R. Multiway calibration. Multilinear PLS. \u003cem\u003eJ. Chemom. \u003c/em\u003e\u003cstrong\u003e10\u003c/strong\u003e, 47\u0026ndash;61 (1996).\u003c/li\u003e\n\u003cli\u003evan der Ploeg, G. R., Westerhuis, J. A., Heintz-Buschart, A. \u0026amp; Smilde, A. K. parafac4microbiome: Exploratory analysis of longitudinal microbiome data using Parallel Factor Analysis. Preprint at https://doi.org/10.1101/2024.05.02.592191 (2024).\u003c/li\u003e\n\u003cli\u003eArenz, S., R\u0026uuml;ckerl, R., Koletzko, B. \u0026amp; Von Kries, R. Breast-feeding and childhood obesity--a systematic review. \u003cem\u003eInt. J. Obes. Relat. Metab. Disord. J. Int. Assoc. Study Obes. \u003c/em\u003e\u003cstrong\u003e28\u003c/strong\u003e, 1247\u0026ndash;1256 (2004).\u003c/li\u003e\n\u003cli\u003eFan, Y., Vinjamuri, A., Tu, D., Lebrilla, C. B. \u0026amp; Donovan, S. M. Determinants of human milk oligosaccharides profiles of participants in the STRONG kids 2 cohort. \u003cem\u003eFront. Nutr. \u003c/em\u003e\u003cstrong\u003e10\u003c/strong\u003e, (2023).\u003c/li\u003e\n\u003cli\u003eDuranti, S. \u003cem\u003eet al.\u003c/em\u003e Bifidobacterium bifidum and the infant gut microbiota: an intriguing case of microbe-host co-evolution. \u003cem\u003eEnviron. Microbiol. \u003c/em\u003e\u003cstrong\u003e21\u003c/strong\u003e, 3683\u0026ndash;3695 (2019).\u003c/li\u003e\n\u003cli\u003eWalker, W. A. \u0026amp; Iyengar, R. S. Breast milk, microbiota, and intestinal immune homeostasis. \u003cem\u003ePediatr. Res. \u003c/em\u003e\u003cstrong\u003e77\u003c/strong\u003e, 220\u0026ndash;228 (2015).\u003c/li\u003e\n\u003cli\u003eStuivenberg, G. A., Burton, J. P., Bron, P. A. \u0026amp; Reid, G. Why Are Bifidobacteria Important for Infants? \u003cem\u003eMicroorg. 2022 Vol 10 Page 278 \u003c/em\u003e\u003cstrong\u003e10\u003c/strong\u003e, 278 (2022).\u003c/li\u003e\n\u003cli\u003eVu, K. \u003cem\u003eet al.\u003c/em\u003e From Birth to Overweight and Atopic Disease: Multiple and Common Pathways of the Infant Gut Microbiome. \u003cem\u003eGastroenterology \u003c/em\u003e\u003cstrong\u003e160\u003c/strong\u003e, 128-144.e10 (2021).\u003c/li\u003e\n\u003cli\u003eFull article: Functional effects of human milk oligosaccharides (HMOs). https://www.tandfonline.com/doi/full/10.1080/19490976.2023.2186115.\u003c/li\u003e\n\u003cli\u003eWinter, S. E. \u0026amp; B\u0026auml;umler, A. J. Gut dysbiosis: Ecological causes and causative effects on human disease. \u003cem\u003eProc. Natl. Acad. Sci. \u003c/em\u003e\u003cstrong\u003e120\u003c/strong\u003e, e2316579120 (2023).\u003c/li\u003e\n\u003cli\u003eB\u0026auml;ckhed, F. \u003cem\u003eet al.\u003c/em\u003e Dynamics and Stabilization of the Human Gut Microbiome during the First Year of Life. \u003cem\u003eCell Host Microbe \u003c/em\u003e\u003cstrong\u003e17\u003c/strong\u003e, 690\u0026ndash;703 (2015).\u003c/li\u003e\n\u003cli\u003eJu, T. \u003cem\u003eet al.\u003c/em\u003e The Gut Commensal Escherichia coli Aggravates High-Fat-Diet-Induced Obesity and Insulin Resistance in Mice. \u003cem\u003eAppl. Environ. Microbiol. \u003c/em\u003e\u003cstrong\u003e89\u003c/strong\u003e, e01628-22 (2023).\u003c/li\u003e\n\u003cli\u003eF\u0026aring;k, F., Karlsson, C. L. J., Ahrn\u0026eacute;, S., Molin, G. \u0026amp; Westr\u0026ouml;m, B. Effects of a high-fat diet during pregnancy and lactation are modulated by E. coli in rat offspring. \u003cem\u003eInt. J. Obes. \u003c/em\u003e\u003cstrong\u003e36\u003c/strong\u003e, 744\u0026ndash;751 (2012).\u003c/li\u003e\n\u003cli\u003eAzad, M. B., Brockway, M. M. \u0026amp; Reyes, S. M. Human milk composition and infant anthropometrics: overview of a systematic review with clinical and research implications. \u003cem\u003eInt. Breastfeed. J. \u003c/em\u003e\u003cstrong\u003e19\u003c/strong\u003e, 45 (2024).\u003c/li\u003e\n\u003cli\u003eLeghi, G. E. \u003cem\u003eet al.\u003c/em\u003e The Impact of Maternal Obesity on Human Milk Macronutrient Composition: A Systematic Review and Meta-Analysis. \u003cem\u003eNutrients \u003c/em\u003e\u003cstrong\u003e12\u003c/strong\u003e, 934 (2020).\u003c/li\u003e\n\u003cli\u003eJuncker, H. G. \u003cem\u003eet al.\u003c/em\u003e Maternal stress in the early postpartum period is associated with alterations in human milk microbiome composition. \u003cem\u003eBrain. Behav. Immun. \u003c/em\u003e\u003cstrong\u003e124\u003c/strong\u003e, 74\u0026ndash;84 (2025).\u003c/li\u003e\n\u003cli\u003eAcar, E., Bro, R. \u0026amp; Smilde, A. Data Fusion in Metabolomics Using Coupled Matrix and Tensor Factorizations. \u003cem\u003eProc. IEEE \u003c/em\u003e\u003cstrong\u003e103\u003c/strong\u003e, 1602 (2015).\u003c/li\u003e\n\u003cli\u003eAcar, E. \u003cem\u003eet al.\u003c/em\u003e Structure-revealing data fusion. \u003cem\u003eBMC Bioinformatics \u003c/em\u003e\u003cstrong\u003e15\u003c/strong\u003e, 239 (2014).\u003c/li\u003e\n\u003cli\u003eAcar, E., Kolda, T. G. \u0026amp; Dunlavy, D. M. All-at-once Optimization for Coupled Matrix and Tensor Factorizations. Preprint at http://arxiv.org/abs/1105.3422 (2011).\u003c/li\u003e\n\u003cli\u003eTun, H. M. \u003cem\u003eet al.\u003c/em\u003e Roles of Birth Mode and Infant Gut Microbiota in Intergenerational Transmission of Overweight and Obesity From Mother to Offspring. \u003cem\u003eJAMA Pediatr. \u003c/em\u003e\u003cstrong\u003e172\u003c/strong\u003e, 368\u0026ndash;377 (2018).\u003c/li\u003e\n\u003cli\u003eMueller, N. T. \u003cem\u003eet al.\u003c/em\u003e Does vaginal delivery mitigate or strengthen the intergenerational association of overweight and obesity? Findings from the Boston Birth Cohort. \u003cem\u003eInt. J. Obes. \u003c/em\u003e\u003cstrong\u003e41\u003c/strong\u003e, 497\u0026ndash;501 (2017).\u003c/li\u003e\n\u003cli\u003eCooke, R. \u003cem\u003eet al.\u003c/em\u003e Catch-Up Growth in Infants and Young Children With Faltering Growth: Expert Opinion to Guide General Clinicians. \u003cem\u003eJ. Pediatr. Gastroenterol. Nutr. \u003c/em\u003e\u003cstrong\u003e77\u003c/strong\u003e, 7 (2023).\u003c/li\u003e\n\u003cli\u003eLampl, M., Thompson, A. L. \u0026amp; Frongillo, E. A. Sex Differences in the Relationships Among Weight Gain, Subcutaneous Skinfold Tissue and Saltatory Length Growth Spurts in Infancy. \u003cem\u003ePediatr. Res. \u003c/em\u003e\u003cstrong\u003e58\u003c/strong\u003e, 1238\u0026ndash;1242 (2005).\u003c/li\u003e\n\u003cli\u003eDe Wit, C. C., Sas, T. C. J., Wit, J. M. \u0026amp; Cutfield, W. S. Patterns of Catch-Up Growth. \u003cem\u003eJ. Pediatr. \u003c/em\u003e\u003cstrong\u003e162\u003c/strong\u003e, 415\u0026ndash;420 (2013).\u003c/li\u003e\n\u003cli\u003eTeam, R. C. R: A Language and Environment for Statistical Title. R Foundation for Statistical Computing.\u003c/li\u003e\n\u003cli\u003eOrganization, W. H. \u003cem\u003eWHO Child Growth Standards : Length/Height-for-Age, Weight-for-Age, Weight-for-Length, Weight -for-Height and Body Mass Index-for-Age : Methods and Development\u003c/em\u003e. (World Health Organization, 2006).\u003c/li\u003e\n\u003cli\u003eSundekilde, U. K. \u003cem\u003eet al.\u003c/em\u003e The Effect of Gestational and Lactational Age on the Human Milk Metabolome. \u003cem\u003eNutrients \u003c/em\u003e\u003cstrong\u003e8\u003c/strong\u003e, (2016).\u003c/li\u003e\n\u003cli\u003eAndreas, N. J. \u003cem\u003eet al.\u003c/em\u003e Multiplatform characterization of dynamic changes in breast milk during lactation. \u003cem\u003eElectrophoresis \u003c/em\u003e\u003cstrong\u003e36\u003c/strong\u003e, 2269\u0026ndash;2285 (2015).\u003c/li\u003e\n\u003cli\u003eAndersen-Civil, A. I. S. \u003cem\u003eet al.\u003c/em\u003e Dietary proanthocyanidins promote localized antioxidant responses in porcine pulmonary and gastrointestinal tissues during Ascaris suum-induced type 2 inflammation. \u003cem\u003eFASEB J. \u003c/em\u003e\u003cstrong\u003e36\u003c/strong\u003e, e22256 (2022).\u003c/li\u003e\n\u003cli\u003eHui, Y., Nielsen, D. S. \u0026amp; Krych, L. De novo clustering of long-read amplicons improves phylogenetic insight into microbiome data. \u003cem\u003ebioRxiv\u003c/em\u003e 2023.11.26.568539 (2023) doi:10.1101/2023.11.26.568539.\u003c/li\u003e\n\u003cli\u003eGloor, G. B., Macklaim, J. M., Pawlowsky-Glahn, V. \u0026amp; Egozcue, J. J. Microbiome Datasets Are Compositional: And This Is Not Optional. \u003cem\u003eFront. Microbiol. \u003c/em\u003e\u003cstrong\u003e8\u003c/strong\u003e, 2224 (2017).\u003c/li\u003e\n\u003cli\u003eAitchison, J. The Statistical Analysis of Compositional Data. \u003cem\u003eJ. R. Stat. Soc. Ser. B Methodol. \u003c/em\u003e\u003cstrong\u003e44\u003c/strong\u003e, 139\u0026ndash;160 (1982).\u003c/li\u003e\n\u003cli\u003eBro, R. \u0026amp; Smilde, A. K. Centering and scaling in component analysis. \u003cem\u003eJ. Chemom. \u003c/em\u003e\u003cstrong\u003e17\u003c/strong\u003e, 16\u0026ndash;33 (2003).\u003c/li\u003e\n\u003cli\u003eBro, R. The N-way Toolbox. \u003cem\u003eThe N-way Toolbox \u003c/em\u003ehttps://nl.mathworks.com/matlabcentral/fileexchange/1088-the-n-way-toolbox (2023).\u003c/li\u003e\n\u003cli\u003eWickham, H. Data Analysis. in \u003cem\u003eggplot2\u003c/em\u003e 189\u0026ndash;201 (Springer International Publishing, Cham, 2016). doi:10.1007/978-3-319-24277-4_9.\u003c/li\u003e\n\u003cli\u003eR Core Team, R. R: A language and environment for statistical computing. (2013).\u003c/li\u003e\n\u003cli\u003eMaechler, M. Finding groups in data: Cluster analysis extended Rousseeuw et al. \u003cem\u003eR Package Version \u003c/em\u003e\u003cstrong\u003e2\u003c/strong\u003e, 242\u0026ndash;248 (2019).\u003c/li\u003e\n\u003cli\u003eKassambara, A. \u0026amp; Mundt, F. Package \u0026lsquo;factoextra\u0026rsquo;. \u003cem\u003eExtr. Vis. \u003c/em\u003e\u003cem\u003eResults Multivar. Data Anal. \u003c/em\u003e\u003cstrong\u003e76\u003c/strong\u003e, (2017).\u003c/li\u003e\n\u003cli\u003eWickham, H., Chang, W. \u0026amp; Wickham, M. H. Package \u0026lsquo;ggplot2\u0026rsquo;. \u003cem\u003eCreate Elegant Data Vis. Using Gramm. Graph. Version \u003c/em\u003e\u003cstrong\u003e2\u003c/strong\u003e, 1\u0026ndash;189 (2016).\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-6244750/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6244750/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eMaternal obesity is a key risk factor for excessive foetal growth and childhood obesity, yet its influence on human milk (HM) composition and the infant gut microbiome development remains unclear. This study examined 169 mother-infant dyads analyzing 570 HM metabolome, 495 HM microbiome, and 348 infant faecal microbiome samples over three months of exclusive breastfeeding, alongside infant anthropometric data through three years postpartum. While BMI was not directly correlated with infant growth (weight-for-length/height z-score), N-way Partial Least Squares modelling revealed microbial and metabolite signatures linked to maternal ppBMI and infant growth. High maternal ppBMI and infant growth were associated with altered HM oligosaccharides and enrichment of \u003cem\u003eBifidobacterium\u003c/em\u003e spp. in the infant gut. In contrast, elevated HM simple sugars, amino acid derivatives, and gut \u003cem\u003eKlebsiella\u003c/em\u003e and \u003cem\u003eEscherichia\u003c/em\u003e spp. relative abundance linked to slower growth. These findings highlight maternal-infant nutritional dynamics, informing targeted strategies to support infant growth.\u003c/p\u003e","manuscriptTitle":"Supervised Modelling of Longitudinal Human Milk and Infant Gut Microbiome Reveal Maternal Pre-Pregnancy BMI and Early Life Growth Interactions","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-03-31 11:43:54","doi":"10.21203/rs.3.rs-6244750/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"9b3c268c-2969-4aed-969a-f63b28921d46","owner":[],"postedDate":"March 31st, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":46178258,"name":"Biological sciences/Microbiology/Microbial communities/Microbiome"},{"id":46178259,"name":"Biological sciences/Systems biology/Time series"},{"id":46178260,"name":"Biological sciences/Computational biology and bioinformatics/Data integration"}],"tags":[],"updatedAt":"2025-03-31T11:43:54+00:00","versionOfRecord":[],"versionCreatedAt":"2025-03-31 11:43:54","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6244750","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6244750","identity":"rs-6244750","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}