Dietary impact on infants' gut microbiota and its capacity in SCFA metabolism: a longitudinal Dutch cohort study

Dietary impact on infants' gut microbiota and its capacity in SCFA metabolism: a longitudinal Dutch cohort study | 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 Research Article Dietary impact on infants' gut microbiota and its capacity in SCFA metabolism: a longitudinal Dutch cohort study Evgenia Dikareva, Michał Skawiński, Liene Bervoets, David Barnett, and 6 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6673441/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 Background: The infant gut microbiota (GM) matures rapidly during early life with complementary feeding marking a pivotal dietary shift that can shape long-term health trajectories. Detailed insight into how timing and composition of solid food introduction influences the composition and metabolic potential of the infant microbiome remains, however, limited. This study aimed to evaluate the impact of complementary feeding dynamics on the infant GM. Methods: We conducted longitudinal whole metagenomic sequencing of fecal samples collected at 4, 5, 6, 9, 11, and 14 months of age from 112 Dutch infants in the LucKi Gut cohort. Based on dietary questionnaires, infants were grouped into three distinct dietary classes of complementary feeding. Results: Infants introduced earlier to a wider variety of solid foods exhibited more diverse and mature gut microbiota already at 4 months, with increased abundance of butyrate-producing taxa such as Flavonifractor plautii . Their microbiomes also showed enhanced capacity to degrade dietary fibers like xylan and rhamnogalacturonan, suggesting accelerated development of metabolic functionality. Functional profiling revealed early enrichment in genes involved in butyrate synthesis, pointing to a link between early feeding diversity and SCFA-producing potential. Conclusions: Our findings highlight that early and diverse complementary feeding fosters a functionally mature microbiota with enhanced fiber degradation and butyrate production capacity. These microbial trajectories may influence immune and metabolic development, underscoring the importance of timely dietary diversification in infancy. Complimentary food introduction of solids dietary fiber xylan butyrate Flavonifractor plautii Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Infancy represents a critical period for gut microbiome development, which is pivotal for immune system maturation and overall long-term health. The maturation of the infant gut microbiome is a dynamic process that undergoes rapid transformations influenced, for instance, by mode of delivery ( 1 – 6 ), type of milk being fed ( 1 , 6 ) and the time and type of complementary food introduction ( 7 ). While several other factors, including environmental exposures ( 1 , 2 , 8 – 10 ) and host genetics ( 11 ) contribute to the shaping of the microbial community, diet stands out as the most profound driver of microbial assembly and maturation in early life ( 12 – 14 ). The transition from a milk-based diet to solid foods heralds a shift in the microbial maturation trajectories, eventually resulting in an adult-like gut microbiota (GM) composition at around 2 years of age ( 15 ). During this transition, the diversity of bacterial taxa capable of metabolizing complex carbohydrates expands and impacts metabolic ( 16 ) and immune responses ( 17 ). Animal studies have shown that weaning impacts the development of immune responses and host metabolism, with effects lasting into adulthood ( 18 , 19 ). Several dietary studies have indicated that the diversity of foods consumed may lead to significant variations in the composition of the adult GM ( 20 – 24 ). On the contrary, only a few human studies have looked into the effects of individual dietary components ( 25 ). To our knowledge, only one previous study investigated associations between specific food items or patterns and microbiota maturation during the weaning period ( 26 ). With evidence accumulating for a link between early-life microbial imbalances and later health outcomes such as allergies ( 27 ), autoimmune diseases ( 28 ) and neurodevelopmental disorders ( 29 ), there is an urgent need to delve into these dietary influences that shape the infant gut microbiome. Investigating these influences during the critical developmental period is vital, given their potential to uncover preventative strategies aimed at mitigating the risk of later-life health conditions. Here, we explore early-life feeding patterns and infants’ microbiota development and maturation within the LucKi-Gut cohort. Longitudinally collected (4, 5, 6, 9, 11, and 14 months of age) fecal samples, profiled by whole metagenome shotgun sequencing, were combined with dietary data. We showed that not only the type of solid foods, but also the time of introduction and food item diversity, shapes the infant microbiota composition and functional capacity. Methods Study population and design The LucKi Gut Study is an ongoing longitudinal birth cohort study aimed at monitoring microbiome development during infancy and early childhood. Pregnant women living in the South Limburg region of the Netherlands were recruited through obstetric practices, gynaecology departments, during lactation information sessions, and via advertisements at venues for pregnancy yoga, baby clothes stores and social media. Infants born before 36 weeks of gestation were excluded from the present analyses. Infant fecal samples were collected at age 1–2 weeks and 1, 2, 4, 5, 6, 9, 11, and 14 months postpartum. Parents were instructed to collect the infant’s fecal samples from the diaper and immediately freeze the samples at -20⁰C in special transport containers (Sarstedt, Hilden, Germany) in their home freezer. All samples were transported to the laboratory of the department of Medical Microbiology, Infectious Diseases and Infection Prevention at Maastricht University Medical Centre and were aliquoted and frozen at -80°C until further use. At every fecal sampling time-point, parents were additionally asked to complete a questionnaire. Questionnaires collected data on infants’ diet, health and developmental status, medication use, as well as on maternal health (during pregnancy), medication use, lifestyle and diet. The questionnaires included detailed questions on the type of infant feeding and the introduction of complementary foods. For the present study, we included 112 infants for whom, at time of analyses, dietary data were available for the first 14 months of life (Supplementary Table 1A-C). Of these 112 infants, microbial profiling data were available for 105 infants. We restricted our analysis to fecal samples collected between 4 and 14 months of age (n = 389), as complementary food was introduced the earliest around 4 months of age (Supplementary Table 2A-B). Shotgun metagenomics sequencing Total metagenomic DNA was extracted by mechanical lysis on a FastPrep-96 homogenizer (MP Biomedicals, USA) followed by DNA purification using the MagPure Stool DNA Kit (Magen Biotechnology Ltd, China). Sequencing was performed on the BGISEQ-500 platform following the standard protocol ( 30 ) at MGI (Riga, Latvia). To standardize the pipeline, the workflow manager Snakemake v5.14.0 was used ( 31 ). The pre-processing comprised quality control with Fastp v0.20.1 (quality phred score: 15; minimal read length, 60 bp) ( 32 ). In addition, Fastp was used to trim the BGI-SEQ adapters “AAGTCGGAGGCCAAGCGGTCTTAGGAAGACAA” for forward and “AAGTCGGATCGTAGCCATGTCGTTCTGTGAGCCAAGGAGTTG” for reverse reads. Human reads were removed by aligning against chm13.draft_v1.0_plusY (downloaded 14.10.2020) with Bowtie 2 v2.3.5.1 (maximum insert size 600 bp) ( 33 ) and Samtools v.1.9 ( 34 ). Subsequently, the forward and reverse reads were used to identify taxonomic composition using MetaPhlAn v3.0 (species-markers database from January 2019 CHOCOPhlAn v30: mpa_v30_CHOCOPhlAn_201901) ( 35 ) by aligning reads to the reference database of marker genes. KEGG-based annotation of short-chain fatty acids (SCFAs) production capacity We used HUMAnN (v3.9) ( 35 ) with the Enzyme Commission (EC) filtered UniRef90 database with default search settings to quantify features of the functional capacity of each sample that are relevant to short chain fatty acid metabolism. UniRef features were regrouped into KEGG orthology (KO) terms using the “humann_regroup_table” utility script provided by HUMAnN. Relevant KO identifiers (n = 65, Supplementary Table 3)( 36 – 39 ) for SCFA production were manually selected from the KEGG database. Data analysis The analysis unfolded in three sequential steps: ( 1 ) identification of groups of infants exhibiting similar longitudinal feeding patterns and separating them into distinct dietary classes, ( 2 ) exploration of the differential effects of identified classes on microbiome development and ( 3 ) determination of significant relations between dietary classes and bacterial taxa, as well as between dietary classes, bacterial taxa and dietary fiber degrading capacity. All analyses were conducted using R version 4.1.2 ( 40 ). Missing values (yes/no) for food items, breastfeeding and formula feeding were imputed based on the answers from the adjacent time points. If this was not possible, the mode across all answers at the given time point was used. The following variables were used: delivery type [vaginal/C-section], delivery place [at home/hospital], sex [male/female], presence of furry pets [none/at home and/or at daycare/solely at daycare], formula [yes/no] and breastfeeding [yes/no] at sample collection, older siblings [none/1/2/3], duration of breastfeeding (in months), age when solids were introduced (in weeks), gestational age (in weeks) and birth weight (in grams). Identification and characterization of dietary classes A list covering regularly used food items was included in each questionnaire. At an infant’s age of 4 to 6 months, ‘bread’, ‘meat’ and ‘fish’ were separate yes/no items, followed by an open-ended question in which the type of bread, meat or fish, respectively, was entered by the parents. From age 9 months onwards, all food items were asked about separately (yes/no answer option). Food items from all questionnaires were harmonized and merged when too few infants consumed individual food items. For this reason, meat, chicken, pork and beef were merged into one category – “meat”. Also, bread, white bread, brown bread and wholegrain bread were merged into “bread”, as between 4–6 months parents not always specified the type of bread given. For some food items, closed questions were only available at later time-points, including pudding (from 6 months onwards), melon, peach, tomato, cheese, butter, margarine, egg and soy products (from 9 months onwards). For these food items, “no” was imputed for the earlier time-points, unless these food items were listed by parents in the open-ended question (separate options for “Other fruit”,”Other vegetables” and ”something else namely”) in which case “yes” was imputed for the listed food item. For all other food items, data were available for all time points. A variation of Principal Component Analysis (PCA), namely logistic PCA (logistic PCA v0.2) ( 41 ), was applied to all dietary data at each time point to reduce the dimensionality. With cv.lpca function, the value of parameter m (m = 3 for each model) was estimated, necessary for the approximation of the natural parameter. Resulting PC scores are the linear combinations of those parameters, projected from the Bernoulli saturated model. Each infant was represented as a sequence (of length 6) of PC scores (1 and 2) for each subsequent time point, which allowed tracking the change of infants’ diets with time in the PC subspace. The differences between infants were quantified with the multivariate dynamic time warping (DTW) distance with ‘dtw’ package (v1.23-1) ( 42 ). This similarity metric accounts for the speed of a change and aims to find an optimal match between input time series. Groups of longitudinal feeding patterns (referred to as dietary classes) and their hierarchical representation were further identified with a hierarchical clustering algorithm, with complete linkage and DTW distance matrix. Dendrogram visualization was constructed with ‘dendextend’ package (v1.17.1) ( 43 ). To characterize dietary classes, proportions of infants who received a given solid food, aggregated per dietary class, were plotted for each time point in the form of stacked bar plots with ggplot2 package (v3.4.4) ( 44 ). Only food items with significantly different proportions ( prop.test , p-value cutoff at 0.05) between dietary classes were plotted. To additionally visualize changes in the aforementioned proportions over time, slope graphs were constructed using newggslopegraph function from ‘CGPfunctions’ package (v0.6.3) ( 45 ). Moreover, the time when a solid food was commonly introduced in each dietary class was calculated. A food item was considered to be widely introduced when more than 50% of the infants in a dietary class received that food item at the respective time point, but not before. To assess the diversity of the infants’ complementary food, various scores were calculated for each time point per infant. Food Variety Score (FVS) ( 46 ) reflects the number of unique food items introduced to the infant between 4–14 months (maximum of 28). Dietary Diversity Score (DDS) ( 46 ) reflects the number of unique food groups introduced. To this end, food items were grouped into eight groups representing vegetables, fruits, legumes/nuts, meat, fish, eggs, dairy, and grains as was described previously. Finally, the Food Allergen Diversity (FAD) was calculated as an index that accounts for the number of allergens introduced (milk, egg, wheat, fish, soy, nuts) up to a maximum of 6 ( 47 ). Clustering of the infants into dietary classes was performed independently of their environmental and lifestyle data. However, to exclude the effect of certain environmental and lifestyle factors on these dietary classes, we explored their variations between the classes. For instance, the effect of breast- and formula-feeding was compared longitudinally as proportions of infants in a dietary class, who were breast or formula-fed at each time point. Linear regression models for breast and formula-feeding were separately fitted between proportion and time (numerical) with interaction between time and dietary class with 95% confidence intervals. Exploration of the differential effects of dietary classes on microbiome development Alpha and beta diversity calculations The ecological alpha diversity metrics for microbiota communities (observed number of species, Shannon and Chao1 indices) were calculated for each sample and aggregated by dietary class and time point with function specnumber , diversity and estimateR from ‘vegan’ package (v2.6-4) ( 48 ), respectively. Both boxplots and loess regression lines fitted between days from sample collection and diversity metrics, using geom_smooth layer of the ggplot function, were plotted. Between-sample microbiota diversity (beta diversity) was investigated by performing the Principal Coordinates Analysis (PCoA) of Aitchison’s dissimilarity indices on a species level using all samples together. Progression of microbiome community structure with age was represented as a boxplot of the first principal coordinate (PCo) scores, coloured by dietary class across time points. PCoA was performed using the ‘microViz’ package (v0.11.0) ( 49 ). Microbial composition was determined by compositional transformation of the read count data at both the genus and species levels for each sample using the ‘microViz’ package. Taxa with an abundance greater than 1% were retained for further analysis. Following this selection, the mean relative taxa abundance was calculated for all dietary classes at each time point. Compositional heatmaps were constructed using the ‘ComplexHeatmap’ package (v2.18.0) ( 50 ). Furthermore, loess regression lines were fitted between days from the sample collection and the relative abundance of taxa (sample-wise), for each dietary class using geom_smooth with default parameters from the ‘ggplot2’ package. Dirichlet Multinomial Mixture ( DMM) clustering The progression of microbiome communities with age was further compared across dietary classes by constructing Dirichlet Multinomial Mixture (DMM) transition graphs ( 51 ). DMM clustering was conducted using the ‘DirichletMultinomial’ package (v1.44.0) ( 52 ), and the number of resulting clusters, known as enterotypes or ‘metacommunites’ was determined using the Laplace approximation score. Samples that exhibited values that differed by more than 41 days from their respective time points were filtered out before the analysis. The transition of infants through DMM clusters with age was visualized separately for each dietary class, as previously described ( 6 , 27 ). Differential abundance analysis To identify bacterial taxa (genus and species level) that differed over time between the dietary classes, linear models for differential abundance analysis (LinDA) were conducted for each taxon using the ‘LinDA’ package (v0.1.0) ( 53 ). LinDA uses centred log2-ratio transformed taxon abundances (read counts) as the response variable. An interaction between time points (categorical, with the first time point as reference level) and dietary classes (with the first class as reference) was included as a predictor variable. Additionally, duration of breastfeeding (numeric, months), age of the introduction of solid foods (numeric, weeks), birth weight (numeric, grams), sex (male as reference), delivery place (home/hospital with hospital as reference), and delivery type (C-section/vaginal with vaginal delivery as reference) were used as confounders. Infants’ random intercepts were modelled for each infant by specifying Infant IDs as a grouping factor. Taxa with a prevalence of more than 5% and an abundance greater than 1% for each time point were retained for the analysis. Microbial maturation Microbial age was calculated as initially described by Subramanian and colleagues ( 54 ). The Random Forest Cross-Validation regression was performed using the rfcv function from the ‘randomForest’ package (v4.7-1.1) ( 55 ). As training data, a set of 66% of the metagenomic read counts, balanced with respect to the dietary classes, was used. The response variable for regression was the age of sample collection. To predictive performance cross-validation error was visualized for each tested model. The best model had the minimal number of top-ranking age-discriminatory taxa. The importance of top-ranking taxa was assessed using metrics such as an increase in model squared error and node purity. Next, the sparse model of top-ranking taxa was used for microbial age prediction, with 100-cross validation of train/test split (66%/33%). The mean across all predictions was computed for each sample, representing the final predicted microbial age. Relative microbiota maturity (RMM) and microbiota-for-age Z-score Relative microbiota maturity (RMM) and the microbiota-for-age Z-score (MAZ) were computed following the methodology outlined in the aforementioned publication ( 54 ). The results were visualized using boxplots for each time point and dietary class. To establish a reference level for microbiota maturity, a fitted loess regression line was calculated, which represents the average change in microbiota maturity with age. In contrast, the references for the MAZ were calculated as the means of all samples' microbial ages at each specific time point. Statistical differences between dietary classes for microbial age, RMM and MAZ were calculated using the method described in Statistical analysis. To investigate the relationship between the RMM and time (as a multi-level categorical variable), a linear mixed-effect model was fitted with random intercepts for each child and an interaction term between dietary classes and time points with ‘lme4’ (v1.1-35.1) package. The significance of the coefficients was tested with ‘lmerTest’ package (v3.1-3). Adjustments were made for confounders, including sex, delivery type, delivery place, number of older siblings, presence of furry pets, birth weight, duration of breastfeeding, and age at introduction of solid foods. The regression results were visualized using a heatmap. Determination of significant relations between dietary classes, bacterial taxa and dietary fiber degradation capacities Inferred fiber degradation profiles To infer the metabolic capabilities of a microbiome community to degrade dietary fibers (DFs), a functional analysis of the microbiome’s inferred fiber degradation profiles (IFDP) was performed on metagenomic samples. These profiles reflect the relative capacity of the metagenomic community to degrade specific dietary fibers. A previously described computational framework ( 56 ) was used to connect metagenomic sequences with annotated polysaccharide-degrading enzymes and the chemical structures of dietary fibers. It has been shown that IFDPs are linked to the host diet and accurately reflect dietary classes. To calculate the IFDPs, contigs’ sequences for each sample were mapped to the functional database containing fiber-degrading enzymes’ amino acid sequences, using Diamond blastx (v.2.1.8) ( 57 ). Further, the mapped counts were multiplied by the enzyme-fiber interaction matrix to obtain the IFDPs, which were used for further analysis. t-Distributed Stochastic Neighbour Embedding (t-SNE) dimensionality reduction was performed on all relative DF degradation capabilities (IFDPs), expressed in percentages, to examine the overall sources of variation. The heatmap for DF, time point and diet was created using mean relative DF degradation capacity for each dietary class at a given time point. Correlations Correlations between the degradation capacities for each DF and bacterial species were analysed at each time point using the ‘stats’ package (v4.1.3). The 'cor.test' function was employed for paired correlation analysis using Spearman’s method, without exact p-value computation. Taxa with an average abundance of > 1% at any time point were included for these analyses. We focused on correlations with an estimate > 0.3, deemed statistically significant (q < 0.05), and meeting criteria across at least two time points. This resulted in identifying 15 fibers and 19 bacterial species. For each of the 15 fibers, the total sum of relative reads of all degraders was calculated. The same selected bacterial species were used in Spearman correlations with food items. Statistical analysis To statistically examine the associations between dietary classes and numerical variables measured (including infant characteristics, alpha diversity, food diversity scores, microbial age, RMM and MAZ) between independent samples at the same time point, an adequate parametric ANOVA with posthoc pairwise t-test, or a non-parametric Kruskal-Wallis test with posthoc pairwise Wilcoxon Rank Sum Test were applied to examine the differences between means (parametric) or distributions (non-parametric) of tested groups. Adequate functions from ‘rstatix’ package (v0.7.2) were used. To statistically compare differences between numerical variables for dependent samples, measured for the same infant at multiple time points, a parametric two-way repeated measure ANOVA with posthoc paired pairwise t-test or non-parametric Friedman test with posthoc Conover test was applied. We used the Kruskal-Wallis test to identify significant differences between dietary classes for each alpha diversity measure, between DMM enterotypes for each bacterial species, between time points for each fiber, between dietary classes for each fiber and between dietary classes for each KO group. Subsequently, the Dunn’s test was applied to determine which pairs of dietary classes or enterotypes differed significantly. P-values were adjusted using the Benjamini-Hochberg (BH) procedure ( 58 ), to control for false discovery rate, with a significance level set at 0.05 for alpha diversity, DMM species and fibers between time points. Non-parametric tests were applied in place of parametric versions after confirming statistical differences in variance between tested groups with Bartlett's test and rejecting the null hypothesis with the Shapiro-Wilk test. The BH correction for multiple testing was applied in each case. Unless specified otherwise, a threshold of 0.05 was used to reject the null hypothesis. BH procedure was also used for p-values in LinDA analysis (significance level set at 0.2) and in beta diversity for PCoA scores between dietary classes at each time point (significance level set at 0.05). To look for differences between IFDPs and dietary classes, the DF degradation capacity in each dietary class was compared to its degradation capacity in all other classes at a specific time point using a Welch Two-Sample t-test with BH correction (significance level set at 0.2). The significance threshold for FDR q values was chosen to be 0.2, as this allows for balancing the discovery of true positives against the control of false positives in this large-scale exploratory study. Results Solid food introduction in infants varies in timing, nature and diversity For 112 participants an extensive set of dietary data, including 46 different food items, was collected through repeated questionnaires at 4, 5, 6, 9, 11, and 14 months postpartum (Supplementary Table 1A). Excluding variables that were included only in the questionnaires at the early time-points (e.g., tea, rice cake, breadstick) or that were given to fewer than 7.5 % of the infants (buttermilk, soymilk, red beets)and aggregating certain food items from the later, more detailed, questionnaires into the broader categories (e.g., different species of fish as well as meat) as were used in the earlier questionnaires at ages 4 to 6 months (Supplementary Table 4), resulted in 28 complementary food items (Supplementary Table 5A-B). By applying logistic PCA with a longitudinal distance metric on these dietary data, three distinct dietary classes were identified (Figure 1A). The impact of individual food items on the first two principal components (PC) varied across different ages. Overall, at ages 4 and 5 months, fruits and vegetables contributed the most to both PC1 and PC2, while at later time points, animal-derived food items gained in importance (Figure 1B). Specifically, at 4 months, carrot, banana and pear had the largest impact on PC1 and PC2, whereas at 5 months cauliflower, pear, and potato had the highest summed loadings on the first two PCs. Furthermore, meat, beans and pasta became more influential at 6 months and fish, margarine and strawberry at 9 months. Finally, when the diet became more adult-like, butter, strawberry and peach had the largest impact at 11 months and butter, margarine and orange at 14 months (Figure 1B). Dietary classes exhibit variation across different ages, and these differences become more pronounced over time (Supplementary Figure 3). Next, we examined how the timing of introduction of individual food items differed between the three dietary classes. The introduction of new food items started early on for infants in Dietary Class (DC) 1 (median = 17 weeks of age, Interquartile Range (IQR) [16-18]) with the age at first introduction being significantly lower as compared to infants in DC2 (median = 20, IQR [17-24], p-value = 0.001) and DC3 (median = 18, IQR [17-23], p-value = 0.018) (Figure 1C, Supplementary Table 1B-C). While few food items had been introduced by the age of 4 months, the proportion of infants already receiving broccoli (χ2=10.273, p-value = 0.006), carrot (χ2=10.167, p-value = 0.006) and pear (χ2=6.984, p-value = 0.03) significantly differed between the dietary classes (Supplementary Tables 6-7). Across all cases, infants in DC1 most frequently received these food items — for instance, broccoli (20% in DC1 vs. 0% in DC2 and 3% in DC3), carrot (39% vs. 15% and 12), and pear (20% vs. 0% and 9%) (Figure 1D). By 6 months of age, the proportion of infants receiving these and other fruits and vegetables, including apple, banana, beans and cauliflower, gradually increases. Again, the proportion of infants in DC1 was highest and in DC2 lowest for all of these food items (all p-value < 0.05, Supplementary Table 7). By 9 months of age, more than half of the infants in DC1 were already introduced to 75% (21/28) of the measured food items, in comparison to 54% (15/28) and 61% (17/28) of food items being received by at least half of the infants in DC2 and DC3, respectively (Supplementary Table 5A-B, Supplementary Figure 4). Interestingly, while most food items were introduced more frequently in DC1, certain items—like melon and butter—were given to a higher proportion of infants in DC3 despite some – like margarine and porridge —given in lower proportions to the same infants (Figure 1D, Supplementary Figure 5-6, Supplementary Tables 6-7). In line with the earlier introduction of many food items, the diversity of items being introduced, expressed as the FVS, and to a lesser extent the DDS, was also highest among infants in DC1, whereas infants in DC2 had the lowest FVS, particularly from 6 months onwards. In general, the FVS and DDS gradually increased over time in all DCs, with the largest shift between the ages of 6 and 9 months (pairwise t-test and Wilcoxon test, adjusted p-value (q) < 0.001 for each of the dietary classes) (Supplementary Figure 7, Supplementary Table 8). The food allergen diversity (FAD, milk, egg, wheat, fish, soy and nuts) was higher among infants in DC1 as compared to DC3 at 5 (pairwise t-test, p-value = 0.019), 6 (pairwise t-test, p-value = 0.041) and 14 (pairwise t-test, p-value = 0.045) months of age. Except for modest differences in the proportion of breastfed infants at 4 months of age (Chi-square test, χ2=5.99, p-value = 0.05) and formula-fed infants at 5, 6 and 11 months, no associations were found between dietary classes and perinatal data, such as birth mode and place, gestational age, birth weight, sex, furry pet exposure and number of older siblings (Supplementary Figures 1-2). Overall, the age at introduction of food items and the diversity of food items being introduced contributed the most to the formation of the three dietary classes. DC1 had a significantly higher amount of fruits (apple, banana, kiwi, melon, orange, peach, pear, and strawberry) than DC2 at all ages, and higher than DC3 at 5 and 11 months of age (Supplementary Figure 8, Supplementary Tables 9A-9B). Therefore, we descriptively named infants in DC1 as “ early and diverse fruit introducers ”, DC2 as “ partial introducers with reduced fruit diversity” and DC3 as “ delayed introducers with frequent butter consumption” . Dietary classes associate with infant microbiome development Fecal samples, collected at 4, 5, 6, 9, 11 and 14 months of age, were available for metagenomic profiling in 105 out of the 112 infants (Supplementary Table 2A). At the age of 4 months, microbial diversity, measured by the Shannon index, was significantly higher among infants in DC1 as compared to DC2 (pairwise t-test, p-value = 0.011) and DC3 (pairwise t-test, p-value = 0.005). Moreover, these “early and diverse food introducers” also had the highest observed and estimated (Chao1) species richness indices at 4, 6 and 11 months (Figure 2A, Supplementary Table 10). No significant differences in microbial richness and diversity were observed between DC2 and DC3. Next, we examined the progression in the microbiome community structure and across dietary classes by conducting a Principal Coordinates Analysis (PCoA) on a species level Aitchison’s dissimilarity matrix (Supplementary Figure 9). Ordination along the first principal coordinate significantly differed between DC1 and DC2 (pairwise t-test, adjusted p-value = 0.027), and DC1 and DC3 (pairwise t-test, q = 0.005) at 4 months, and also between DC1 and DC3 (pairwise t-test, q = 0.002) at 5, and at 6 months (pairwise t-test, q = 0.004) (Figure 2B, Supplementary Table 11). With PCo1 scores explaining 13.8% of variance, thus capturing high variability, these results suggest distinct microbial community structures of infants across dietary classes particularly between 4 and 6 months of life. Permutational analysis of variance confirmed that dietary classes were significantly associated with the microbial community structure at 5 months of age even upon adjustment for potential confounding factors, including breastfeeding status (PERMANOVA, p-value = 0.008, Supplementary Table 12). Overall, infants in DC1 had the richest and most diverse GM, with a microbial community structure distinct from those of the other two dietary classes. RMM and MAZ were used to compare the GM development trajectories between the dietary classes and age. The age-related GM development was similar across the dietary classes, except for a slightly more mature microbiota among infants in DC1 as compared to DC3 at 4 months of age (pairwise Wilcoxon test, adjusted p-value = 0.05, Figure 2C, Supplementary Table 13). To identify compositional states (enterotypes) of the infants, and to compare the transitions between these states in different dietary classes, DMM clustering was performed. DMM clustering resulted in 6 enterotypes with different predominant bacterial species (Supplementary Figure 10, Supplementary Table 14) and significantly increasing species richness and diversity across clusters (Supplementary Figure 11, Supplementary Tables 15A-B). Enterotype (E) 1 to 5 were mainly driven by the abundance of various Bifidobacterium species, where E1 and E2 were dominated by Bifidobacterium longum , and Bifidobacterium bifidum, respectively, while a more balanced distribution of multiple Bifidobacterium species, including Bifidobacterium breve and B. longum (E3) and B. longum (E4), was observed for the other enterotypes. Faecalibacterium prausnitzii abundance increased across enterotypes and was significantly higher in E6 as compared to all of the other enterotypes (both q < 0.001). E6 was also characterized by an increased abundance of Prevotella copri and Bacteroides uniformis as compared to all other enterotypes, (q ≤ 0.05 for all) (Supplementary Tables 16A-B). With age, infants’ microbiota transitioned from the first four enterotypes to enterotypes five and six. The most profound transition is noticeable between 6 and 9 months for every dietary class (Figure 2D). Infants in DC2 had all transitioned to E6 at 14 months, while some infants in DC1 and almost half of the infants in DC3 remained in E5 until the age of 14 months. Interestingly, in DC2, the highest rate of transition from E2 to E6 occurred between 6 and 9 months of age when the highest amount of new food items were introduced. The same rapid transition was visible for DC3 at the same ages, whereas infants in DC1 had a more gradual and smoother transition between the DMM clusters, likely due to the early introduction of a wide variety of solid foods in comparison with the other two classes. Dietary classes drive the abundance of specific microbial species To further investigate these age-dependent associations of dietary classes with the relative abundance of specific taxa, differential abundance analysis was performed, including the interaction between dietary classes and age, and adjusting for relevant confounding factors such as birth weight, sex, birth mode, birthplace, duration of breastfeeding, and the age of the first introduction of solids. Age was the strongest driver of significant differences in the abundance of many bacterial species (Figure 3) with a significant decrease in many Bifidobacterium species , Veillonella parvula and Enterobacterales ( Escherichia coli , Klebsiella pneumoniae ) and a strong increase in members of the order Clostridia , including Roseburia intestinalis , Roseburia faecis , Roseburia hominis, F. prausnitzii , Blautia wexlerae, Fusicatenibacter saccharivorans, Flavonifractor plautii and Eubacterium rectale (LinDA, all q < 0.05 at 14 months, Supplementary Table 17). Additionally, many previously reported associations (6, 59-61), including a lower abundance of Bacteroides among infants born by C-section and a higher abundance of Bifidobacterium species with increasing duration of breastfeeding, were confirmed in our study (Figure 3, Supplementary Table 17). When examining the differential abundance of bacterial species among infants in the different dietary classes, most associations appeared to be age-dependent (i.e., an age*dietary class interaction) in DC3 in comparison to DC1. The exception was the age-independent decreased abundance of B. longum (LinDA, q = 0.02, Supplementary Table 17) among infants in DC3 as compared to infants in DC2. Most age-dependent differences in bacterial species abundance were observed at ages 9 to 14 months and predominantly in DC3 as compared to DC1. Here, the abundance of Bacteroides fragilis, Bifidobacterium adolescentis , B. bifidum , B. breve , Bifidobacterium kashiwanohense , B. longum , Collinsella aerofaciens , E. coli , F. plautii and V. parvula at 11 months and B. fragilis and C. aerofaciens at 9 months was significantly enriched (LinDA, all q < 0.2) in DC3 in comparison to DC1, whereas Bacteroides ovatus was significantly enriched (q = 0.126) in DC2 in comparison to DC1 at 14 months (Figure 3, Supplementary Table 17). Altogether, our results demonstrate that the impact of feeding patterns on bacterial species abundance is most pronounced from the age of 9 months onwards and, in line with the enterotype transition trajectories, reveal a stunted shift towards keystone members of the adult gut microbiome among infants in DC3. Dietary classes are linked to the microbiome’s fiber-degrading capacity Next, we inspected the capability of the infants’ microbiomes to degrade dietary fibers (DF) within and between the dietary classes. After inferring fiber degradation profiles (IFDPs), t-SNE dimensionality reduction was performed. This showed that the variation in IFDP’s was mainly driven by the infants’ age (Supplementary Figure 12). Overall, the infant’s GM fiber-degrading capacity was highest for fibers such as xanthan, xyloglucan, chitin, beta-glucan, cellulose and xylan across all time points (Supplementary Figure 13, Supplementary Table 18). Notably, the degradation capacity increased with age for gellan (Dunn’s test, 6-9 mon, q = 0.011; 9-11 mon, q = 0.001) and laminaran (9-11 mon, q = 0.001) (Supplementary Tables 19A-B). Galactoglucomannan (q = 0.040), galactomannan (q = 0.040) and dextran (q = 0.024) had a peak in degradation capacity between 6 and 9 months (Supplementary Tables 19B). Subsequently, we explored the connection between dietary classes and microbial fiber degradation potential. The T-statistic heatmap (Figure 4A, Supplementary Table 20) highlights that the microbiome of infants in DC1 had amongst others statistically significantly higher degradation capabilities of arabinoxylan, rhamnogalacturonan and arabinogalactan at ages 4, 5 and 6 months and also higher degradation capacity of xylan, pectin and arabinan at 5 and 6 months of age. The microbiome of infants in DC3 was characterized by a higher capacity to degrade chitin, laminaran, xanthan and gellan at various ages (Figure 4A, Supplementary Table 20). To identify the bacterial species driving these differences in fiber-degrading capacity of the infant microbiome in across dietary classes, we first analyzed longitudinal correlations between IFDPs and bacterial species abundances. The analysis revealed two significant positive correlation patterns between fibers rhamnogalacturonan, arabinogalactan, pectin, arabinan, arabinoxylan, xylan, xyloglucan, mannan, carrageenan, dextran, and galactomannan and bacteria B. uniformis , B. ovatus , Parabacteroides distasonis , Bacteroides vulgatus , C. aerofaciens , Bacteroides dorei , B. fragilis , P. copri , Alistipes finegoldii , Ruminococcus gnavus, and. F. plautii at 4-6 months. And between fibers laminaran, xanthan, gellan, resistant starch, inulin, and chitin and bacteria R. gnavus , V. parvula, and E.coli at 9-14 months (Supplementary Figure 14, Supplementary Table 21). Next, we investigated degradation capacity of the bacterial communities. For that we selected species with a statistically significant (q 0.3) with a given fiber at least at two time points. Then, we summed the compositionally transformed reads for all selected bacteria capable of potentially degrading that specific fiber (Supplementary Table 22). Out of 15 fiber degradation profiles, 10 profiles correlated with more than one bacterial species based upon our selection criteria. Afterwards, we examined potential differences in the abundance of inferred fiber degrading bacteria between dietary classes at each time point (Supplementary Table 23A). At the age of 4 months, the total abundance of bacteria associated with arabinoxylan degradation differed between DC1-DC2 (Dunn’s test, q = 0.024), with the degradation of rhamnogalacturonan DC1-DC3 (Dunn’s test, q = 0.044) and xylan DC1-DC2 (Dunn’s test, q = 0.046) and DC1-DC3 (Dunn’s test, q = 0.046) (Figure 4B, Supplementary Table 23B). At 5 months the abundance of bacteria associated with rhamnogalacturonan degradation differed between DC2-DC3 (Dunn’s test, q = 0.049), with dextran, galactoglucomannan, galactomannan between DC1-DC3 (Dunn’s test, q = 0.027 for all). Finally, at 6 months with xylan degradation between DC1-DC3 (Dunn’s test, q = 0.036). Given the observed differences in fiber degradation capacity among infants in the three dietary classes, we next explored whether these shifts might translate into altered metabolic potential. We focused on xylan-degrading microbes at 4 months of age because xylan, although more abundant in grains, broccoli, carrots, and cauliflower, is also present in apples (62), pears (63) and strawberries (64) — foods introduced significantly earlier and more frequently in DC1 than in DC2 and DC3. Indeed, xylan-degrading capacity at 4 months was notably enriched in DC1 (Figure 4B), and correlated strongly (q < 0.001) with F. plautii (Supplementary Table 21), reflecting this species’ metabolic potential to utilize xylan at this early age. Timing and diversity of complementary feeding can modulate butyrate production capacity To further assess whether these microbiome differences might influence SCFA production, we compared the gene abundances for key enzymes involved in butyrate, acetate, and propionate biosynthesis at 4 months (Supplementary Tables 24A-B). HUMAnN detected genes belonging to 49 KO identifiers related to the metabolism of propionate (n = 25), acetate (n = 19), and butyrate (n = 10). Among these, five identifiers were associated with two SCFAs. Strikingly, the buk gene (encoding butyrate kinase; K00929) was significantly enriched in DC1 relative to both DC2 and DC3 (Figure 4C), indicating an enhanced potential for butyrate formation in infants exposed to earlier and more diverse complementary foods. Several additional butyrate-related genes, including croR (3-hydroxybutyryl-CoA dehydratase; K17865), abfD (4-hydroxybutyryl-CoA dehydratase, K14534), hbd (3-hydroxybutyryl-CoA dehydrogenase; K00074), mcmA1 (methylmalonyl-CoA mutase, N-terminal domain; K01848), were similarly elevated in DC1 (Figure 4C). Taxon distribution analyses pointed to F. plautii as the principal carrier of these genes at 4 months (Figure 4D), followed by Acidaminococcus intestini , consistent with their documented role in butyrate fermentation (65-67). Intriguingly, however, F. plautii peaked later in DC3 at 11 months (Figure 3, Supplementary Table 17), while in DC1, F. prausnitzii rapidly expanded from 9 to 14 months and dominated butyrate production pathways at older ages. This dynamic suggests that, although F. plautii initially contributes substantially to butyrate-forming potentials in infants who consume xylan-rich vegetables early, it is eventually supplanted by F. prausnitzii as the diet diversifies further, particularly in DC1. Even though both acetate and propionate are commonly converted upon fiber degradation in the intestine, no substantial differences were observed between the three dietary classes (Supplementary Table 24A-B). Overall, our findings highlight how the timing and diversity of complementary food introduction can modulate infant gut fiber metabolism and, by extension, the capacity for butyrate production, potentially influencing immune and metabolic maturation in early life. Discussion This study aimed to explore the influence of feeding patterns on the development of the infant’s GM community, taking into account confounding factors, and to examine its association with the degradation capacity of various dietary fibers into e.g. SCFA’s. We identified three DCs based on the timing and types of solid foods introduced. Infants in DC1 received solid foods earlier and had a more diverse intake of fruits; therefore, this group was labeled 'early and diverse fruit introducers ’, subsequently DC2 was labeled “ partial introducers with reduced fruit diversity ” and DC3 “ delayed introducers with frequent butter consumption ”, due to higher consumption of butter, cheese, and milk that do not contain dietary fibers. Early introducers had higher early-life microbial diversity and richness, and their microbiome structure between 4 and 6 months of age differed significantly from the delayed introducers. The same trend was reported in a Danish cohort ( 68 ), that observed lower alpha and beta diversity in infants receiving solids later on in life. However, in the present study the effect of the early introduction of the solids diminished over time, and by 14 months, it was no longer distinguishable in the GM alpha or beta diversity. Early introducers also had more mature microbiome than delayed introducers. Although the exact timing of how and when this is essential for the healthy development of infants is still unclear, previous studies have linked immature GM to negative health outcomes, e.g. asthma development later in life ( 54 )( 69 ). Interestingly, our findings contrast with findings from other studies ( 6 , 70 , 71 ) by indicating that the type and timing of the introduction of solid foods, rather than the cessation of breast milk, primarily drives the development of the infant's GM, which was also confirmed by Laursen et al. ( 68 ). Unlike those earlier studies—some of which used more limited dietary questionnaires or aggregated complementary feeding data, and primarily relied on 16S rRNA sequencing—our approach integrated comprehensive longitudinal metagenomic profiling with detailed dietary diversity information to capture the nuanced impacts of specific food items on gut microbiome maturation. Despite the age being the main dominant driver of microbiome shifts, delayed introducers showed delayed microbial maturation, resulting in higher levels of early-life bacteria at later age compared to early introducers ( 72 ). The only non-age-driven difference in species abundance was B. longum , higher in partial introducers as compared to delayed introducers. Thus, our study contributes valuable evidence for refining current guidelines on the introduction of solid food. Due to earlier introduction of the solids, earlier introducers had a smoother transition between the enterotypes. This finding supports previous research indicating that the development of infants' GM is influenced by the introduction of solid foods ( 7 , 68 , 73 ). Nevertheless, by age 14 months, only partial introducers were fully transitioned to enterotype 6, characterized by the highest abundance of bacteria F. prausnitzii , a bacterium which typically increases with age, notably after 6 months ( 74 , 75 ). In contrast, delayed introducers exhibited lower abundance of this bacterium compared to other classes. It’s worth noting that F. prausnitzii depletion is associated with increase in inflammatory responses, and related with atopic dermatitis in infants ( 76 ) and with inflammatory bowel disease in adults ( 77 , 78 ). The feeding patterns observed in delayed introducers resulted in decreased abundance of this bacterium, suggesting a potential risk for AD development in the future ( 27 ). However, reverse causation cannot be excluded during this time window- given that atopic dermatitis commonly manifests within the first 1–2 years of life ( 79 ). Changes in gut community structure triggered by different types of diet, will lead towards different community functionality, in our study manifested in DF degradation capacity. Earlier introducers had better capacity in degrading several DF (arabinan containing fibers, pectin, and xylan) at earlier ages, while delayed introducer had higher capacity at various ages for xanthan, gellan, laminaran ( 80 ) and chitin despite the main source of these fibers like seaweed, kombu, and gluten-free or vegan alternatives for gelatine are unlikely to be given to the infants. The latter could possibly be explained by the fact that structurally similar polysaccharides often require the same enzymatic machinery for degradation, allowing these microbes to respond to familiar substrates, e.g. chitin and cellulose ( 81 ). Moreover, some of the more exotic fibers — such as gellan and xanthan are widely used as food additive ( 82 , 83 ). Chitin, in comparison, is a component of fungal cell walls ( 84 ), and infant exposure to airborne fungal spores or soil microbes may also shape the GM’s fiber-degrading capabilities. In our study, chitin degradation capability was positive correlated with E. coli and R. gnavus in contrast to a previous finding ( 85 ). Additionally, early introducers had greater xylan-degrading capacity at 4 months compared to the other DCs, which was strongly linked to F. plautii . F. plautii increased with age, and was the highest in delayed introducers at the later ages. This bacterium is known to suppress Th2 immune responses in mice ( 86 ), indicating possible anti-inflammatory properties. Moreover, early introducers had significant enrichment of the buk gene, indicating higher butyrate production potential as early as at 4 months. Our study demonstrates that introducing diverse fiber-containing foods as early as 4–5 months leads to the development of a versatile GM community capable of producing butyrate, which supports the findings of Differding et al. ( 87 ). Interestingly, F. plautii is capable of butyrogenic lysine degradation ( 88 ), and this pathway is more prominent at 4 months in early and diverse introducers. Thus, our study revealed that a diet rich in fruits and vegetables promotes F. plautii , which, in turn, potentially enhances butyrate production. This is in alignment with findings in a synthetic gut bacterial community ( 89 ). Overall, this indicates that early and diverse complementary feeding enhances SCFA-producing potential, particularly for butyrate, potentially supporting immune and metabolic development in infancy. Moving forward, it is essential to investigate infant gut metabolites to understand how theoretical dietary fiber degradation capacity profiles correspond to gut metabolome profiles. Furthermore, it is important to consider its immunological response in the gut, e.g. Nwaru et al. ( 90 ) demonstrated that introduction of solid foods already at 3 months may protect high-risk infants against atopic sensitization. Strengths and limitations A major strength of this study lies in its integration of dense, longitudinal dietary data with high-resolution metagenomic sequencing across a critical developmental window (4–14 months). This allowed us to go beyond taxonomic composition and directly assess functional potential, such as fiber-degrading and butyrate-producing capacity, which are rarely examined in such detail during early infancy. Moreover, the richness of the dietary data—capturing both timing and diversity of specific food introductions—enabled us to define biologically meaningful dietary patterns with strong associations to microbial maturation. Our methodological approach, combining logistic PCA with dynamic time warping and clustering, provided nuanced insight into real-life dietary trajectories and their microbial consequences. While our approach offers novel insights, some limitations warrant consideration. Quantification of food intake was not feasible as validated instruments are lacking for this age group as it is difficult to estimate due to frequent spitting up and regurgitation, which usually resolve around 12 months of age ( 91 ). As such, our dietary analysis focused on the introduction (yes or no) of individual food items. In our clustering approach, we performed separate binary principal component analyses (PCA) at each time-point, and then used the resulting PC scores to compare dietary trajectories using dynamic time warping. A limitation of this method is that the principal component axes inferred from each time-point may represent somewhat different aspects of the underlying dietary information. Consequently, the PC scores are not guaranteed to be on a common scale or orientation, and their use as a longitudinal time series may introduce minor distortions in trajectory comparison that could affect downstream clustering. Despite this potential limitation, we observed biologically plausible and interpretable groupings, suggesting that our approach captured meaningful variation in complementary feeding patterns. Future studies may consider comparing alternative approaches, such as performing PCA on pooled data across time-points to obtain time-invariant component structures. Nonetheless, we believe our analysis offers valuable insights into early dietary development and provides a basis for further exploration. Conclusion Taken together, these results underscore how early complementary feeding decisions, even as early as 4 months, can shape the infant gut’s metabolic capacity. Introducing xylan-containing grains, vegetables and fruits at this age appears to foster butyrate-producing taxa, including F. plautii . Over time, F. prausnitzii eventually predominates in the microbiome of children who were fed rapidly diversifying diets. These shifts in fiber-degrading and butyrate-producing capabilities may significantly impact immune development and long-term metabolic health, highlighting the need for further research into optimal early-life dietary strategies to support microbiome maturation. Abbreviations BH Benjamini-Hochberg DC Dietary class DDS Dietary Diversity Score DF Dietary Fibers DMM Dirichlet Multinomial Mixture DTW Dynamic Time Warping FAD Food Allergen Diversity FVS Food Variety Score GM Gut microbiota IFDP Inferred Fiber Degradation Profiles IQR Interquartile Range LinDA LINear models for Differential Abundance analysis MAZ Microbiota-for-Age Z-score MDS Multidimensional Scaling PC Principal Components PCA Principal Component Analysis PCoA Principal Coordinates Analysis RMM Relative microbiota maturity SCFA Short Chain Fatty Acids t-SNE t-Distributed Stochastic Neighbour Embedding Declarations Ethics approval and consent to participate Written informed consent was obtained from both parents/legal caregivers prior to enrolment in the study. This research confirmed to the principles of the Helsinki Declaration. Ethical approval was obtained by the Medical Ethical Committee of Maastricht University Medical Center (study number: METC-15-4-237 ). Consent for publication Not applicable. Availability of data and materials Trimmed and quality filtered reads, with removed human reads can be found at European Nucleotide Archive under project number PRJEB89491. The metadata are not publicly available due to the potentially identifiable nature of the data and privacy concerns by study participants, but are available upon reasonable request from the corresponding authors (NvB and MM). Competing interests No potential conflict of interest was reported by the author(s). Funding The LucKi Gut study was funded by a grant from The Netherlands Organization for Health Research and Development (ZonMw) through the European Union Joint Programming Initiative—A Healthy Diet for a Healthy Life (received by J.P. and M.M.; project #: 529051010). Authors' contributions Conceptualization and design: L.B., J.P., N.v.B. and M.M.; Study coordination: L.B.; Data collection: L.B., M.M., and E.D.; Methodology: E.D., M.S., L.B., J.P., N.v.B. and M.M.; Formal analysis: E.D., M.S., G.L., D.B., G.G., C.D.; Data curation: M.M., and E.D.; writing—original draft preparation: E.D., M.S., N.v.B. and J.P.; Writing—review and editing, all; supervision: N.v.B. and J.P.; project administration: C.D., M.M., and E.D.; funding acquisition: M.M. and J.P. All authors have read and agreed to the published version of the manuscript. Acknowledgements The authors would like to sincerely thank all families for their participation during this study, as well as the participating midwifery practices in the Limburg area of the Netherlands. We also would like to thank members of the John Penders lab for their help with this study. ChatGPT (OpenAI) was used to assist with language editing and improving grammar in this manuscript. The authors take full responsibility for the content and interpretation of the results. Authors' information (optional) References Penders J, Thijs C, Vink C, Stelma FF, Snijders B, Kummeling I, et al. Factors influencing the composition of the intestinal microbiota in early infancy. Pediatrics. 2006;118(2):511-21. Penders J, Gerhold K, Thijs C, Zimmermann K, Wahn U, Lau S, et al. New insights into the hygiene hypothesis in allergic diseases: mediation of sibling and birth mode effects by the gut microbiota. Gut microbes. 2014;5(2):239-44. Jokela R, Korpela K, Jian C, Dikareva E, Nikkonen A, Saisto T, et al. Quantitative insights into effects of intrapartum antibiotics and birth mode on infant gut microbiota in relation to well-being during the first year of life. Gut Microbes. 2022;14(1):2095775. Korpela K. Impact of delivery mode on infant gut microbiota. Annals of Nutrition and Metabolism. 2021;77(Suppl. 3):11-9. Korpela K, Helve O, Kolho K-L, Saisto T, Skogberg K, Dikareva E, et al. Maternal fecal microbiota transplantation in cesarean-born infants rapidly restores normal gut microbial development: a proof-of-concept study. Cell. 2020;183(2):324-34. e5. Stewart CJ, Ajami NJ, O’Brien JL, Hutchinson DS, Smith DP, Wong MC, et al. Temporal development of the gut microbiome in early childhood from the TEDDY study. Nature. 2018;562(7728):583-8. Homann C-M, Rossel CA, Dizzell S, Bervoets L, Simioni J, Li J, et al. Infants’ first solid foods: impact on gut microbiota development in two intercontinental cohorts. Nutrients. 2021;13(8):2639. Parajuli A, Grönroos M, Siter N, Puhakka R, Vari HK, Roslund MI, et al. Urbanization reduces transfer of diverse environmental microbiota indoors. Frontiers in microbiology. 2018;9:84. Panzer AR, Sitarik AR, Fadrosh D, Havstad SL, Jones K, Davidson B, et al. The impact of prenatal dog keeping on infant gut microbiota development. Clinical & Experimental Allergy. 2023. Liu W, Sun Z, Ma C, Zhang J, Ma C, Zhao Y, et al. Exposure to soil environments during earlier life stages is distinguishable in the gut microbiome of adult mice. Gut Microbes. 2021;13(1):1830699. Olivares M, Neef A, Castillejo G, De Palma G, Varea V, Capilla A, et al. The HLA-DQ2 genotype selects for early intestinal microbiota composition in infants at high risk of developing coeliac disease. Gut. 2015;64(3):406-17. Catassi G, Aloi M, Giorgio V, Gasbarrini A, Cammarota G, Ianiro G. The Role of Diet and Nutritional Interventions for the Infant Gut Microbiome. Nutrients. 2024;16(3):400. Wilson AS, Koller KR, Ramaboli MC, Nesengani LT, Ocvirk S, Chen C, et al. Diet and the human gut microbiome: an international review. Digestive diseases and sciences. 2020;65:723-40. Voreades N, Kozil A, Weir TL. Diet and the development of the human intestinal microbiome. Frontiers in microbiology. 2014;5:494. Avershina E, Lundgård K, Sekelja M, Dotterud C, Storrø O, Øien T, et al. Transition from infant‐to adult‐like gut microbiota. Environmental microbiology. 2016;18(7):2226-36. Jian C, Carpén N, Helve O, de Vos WM, Korpela K, Salonen A. Early-life gut microbiota and its connection to metabolic health in children: Perspective on ecological drivers and need for quantitative approach. EBioMedicine. 2021;69. Yang I, Corwin EJ, Brennan PA, Jordan S, Murphy JR, Dunlop A. The infant microbiome: implications for infant health and neurocognitive development. Nursing research. 2016;65(1):76-88. Meng Q, Luo Z, Cao C, Sun S, Ma Q, Li Z, et al. Weaning alters intestinal gene expression involved in nutrient metabolism by shaping gut microbiota in pigs. Frontiers in microbiology. 2020;11:694. Al Nabhani Z, Dulauroy S, Marques R, Cousu C, Al Bounny S, Déjardin F, et al. A weaning reaction to microbiota is required for resistance to immunopathologies in the adult. Immunity. 2019;50(5):1276-88. e5. Bailén M, Bressa C, Martínez-López S, González-Soltero R, Montalvo Lominchar MG, San Juan C, et al. Microbiota features associated with a high-fat/low-fiber diet in healthy adults. Frontiers in nutrition. 2020;7:583608. Fritsch J, Garces L, Quintero MA, Pignac-Kobinger J, Santander AM, Fernández I, et al. Low-fat, high-fiber diet reduces markers of inflammation and dysbiosis and improves quality of life in patients with ulcerative colitis. Clinical Gastroenterology and Hepatology. 2021;19(6):1189-99. e30. Barber C, Mego M, Sabater C, Vallejo F, Bendezu RA, Masihy M, et al. Differential effects of western and mediterranean-type diets on gut microbiota: A metagenomics and metabolomics approach. Nutrients. 2021;13(8):2638. Bowyer RC, Jackson MA, Pallister T, Skinner J, Spector TD, Welch AA, et al. Use of dietary indices to control for diet in human gut microbiota studies. Microbiome. 2018;6:1-11. Del Chierico F, Vernocchi P, Dallapiccola B, Putignani L. Mediterranean diet and health: food effects on gut microbiota and disease control. International journal of molecular sciences. 2014;15(7):11678-99. García-Vega ÁS, Corrales-Agudelo V, Reyes A, Escobar JS. Diet quality, food groups and nutrients associated with the gut microbiota in a nonwestern population. Nutrients. 2020;12(10):2938. Lalli MK, Salo TE, Hakola L, Knip M, Virtanen SM, Vatanen T. Associations between dietary fibers and gut microbiome composition in the EDIA longitudinal infant cohort. The American Journal of Clinical Nutrition. 2025;121(1):83-99. Galazzo G, van Best N, Bervoets L, Dapaah IO, Savelkoul PH, Hornef MW, et al. Development of the microbiota and associations with birth mode, diet, and atopic disorders in a longitudinal analysis of stool samples, collected from infancy through early childhood. Gastroenterology. 2020;158(6):1584-96. Okada H, Kuhn C, Feillet H, Bach J-F. The ‘hygiene hypothesis’ for autoimmune and allergic diseases: an update. Clinical & Experimental Immunology. 2010;160(1):1-9. Ahrens AP, Hyötyläinen T, Petrone JR, Igelström K, George CD, Garrett TJ, et al. Infant microbes and metabolites point to childhood neurodevelopmental disorders. Cell. 2024;187(8):1853-73. e15. Fang C, Zhong H, Lin Y, Chen B, Han M, Ren H, et al. Assessment of the cPAS-based BGISEQ-500 platform for metagenomic sequencing. Gigascience. 2018;7(3):gix133. Mölder F, Jablonski KP, Letcher B, Hall MB, Tomkins-Tinch CH, Sochat V, et al. Sustainable data analysis with Snakemake. F1000Research. 2021;10. Chen S, Zhou Y, Chen Y, Gu J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics. 2018;34(17):i884-i90. Langmead B, Salzberg SL. Fast gapped-read alignment with Bowtie 2. Nature methods. 2012;9(4):357-9. Danecek P, Bonfield JK, Liddle J, Marshall J, Ohan V, Pollard MO, et al. Twelve years of SAMtools and BCFtools. Gigascience. 2021;10(2):giab008. Beghini F, McIver LJ, Blanco-Míguez A, Dubois L, Asnicar F, Maharjan S, et al. Integrating taxonomic, functional, and strain-level profiling of diverse microbial communities with bioBakery 3. elife. 2021;10:e65088. Louis P, Flint HJ. Formation of propionate and butyrate by the human colonic microbiota. Environmental microbiology. 2017;19(1):29-41. Reichardt N, Duncan SH, Young P, Belenguer A, McWilliam Leitch C, Scott KP, et al. Phylogenetic distribution of three pathways for propionate production within the human gut microbiota. The ISME journal. 2014;8(6):1323-35. Singh V, Lee G, Son H, Koh H, Kim ES, Unno T, et al. Butyrate producers,“The Sentinel of Gut”: Their intestinal significance with and beyond butyrate, and prospective use as microbial therapeutics. Frontiers in microbiology. 2023;13:1103836. Zhang B, Lingga C, Bowman C, Hackmann TJ. A new pathway for forming acetate and synthesizing ATP during fermentation in bacteria. Applied and environmental microbiology. 2021;87(14):e02959-20. Team RDC. R: A language and environment for statistical computing. (No Title). 2010. Landgraf AJ, Lee Y. Dimensionality reduction for binary data through the projection of natural parameters. Journal of Multivariate Analysis. 2020;180:104668. Giorgino T. Computing and visualizing dynamic time warping alignments in R: the dtw package. Journal of statistical Software. 2009;31:1-24. Galili T. dendextend: an R package for visualizing, adjusting and comparing trees of hierarchical clustering. Bioinformatics. 2015;31(22):3718-20. Wickham H, Wickham H. Data analysis: Springer; 2016. Powell C. CGPfunctions: powell miscellaneous functions for teaching and learning statistics. R package version 06. 2020;3. Rathnayake KM, Madushani P, Silva K. Use of dietary diversity score as a proxy indicator of nutrient adequacy of rural elderly people in Sri Lanka. BMC research notes. 2012;5:1-6. Lee BR, Jung H-I, Kim SK, Kwon M, Kim H, Jung M, et al. Dietary Diversity during early infancy increases microbial diversity and prevents egg allergy in high-risk infants. Immune Network. 2022;22(2). Jari Oksanen F, Friendly M, Kindt R, Legendre P, McGlinn D, Minchin PR, et al. Vegan: community ecology package. R package version. 2018;2(6). Barnett DJ, Arts IC, Penders J. microViz: an R package for microbiome data visualization and statistics. Journal of Open Source Software. 2021;6(63):3201. Gu Z, Eils R, Schlesner M. Complex heatmaps reveal patterns and correlations in multidimensional genomic data. Bioinformatics. 2016;32(18):2847-9. Holmes I, Harris K, Quince C. Dirichlet multinomial mixtures: generative models for microbial metagenomics. PloS one. 2012;7(2):e30126. Morgan M. DirichletMultinomial: Dirichlet-Multinomial Mixture Model Machine Learning for Microbiome Data. 2021. p. R package. Zhou H, He K, Chen J, Zhang X. LinDA: linear models for differential abundance analysis of microbiome compositional data. Genome biology. 2022;23(1):95. Subramanian S, Huq S, Yatsunenko T, Haque R, Mahfuz M, Alam MA, et al. Persistent gut microbiota immaturity in malnourished Bangladeshi children. Nature. 2014;510(7505):417-21. Liaw A, Wiener M. Classification and regression by randomForest. R news. 2002;2(3):18-22. Cohen Y, Borenstein E. The microbiome’s fiber degradation profile and its relationship with the host diet. BMC biology. 2022;20(1):266. Buchfink B, Reuter K, Drost H-G. Sensitive protein alignments at tree-of-life scale using DIAMOND. Nature methods. 2021;18(4):366-8. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal statistical society: series B (Methodological). 1995;57(1):289-300. Shao Y, Forster SC, Tsaliki E, Vervier K, Strang A, Simpson N, et al. Stunted microbiota and opportunistic pathogen colonization in caesarean-section birth. Nature. 2019;574(7776):117-21. Reyman M, van Houten MA, van Baarle D, Bosch AA, Man WH, Chu MLJ, et al. Impact of delivery mode-associated gut microbiota dynamics on health in the first year of life. Nature communications. 2019;10(1):4997. Matharu D, Ponsero AJ, Dikareva E, Korpela K, Kolho K-L, de Vos WM, et al. Bacteroides abundance drives birth mode dependent infant gut microbiota developmental trajectories. Frontiers in Microbiology. 2022;13:953475. Borujeni NE, Alavijeh MK, Denayer JF, Karimi K. A novel integrated biorefinery approach for apple pomace valorization with significant socioeconomic benefits. Renewable Energy. 2023;208:275-86. Jermyn M, Isherwood F. Changes in the cell wall of the pear during ripening. Biochemical Journal. 1956;64(1):123. Hotchkiss AT, Chau HK, Strahan GD, Nuñez A, Harron A, Simon S, et al. Structural characterization of strawberry pomace. Heliyon. 2024;10(9). Jumas-Bilak E, Carlier J-P, Jean-Pierre H, Mory F, Teyssier C, Gay B, et al. Acidaminococcus intestini sp. nov., isolated from human clinical samples. International journal of systematic and evolutionary microbiology. 2007;57(10):2314-9. Cardoso BB, Amorim C, Franco-Duarte R, Alves JI, Barbosa SG, Silvério SC, et al. Epilactose as a promising butyrate-promoter prebiotic via microbiota modulation. Life. 2024;14(5):643. Cui J, Ramesh G, Wu M, Jensen ET, Crago O, Bertoni AG, et al. Butyrate-producing bacteria and insulin homeostasis: the microbiome and insulin longitudinal evaluation study (MILES). Diabetes. 2022;71(11):2438-46. Laursen MF, Andersen LB, Michaelsen KF, Mølgaard C, Trolle E, Bahl MI, et al. Infant gut microbiota development is driven by transition to family foods independent of maternal obesity. Msphere. 2016;1(1):10.1128/msphere. 00069-15. Stokholm J, Blaser MJ, Thorsen J, Rasmussen MA, Waage J, Vinding RK, et al. Maturation of the gut microbiome and risk of asthma in childhood. Nature communications. 2018;9(1):141. Bäckhed F, Roswall J, Peng Y, Feng Q, Jia H, Kovatcheva-Datchary P, et al. Dynamics and stabilization of the human gut microbiome during the first year of life. Cell host & microbe. 2015;17(5):690-703. Bergström A, Skov TH, Bahl MI, Roager HM, Christensen LB, Ejlerskov KT, et al. Establishment of intestinal microbiota during early life: a longitudinal, explorative study of a large cohort of Danish infants. Applied and environmental microbiology. 2014;80(9):2889-900. Laursen MF, Sakanaka M, von Burg N, Mörbe U, Andersen D, Moll JM, et al. Bifidobacterium species associated with breastfeeding produce aromatic lactic acids in the infant gut. Nature microbiology. 2021;6(11):1367-82. Laursen MF. Gut microbiota development: influence of diet from infancy to toddlerhood. Annals of Nutrition and Metabolism. 2021;77(Suppl. 3):21-34. Robertson RC, Manges AR, Finlay BB, Prendergast AJ. The human microbiome and child growth–first 1000 days and beyond. Trends in microbiology. 2019;27(2):131-47. Koga Y, Tokunaga S, Nagano J, Sato F, Konishi K, Tochio T, et al. Age-associated effect of kestose on Faecalibacterium prausnitzii and symptoms in the atopic dermatitis infants. Pediatric Research. 2016;80(6):844-51. Candela M, Rampelli S, Turroni S, Severgnini M, Consolandi C, De Bellis G, et al. Unbalance of intestinal microbiota in atopic children. BMC microbiology. 2012;12:1-9. Sokol H, Seksik P, Furet J, Firmesse O, Nion-Larmurier I, Beaugerie L, et al. Low counts of Faecalibacterium prausnitzii in colitis microbiota. Inflammatory bowel diseases. 2009;15(8):1183-9. Machiels K, Joossens M, Sabino J, De Preter V, Arijs I, Eeckhaut V, et al. A decrease of the butyrate-producing species Roseburia hominis and Faecalibacterium prausnitzii defines dysbiosis in patients with ulcerative colitis. Gut. 2014;63(8):1275-83. Illi S, von Mutius E, Lau S, Nickel R, Grüber C, Niggemann B, et al. The natural course of atopic dermatitis from birth to age 7 years and the association with asthma. Journal of Allergy and Clinical Immunology. 2004;113(5):925-31. Sharma PP, Baskaran V. Polysaccharide (laminaran and fucoidan), fucoxanthin and lipids as functional components from brown algae (Padina tetrastromatica) modulates adipogenesis and thermogenesis in diet-induced obesity in C57BL6 mice. Algal research. 2021;54:102187. Latańska I, Kolesińska B, Draczyński Z, Sujka W. The use of chitin and chitosan in manufacturing dressing materials. Progress on Chemistry and Application of Chitin and its Derivatives. 2020(25):16-36. Do MH, Lee HHL, Lee J-E, Park M, Oh M-J, Lee H-B, et al. Gellan gum prevents non-alcoholic fatty liver disease by modulating the gut microbiota and metabolites. Food Chemistry. 2023;400:134038. Ostrowski MP, La Rosa SL, Kunath BJ, Robertson A, Pereira G, Hagen LH, et al. Mechanistic insights into consumption of the food additive xanthan gum by the human gut microbiota. Nature microbiology. 2022;7(4):556-69. Lopez-Santamarina A, Mondragon AdC, Lamas A, Miranda JM, Franco CM, Cepeda A. Animal-origin prebiotics based on chitin: An alternative for the future? a critical review. Foods. 2020;9(6):782. Rodriguez J, Neyrinck AM, Zhang Z, Seethaler B, Nazare J-A, Robles Sánchez C, et al. Metabolite profiling reveals the interaction of chitin-glucan with the gut microbiota. Gut Microbes. 2020;12(1):1810530. Ogita T, Yamamoto Y, Mikami A, Shigemori S, Sato T, Shimosato T. Oral administration of Flavonifractor plautii strongly suppresses Th2 immune responses in mice. Frontiers in immunology. 2020;11:379. Differding MK, Benjamin-Neelon SE, Hoyo C, Østbye T, Mueller NT. Timing of complementary feeding is associated with gut microbiota diversity and composition and short chain fatty acid concentrations over the first year of life. BMC microbiology. 2020;20:1-13. Shetty SA, Kuipers B, Atashgahi S, Aalvink S, Smidt H, de Vos WM. Inter-species metabolic interactions in an in-vitro minimal human gut microbiome of core bacteria. npj Biofilms and Microbiomes. 2022;8(1):21. Weiss AS, Niedermeier LS, von Strempel A, Burrichter AG, Ring D, Meng C, et al. Nutritional and host environments determine community ecology and keystone species in a synthetic gut bacterial community. Nature Communications. 2023;14(1):4780. Nwaru BI, Takkinen HM, Niemelä O, Kaila M, Erkkola M, Ahonen S, et al. Introduction of complementary foods in infancy and atopic sensitization at the age of 5 years: timing and food diversity in a Finnish birth cohort. Allergy. 2013;68(4):507-16. Hegar B, Dewanti NR, Kadim M, Alatas S, Firmansyah A, Vandenplas Y. Natural evolution of regurgitation in healthy infants. Acta Paediatrica. 2009;98(7):1189-93. Additional Declarations No competing interests reported. Supplementary Files SupplementaryFigures.pdf SupplementaryTables.xlsx 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-6673441","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":465235212,"identity":"b48d65e2-d7c5-49c3-9b58-d1f5dc33d216","order_by":0,"name":"Evgenia Dikareva","email":"","orcid":"","institution":"Maastricht University Medical Centre+","correspondingAuthor":false,"prefix":"","firstName":"Evgenia","middleName":"","lastName":"Dikareva","suffix":""},{"id":465235214,"identity":"86ab9ec2-125d-4e37-b143-b820b811c808","order_by":1,"name":"Michał Skawiński","email":"","orcid":"","institution":"Maastricht University","correspondingAuthor":false,"prefix":"","firstName":"Michał","middleName":"","lastName":"Skawiński","suffix":""},{"id":465235219,"identity":"97504cdf-4c87-46cb-96be-9cb3533b3ac5","order_by":2,"name":"Liene Bervoets","email":"","orcid":"","institution":"KU Leuven","correspondingAuthor":false,"prefix":"","firstName":"Liene","middleName":"","lastName":"Bervoets","suffix":""},{"id":465235222,"identity":"40ea02ee-d458-4a31-9d70-71b1f8e114bb","order_by":3,"name":"David Barnett","email":"","orcid":"","institution":"Maastricht University Medical Centre+","correspondingAuthor":false,"prefix":"","firstName":"David","middleName":"","lastName":"Barnett","suffix":""},{"id":465235223,"identity":"080ea841-1939-4478-b562-dde781cc4dcc","order_by":4,"name":"Gianluca Galazzo","email":"","orcid":"","institution":"Maastricht University Medical Centre+","correspondingAuthor":false,"prefix":"","firstName":"Gianluca","middleName":"","lastName":"Galazzo","suffix":""},{"id":465235226,"identity":"7d0a3dbe-3a04-4225-bd2a-616103cde362","order_by":5,"name":"Giang Le","email":"","orcid":"","institution":"Biomedical Primate Research Centre","correspondingAuthor":false,"prefix":"","firstName":"Giang","middleName":"","lastName":"Le","suffix":""},{"id":465235227,"identity":"5bd5cbe5-af93-4f28-8446-3634d04563f9","order_by":6,"name":"Christel Driessen","email":"","orcid":"","institution":"Maastricht University Medical Centre+","correspondingAuthor":false,"prefix":"","firstName":"Christel","middleName":"","lastName":"Driessen","suffix":""},{"id":465235228,"identity":"caac0897-d986-475a-a2dd-f4d92aeacb7a","order_by":7,"name":"John Penders","email":"","orcid":"","institution":"Maastricht University Medical Centre+","correspondingAuthor":false,"prefix":"","firstName":"John","middleName":"","lastName":"Penders","suffix":""},{"id":465235234,"identity":"ca46be49-4a23-471a-8a87-a6360613a27d","order_by":8,"name":"Monique Mommers","email":"","orcid":"","institution":"Maastricht University","correspondingAuthor":false,"prefix":"","firstName":"Monique","middleName":"","lastName":"Mommers","suffix":""},{"id":465235236,"identity":"48724aca-a6e1-4629-8954-869046efa7b7","order_by":9,"name":"Niels Best","email":"data:image/png;base64,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","orcid":"","institution":"Maastricht University Medical Centre+","correspondingAuthor":true,"prefix":"","firstName":"Niels","middleName":"","lastName":"Best","suffix":""}],"badges":[],"createdAt":"2025-05-15 14:08:20","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6673441/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6673441/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":84184860,"identity":"e3ff9482-769e-49ae-af82-5025711880d9","added_by":"auto","created_at":"2025-06-09 05:11:05","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":215446,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eLongitudinal dietary classes of infants from 4 to 14 months of age.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA\u003c/strong\u003e. Dendrogram of clustering results for the longitudinal dietary classes of infants. The colors of the branches represent different dietary classes.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eB\u003c/strong\u003e. Barplots of logistic PCA models loadings, indicating the impact of food items on PC1 and PC2 at different time points.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eC\u003c/strong\u003e. Violin plots with the distribution of ages when solid food were first introduced. Colors represent dietary classes, and significant differences are displayed over brackets with adjusted p-values.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eD\u003c/strong\u003e. Proportions of infants being fed specific food items within each of the dietary classes over time. Only food items significantly different between dietary classes are presented (n=18).\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-6673441/v1/00673cf40c397e88add510dd.png"},{"id":84184859,"identity":"65fc545c-8864-45c7-9706-3a274f8ec504","added_by":"auto","created_at":"2025-06-09 05:11:05","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":73934,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMicrobial maturation trajectories of infants within each of the dietary classes.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA\u003c/strong\u003e. Microbial richness and diversity (Shannon diversity, Species richness, Chao1 estimated richness) for infants within each dietary class per time point. Statistically significant differences between dietary classes are indicated by adjusted p-values above the box-plots.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eB\u003c/strong\u003e. Changes in infants’ gut microbiota profiles over time within each dietary class. Box-plots represent scores of the first PCoA component based upon species-level Aitchison’s distance. Statistically significant differences between dietary classes are indicated by adjusted p-values above the box-plots.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eC\u003c/strong\u003e. Random Forest regression model’s predictions of microbiota maturation. Predicted microbial age in days with fitted loess regression line as ”baseline” between chronological and microbial age. Calculated from the previous parameters relative microbiota maturity and MAZ score calculated for each sample. Baseline is indicated with a dashed horizontal line\u003cstrong\u003e.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eD\u003c/strong\u003e. Infants' microbiota transition across DMM clusters in each dietary class over time. The line thickness represents the transition frequency, and circle color represents the enterotypes (y axes), from the darkest color (enterotype 1) to light blue (enterotype 6). The size of the circle represents the number of infants in that particular cluster per time point.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-6673441/v1/e33a9748907289a78a28681e.png"},{"id":84184863,"identity":"faa8f18f-d2ad-4113-920c-fe9556a5d9f8","added_by":"auto","created_at":"2025-06-09 05:11:05","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":551016,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eLongitudinal changes in the abundance of bacterial species in each of the classes.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA\u003c/strong\u003e. Heatmap of fold change in abundance of most abundant species derived from differentiation abundance analysis. Models were adjusted for sex, duration of breastfeeding, age of introduction of solid foods, birth weight, delivery place, and delivery type. The significant p-values after adjustment with Benjamin-Hochberg procedure are marked with \"**\" for \u0026lt;0.05, “*” between 0.05 and 0.2. Ref. 1 indicates Class 1 was compered to classes 2 and 3, and ref. 2 indicates that Class 2 to compared to classes 1 and 3.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-6673441/v1/b4d451900e910d492c845357.png"},{"id":84184864,"identity":"5a1baac7-68bf-4e31-ac7a-f843a68108f9","added_by":"auto","created_at":"2025-06-09 05:11:05","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":409901,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eDietary Class Influence on Microbial Degradation Capacity and Functional Potential\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eA\u003c/strong\u003e. The heatmap illustrates T-statistics derived from t-tests, which compare the degradation capacity of DFs within each dietary class to their capacity across two other classes at a particular time point. The significant p-values after adjustment with Benjamini-Hochberg procedure are marked with \"**\" for \u0026lt;0.05, “*” between 0.05 and 0.2.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eB. \u003c/strong\u003eStacked barplots displaying the relative abundance of significant fiber degraders at 4 months of age, with asterisks denoting significant differences between dietary classes in fiber degradation capacity.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eC. \u003c/strong\u003eBoxplot of log10-transformed read counts across diet classes for different KEGG codes. Each boxplot colored according to the respective diet class. Statistical significance is assessed using Dunn’s test, with adjusted p-values displayed above the boxplots. Non-significant comparisons are hidden for clarity.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eD. \u003c/strong\u003eStacked bar plot of relative abundance across diet classes for butyrate KEGG codes at 4 months. The stacked bars represent contributions from the top 10 species. Statistical significance is assessed using Dunn’s test, with adjusted p-values displayed above the bars. Non-significant comparisons are hidden for clarity.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-6673441/v1/ad4deaa55c9240de99d67383.png"},{"id":94824933,"identity":"3a2919bd-ac4a-4aa9-8540-54d03976fa90","added_by":"auto","created_at":"2025-10-31 06:49:35","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2655530,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6673441/v1/ac368d4a-4083-457b-b04a-a795d0d99b95.pdf"},{"id":84184862,"identity":"e60b2234-04a3-4727-89b3-9124d32a3c2b","added_by":"auto","created_at":"2025-06-09 05:11:05","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":1181019,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryFigures.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6673441/v1/2fb57e4663fa98869d04c82a.pdf"},{"id":84184865,"identity":"b6a1bb4f-1a19-4f33-a6f1-3571211b7326","added_by":"auto","created_at":"2025-06-09 05:11:05","extension":"xlsx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":775506,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryTables.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-6673441/v1/ca0e9aefc4475aef0863b27a.xlsx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Dietary impact on infants' gut microbiota and its capacity in SCFA metabolism: a longitudinal Dutch cohort study","fulltext":[{"header":"Introduction","content":"\u003cp\u003eInfancy represents a critical period for gut microbiome development, which is pivotal for immune system maturation and overall long-term health. The maturation of the infant gut microbiome is a dynamic process that undergoes rapid transformations influenced, for instance, by mode of delivery (\u003cspan additionalcitationids=\"CR2 CR3 CR4 CR5\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e), type of milk being fed (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e) and the time and type of complementary food introduction (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e). While several other factors, including environmental exposures (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan additionalcitationids=\"CR9\" citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e) and host genetics (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e) contribute to the shaping of the microbial community, diet stands out as the most profound driver of microbial assembly and maturation in early life (\u003cspan additionalcitationids=\"CR13\" citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe transition from a milk-based diet to solid foods heralds a shift in the microbial maturation trajectories, eventually resulting in an adult-like gut microbiota (GM) composition at around 2 years of age (\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e). During this transition, the diversity of bacterial taxa capable of metabolizing complex carbohydrates expands and impacts metabolic (\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e) and immune responses (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e). Animal studies have shown that weaning impacts the development of immune responses and host metabolism, with effects lasting into adulthood (\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e). Several dietary studies have indicated that the diversity of foods consumed may lead to significant variations in the composition of the adult GM (\u003cspan additionalcitationids=\"CR21 CR22 CR23\" citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e). On the contrary, only a few human studies have looked into the effects of individual dietary components (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e). To our knowledge, only one previous study investigated associations between specific food items or patterns and microbiota maturation during the weaning period (\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e). With evidence accumulating for a link between early-life microbial imbalances and later health outcomes such as allergies (\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e), autoimmune diseases (\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e) and neurodevelopmental disorders (\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e), there is an urgent need to delve into these dietary influences that shape the infant gut microbiome. Investigating these influences during the critical developmental period is vital, given their potential to uncover preventative strategies aimed at mitigating the risk of later-life health conditions.\u003c/p\u003e \u003cp\u003eHere, we explore early-life feeding patterns and infants\u0026rsquo; microbiota development and maturation within the LucKi-Gut cohort. Longitudinally collected (4, 5, 6, 9, 11, and 14 months of age) fecal samples, profiled by whole metagenome shotgun sequencing, were combined with dietary data. We showed that not only the type of solid foods, but also the time of introduction and food item diversity, shapes the infant microbiota composition and functional capacity.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStudy population and design\u003c/h2\u003e \u003cp\u003eThe LucKi Gut Study is an ongoing longitudinal birth cohort study aimed at monitoring microbiome development during infancy and early childhood. Pregnant women living in the South Limburg region of the Netherlands were recruited through obstetric practices, gynaecology departments, during lactation information sessions, and via advertisements at venues for pregnancy yoga, baby clothes stores and social media. Infants born before 36 weeks of gestation were excluded from the present analyses. Infant fecal samples were collected at age 1\u0026ndash;2 weeks and 1, 2, 4, 5, 6, 9, 11, and 14 months postpartum. Parents were instructed to collect the infant\u0026rsquo;s fecal samples from the diaper and immediately freeze the samples at -20⁰C in special transport containers (Sarstedt, Hilden, Germany) in their home freezer. All samples were transported to the laboratory of the department of Medical Microbiology, Infectious Diseases and Infection Prevention at Maastricht University Medical Centre and were aliquoted and frozen at -80\u0026deg;C until further use. At every fecal sampling time-point, parents were additionally asked to complete a questionnaire. Questionnaires collected data on infants\u0026rsquo; diet, health and developmental status, medication use, as well as on maternal health (during pregnancy), medication use, lifestyle and diet. The questionnaires included detailed questions on the type of infant feeding and the introduction of complementary foods.\u003c/p\u003e \u003cp\u003eFor the present study, we included 112 infants for whom, at time of analyses, dietary data were available for the first 14 months of life (Supplementary Table\u0026nbsp;1A-C). Of these 112 infants, microbial profiling data were available for 105 infants. We restricted our analysis to fecal samples collected between 4 and 14 months of age (n\u0026thinsp;=\u0026thinsp;389), as complementary food was introduced the earliest around 4 months of age (Supplementary Table\u0026nbsp;2A-B).\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eShotgun metagenomics sequencing\u003c/h3\u003e\n\u003cp\u003eTotal metagenomic DNA was extracted by mechanical lysis on a FastPrep-96 homogenizer (MP Biomedicals, USA) followed by DNA purification using the MagPure Stool DNA Kit (Magen Biotechnology Ltd, China). Sequencing was performed on the BGISEQ-500 platform following the standard protocol (\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e) at MGI (Riga, Latvia). To standardize the pipeline, the workflow manager Snakemake v5.14.0 was used (\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e). The pre-processing comprised quality control with Fastp v0.20.1 (quality phred score: 15; minimal read length, 60 bp) (\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e). In addition, Fastp was used to trim the BGI-SEQ adapters \u0026ldquo;AAGTCGGAGGCCAAGCGGTCTTAGGAAGACAA\u0026rdquo; for forward and \u0026ldquo;AAGTCGGATCGTAGCCATGTCGTTCTGTGAGCCAAGGAGTTG\u0026rdquo; for reverse reads. Human reads were removed by aligning against chm13.draft_v1.0_plusY (downloaded 14.10.2020) with Bowtie 2 v2.3.5.1 (maximum insert size 600 bp) (\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e) and Samtools v.1.9 (\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e). Subsequently, the forward and reverse reads were used to identify taxonomic composition using MetaPhlAn v3.0 (species-markers database from January 2019 CHOCOPhlAn v30: mpa_v30_CHOCOPhlAn_201901) (\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e) by aligning reads to the reference database of marker genes.\u003c/p\u003e\n\u003ch3\u003eKEGG-based annotation of short-chain fatty acids (SCFAs) production capacity\u003c/h3\u003e\n\u003cp\u003eWe used HUMAnN (v3.9) (\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e) with the Enzyme Commission (EC) filtered UniRef90 database with default search settings to quantify features of the functional capacity of each sample that are relevant to short chain fatty acid metabolism. UniRef features were regrouped into KEGG orthology (KO) terms using the \u0026ldquo;humann_regroup_table\u0026rdquo; utility script provided by HUMAnN.\u003c/p\u003e \u003cp\u003eRelevant KO identifiers (n\u0026thinsp;=\u0026thinsp;65, Supplementary Table\u0026nbsp;3)(\u003cspan additionalcitationids=\"CR37 CR38\" citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e) for SCFA production were manually selected from the KEGG database.\u003c/p\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eData analysis\u003c/h2\u003e \u003cp\u003eThe analysis unfolded in three sequential steps: (\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e) identification of groups of infants exhibiting similar longitudinal feeding patterns and separating them into distinct dietary classes, (\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e) exploration of the differential effects of identified classes on microbiome development and (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e) determination of significant relations between dietary classes and bacterial taxa, as well as between dietary classes, bacterial taxa and dietary fiber degrading capacity. All analyses were conducted using R version 4.1.2 (\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eMissing values (yes/no) for food items, breastfeeding and formula feeding were imputed based on the answers from the adjacent time points. If this was not possible, the mode across all answers at the given time point was used. The following variables were used: delivery type [vaginal/C-section], delivery place [at home/hospital], sex [male/female], presence of furry pets [none/at home and/or at daycare/solely at daycare], formula [yes/no] and breastfeeding [yes/no] at sample collection, older siblings [none/1/2/3], duration of breastfeeding (in months), age when solids were introduced (in weeks), gestational age (in weeks) and birth weight (in grams).\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eIdentification and characterization of dietary classes\u003c/h3\u003e\n\u003cp\u003eA list covering regularly used food items was included in each questionnaire. At an infant\u0026rsquo;s age of 4 to 6 months, \u0026lsquo;bread\u0026rsquo;, \u0026lsquo;meat\u0026rsquo; and \u0026lsquo;fish\u0026rsquo; were separate yes/no items, followed by an open-ended question in which the type of bread, meat or fish, respectively, was entered by the parents. From age 9 months onwards, all food items were asked about separately (yes/no answer option). Food items from all questionnaires were harmonized and merged when too few infants consumed individual food items. For this reason, meat, chicken, pork and beef were merged into one category \u0026ndash; \u0026ldquo;meat\u0026rdquo;. Also, bread, white bread, brown bread and wholegrain bread were merged into \u0026ldquo;bread\u0026rdquo;, as between 4\u0026ndash;6 months parents not always specified the type of bread given.\u003c/p\u003e \u003cp\u003eFor some food items, closed questions were only available at later time-points, including pudding (from 6 months onwards), melon, peach, tomato, cheese, butter, margarine, egg and soy products (from 9 months onwards). For these food items, \u0026ldquo;no\u0026rdquo; was imputed for the earlier time-points, unless these food items were listed by parents in the open-ended question (separate options for \u0026ldquo;Other fruit\u0026rdquo;,\u0026rdquo;Other vegetables\u0026rdquo; and \u0026rdquo;something else namely\u0026rdquo;) in which case \u0026ldquo;yes\u0026rdquo; was imputed for the listed food item. For all other food items, data were available for all time points.\u003c/p\u003e \u003cp\u003eA variation of Principal Component Analysis (PCA), namely logistic PCA (logistic PCA v0.2) (\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e), was applied to all dietary data at each time point to reduce the dimensionality. With \u003cem\u003ecv.lpca\u003c/em\u003e function, the value of parameter m (m\u0026thinsp;=\u0026thinsp;3 for each model) was estimated, necessary for the approximation of the natural parameter. Resulting PC scores are the linear combinations of those parameters, projected from the Bernoulli saturated model. Each infant was represented as a sequence (of length 6) of PC scores (1 and 2) for each subsequent time point, which allowed tracking the change of infants\u0026rsquo; diets with time in the PC subspace. The differences between infants were quantified with the multivariate dynamic time warping (DTW) distance with \u0026lsquo;dtw\u0026rsquo; package (v1.23-1) (\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e). This similarity metric accounts for the speed of a change and aims to find an optimal match between input time series. Groups of longitudinal feeding patterns (referred to as dietary classes) and their hierarchical representation were further identified with a hierarchical clustering algorithm, with complete linkage and DTW distance matrix. Dendrogram visualization was constructed with \u0026lsquo;dendextend\u0026rsquo; package (v1.17.1) (\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eTo characterize dietary classes, proportions of infants who received a given solid food, aggregated per dietary class, were plotted for each time point in the form of stacked bar plots with \u003cem\u003eggplot2\u003c/em\u003e package (v3.4.4) (\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e). Only food items with significantly different proportions (\u003cem\u003eprop.test\u003c/em\u003e, p-value cutoff at 0.05) between dietary classes were plotted. To additionally visualize changes in the aforementioned proportions over time, slope graphs were constructed using \u003cem\u003enewggslopegraph\u003c/em\u003e function from \u0026lsquo;CGPfunctions\u0026rsquo; package (v0.6.3) (\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e). Moreover, the time when a solid food was commonly introduced in each dietary class was calculated. A food item was considered to be widely introduced when more than 50% of the infants in a dietary class received that food item at the respective time point, but not before. To assess the diversity of the infants\u0026rsquo; complementary food, various scores were calculated for each time point per infant. Food Variety Score (FVS) (\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e) reflects the number of unique food items introduced to the infant between 4\u0026ndash;14 months (maximum of 28). Dietary Diversity Score (DDS) (\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e) reflects the number of unique food groups introduced. To this end, food items were grouped into eight groups representing vegetables, fruits, legumes/nuts, meat, fish, eggs, dairy, and grains as was described previously. Finally, the Food Allergen Diversity (FAD) was calculated as an index that accounts for the number of allergens introduced (milk, egg, wheat, fish, soy, nuts) up to a maximum of 6 (\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eClustering of the infants into dietary classes was performed independently of their environmental and lifestyle data. However, to exclude the effect of certain environmental and lifestyle factors on these dietary classes, we explored their variations between the classes. For instance, the effect of breast- and formula-feeding was compared longitudinally as proportions of infants in a dietary class, who were breast or formula-fed at each time point. Linear regression models for breast and formula-feeding were separately fitted between proportion and time (numerical) with interaction between time and dietary class with 95% confidence intervals.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eExploration of the differential effects of dietary classes on microbiome development\u003c/h2\u003e \u003cdiv id=\"Sec9\" class=\"Section3\"\u003e \u003ch2\u003eAlpha and beta diversity calculations\u003c/h2\u003e \u003cp\u003eThe ecological alpha diversity metrics for microbiota communities (observed number of species, Shannon and Chao1 indices) were calculated for each sample and aggregated by dietary class and time point with function \u003cem\u003especnumber\u003c/em\u003e, \u003cem\u003ediversity\u003c/em\u003e and \u003cem\u003eestimateR\u003c/em\u003e from \u0026lsquo;vegan\u0026rsquo; package (v2.6-4) (\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e), respectively. Both boxplots and loess regression lines fitted between days from sample collection and diversity metrics, using \u003cem\u003egeom_smooth\u003c/em\u003e layer of the \u003cem\u003eggplot\u003c/em\u003e function, were plotted.\u003c/p\u003e \u003cp\u003eBetween-sample microbiota diversity (beta diversity) was investigated by performing the Principal Coordinates Analysis (PCoA) of Aitchison\u0026rsquo;s dissimilarity indices on a species level using all samples together. Progression of microbiome community structure with age was represented as a boxplot of the first principal coordinate (PCo) scores, coloured by dietary class across time points. PCoA was performed using the \u0026lsquo;microViz\u0026rsquo; package (v0.11.0) (\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eMicrobial composition was determined by compositional transformation of the read count data at both the genus and species levels for each sample using the \u0026lsquo;microViz\u0026rsquo; package. Taxa with an abundance greater than 1% were retained for further analysis. Following this selection, the mean relative taxa abundance was calculated for all dietary classes at each time point. Compositional heatmaps were constructed using the \u0026lsquo;ComplexHeatmap\u0026rsquo; package (v2.18.0) (\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e). Furthermore, loess regression lines were fitted between days from the sample collection and the relative abundance of taxa (sample-wise), for each dietary class using geom_smooth with default parameters from the \u0026lsquo;ggplot2\u0026rsquo; package.\u003c/p\u003e \u003cp\u003e \u003cb\u003eDirichlet Multinomial Mixture\u003c/b\u003e (\u003cb\u003eDMM) clustering\u003c/b\u003e\u003c/p\u003e \u003cp\u003eThe progression of microbiome communities with age was further compared across dietary classes by constructing Dirichlet Multinomial Mixture (DMM) transition graphs (\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e). DMM clustering was conducted using the \u0026lsquo;DirichletMultinomial\u0026rsquo; package (v1.44.0) (\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e), and the number of resulting clusters, known as enterotypes or \u0026lsquo;metacommunites\u0026rsquo; was determined using the Laplace approximation score. Samples that exhibited values that differed by more than 41 days from their respective time points were filtered out before the analysis. The transition of infants through DMM clusters with age was visualized separately for each dietary class, as previously described (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e\n\u003ch3\u003eDifferential abundance analysis\u003c/h3\u003e\n\u003cp\u003eTo identify bacterial taxa (genus and species level) that differed over time between the dietary classes, linear models for differential abundance analysis (LinDA) were conducted for each taxon using the \u0026lsquo;LinDA\u0026rsquo; package (v0.1.0) (\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e). LinDA uses centred log2-ratio transformed taxon abundances (read counts) as the response variable. An interaction between time points (categorical, with the first time point as reference level) and dietary classes (with the first class as reference) was included as a predictor variable. Additionally, duration of breastfeeding (numeric, months), age of the introduction of solid foods (numeric, weeks), birth weight (numeric, grams), sex (male as reference), delivery place (home/hospital with hospital as reference), and delivery type (C-section/vaginal with vaginal delivery as reference) were used as confounders. Infants\u0026rsquo; random intercepts were modelled for each infant by specifying Infant IDs as a grouping factor. Taxa with a prevalence of more than 5% and an abundance greater than 1% for each time point were retained for the analysis.\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eMicrobial maturation\u003c/h2\u003e \u003cp\u003eMicrobial age was calculated as initially described by Subramanian and colleagues (\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e). The Random Forest Cross-Validation regression was performed using the \u003cem\u003erfcv\u003c/em\u003e function from the \u0026lsquo;randomForest\u0026rsquo; package (v4.7-1.1) (\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e). As training data, a set of 66% of the metagenomic read counts, balanced with respect to the dietary classes, was used. The response variable for regression was the age of sample collection. To predictive performance cross-validation error was visualized for each tested model. The best model had the minimal number of top-ranking age-discriminatory taxa. The importance of top-ranking taxa was assessed using metrics such as an increase in model squared error and node purity. Next, the sparse model of top-ranking taxa was used for microbial age prediction, with 100-cross validation of train/test split (66%/33%). The mean across all predictions was computed for each sample, representing the final predicted microbial age.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eRelative microbiota maturity (RMM) and microbiota-for-age Z-score\u003c/h2\u003e \u003cp\u003eRelative microbiota maturity (RMM) and the microbiota-for-age Z-score (MAZ) were computed following the methodology outlined in the aforementioned publication (\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e). The results were visualized using boxplots for each time point and dietary class. To establish a reference level for microbiota maturity, a fitted loess regression line was calculated, which represents the average change in microbiota maturity with age. In contrast, the references for the MAZ were calculated as the means of all samples' microbial ages at each specific time point. Statistical differences between dietary classes for microbial age, RMM and MAZ were calculated using the method described in Statistical analysis. To investigate the relationship between the RMM and time (as a multi-level categorical variable), a linear mixed-effect model was fitted with random intercepts for each child and an interaction term between dietary classes and time points with \u0026lsquo;lme4\u0026rsquo; (v1.1-35.1) package. The significance of the coefficients was tested with \u0026lsquo;lmerTest\u0026rsquo; package (v3.1-3). Adjustments were made for confounders, including sex, delivery type, delivery place, number of older siblings, presence of furry pets, birth weight, duration of breastfeeding, and age at introduction of solid foods. The regression results were visualized using a heatmap.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eDetermination of significant relations between dietary classes, bacterial taxa and dietary fiber degradation capacities\u003c/h2\u003e \u003cdiv id=\"Sec14\" class=\"Section3\"\u003e \u003ch2\u003eInferred fiber degradation profiles\u003c/h2\u003e \u003cp\u003eTo infer the metabolic capabilities of a microbiome community to degrade dietary fibers (DFs), a functional analysis of the microbiome\u0026rsquo;s inferred fiber degradation profiles (IFDP) was performed on metagenomic samples. These profiles reflect the relative capacity of the metagenomic community to degrade specific dietary fibers. A previously described computational framework (\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e) was used to connect metagenomic sequences with annotated polysaccharide-degrading enzymes and the chemical structures of dietary fibers. It has been shown that IFDPs are linked to the host diet and accurately reflect dietary classes. To calculate the IFDPs, contigs\u0026rsquo; sequences for each sample were mapped to the functional database containing fiber-degrading enzymes\u0026rsquo; amino acid sequences, using \u003cem\u003eDiamond blastx\u003c/em\u003e (v.2.1.8) (\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e). Further, the mapped counts were multiplied by the enzyme-fiber interaction matrix to obtain the IFDPs, which were used for further analysis. t-Distributed Stochastic Neighbour Embedding (t-SNE) dimensionality reduction was performed on all relative DF degradation capabilities (IFDPs), expressed in percentages, to examine the overall sources of variation. The heatmap for DF, time point and diet was created using mean relative DF degradation capacity for each dietary class at a given time point.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eCorrelations\u003c/h2\u003e \u003cp\u003eCorrelations between the degradation capacities for each DF and bacterial species were analysed at each time point using the \u0026lsquo;stats\u0026rsquo; package (v4.1.3). The 'cor.test' function was employed for paired correlation analysis using Spearman\u0026rsquo;s method, without exact p-value computation. Taxa with an average abundance of \u0026gt;\u0026thinsp;1% at any time point were included for these analyses. We focused on correlations with an estimate\u0026thinsp;\u0026gt;\u0026thinsp;0.3, deemed statistically significant (q\u0026thinsp;\u0026lt;\u0026thinsp;0.05), and meeting criteria across at least two time points. This resulted in identifying 15 fibers and 19 bacterial species. For each of the 15 fibers, the total sum of relative reads of all degraders was calculated. The same selected bacterial species were used in Spearman correlations with food items.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eTo statistically examine the associations between dietary classes and numerical variables measured (including infant characteristics, alpha diversity, food diversity scores, microbial age, RMM and MAZ) between independent samples at the same time point, an adequate parametric ANOVA with posthoc pairwise t-test, or a non-parametric Kruskal-Wallis test with posthoc pairwise Wilcoxon Rank Sum Test were applied to examine the differences between means (parametric) or distributions (non-parametric) of tested groups. Adequate functions from \u0026lsquo;rstatix\u0026rsquo; package (v0.7.2) were used. To statistically compare differences between numerical variables for dependent samples, measured for the same infant at multiple time points, a parametric two-way repeated measure ANOVA with posthoc paired pairwise t-test or non-parametric Friedman test with posthoc Conover test was applied.\u003c/p\u003e \u003cp\u003eWe used the Kruskal-Wallis test to identify significant differences between dietary classes for each alpha diversity measure, between DMM enterotypes for each bacterial species, between time points for each fiber, between dietary classes for each fiber and between dietary classes for each KO group. Subsequently, the Dunn\u0026rsquo;s test was applied to determine which pairs of dietary classes or enterotypes differed significantly. P-values were adjusted using the Benjamini-Hochberg (BH) procedure (\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e), to control for false discovery rate, with a significance level set at 0.05 for alpha diversity, DMM species and fibers between time points.\u003c/p\u003e \u003cp\u003eNon-parametric tests were applied in place of parametric versions after confirming statistical differences in variance between tested groups with Bartlett's test and rejecting the null hypothesis with the Shapiro-Wilk test. The BH correction for multiple testing was applied in each case. Unless specified otherwise, a threshold of 0.05 was used to reject the null hypothesis.\u003c/p\u003e \u003cp\u003eBH procedure was also used for p-values in LinDA analysis (significance level set at 0.2) and in beta diversity for PCoA scores between dietary classes at each time point (significance level set at 0.05).\u003c/p\u003e \u003cp\u003eTo look for differences between IFDPs and dietary classes, the DF degradation capacity in each dietary class was compared to its degradation capacity in all other classes at a specific time point using a Welch Two-Sample t-test with BH correction (significance level set at 0.2). The significance threshold for FDR q values was chosen to be 0.2, as this allows for balancing the discovery of true positives against the control of false positives in this large-scale exploratory study.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cp\u003e\u003cstrong\u003e\u003cem\u003eSolid food introduction in infants varies in timing, nature and diversity\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor 112 participants an extensive set of dietary data, including 46 different food items, was collected through repeated questionnaires at 4, 5, 6, 9, 11, and 14 months postpartum (Supplementary Table 1A). Excluding variables that were included only in the questionnaires at the early time-points (e.g., tea, rice cake, breadstick) or that were given to fewer than 7.5 % of the infants (buttermilk, soymilk, red beets)and aggregating certain food items from the later, more detailed, questionnaires into the broader categories (e.g., different species of fish as well as meat) as were used in the earlier questionnaires at ages 4 to 6 months (Supplementary Table 4), resulted in 28 complementary food items (Supplementary Table 5A-B).\u003c/p\u003e\n\u003cp\u003eBy applying logistic PCA with a longitudinal distance metric on these dietary data, three distinct dietary classes were identified (Figure 1A). The impact of individual food items on the first two principal components (PC) varied across different ages. Overall, at ages 4 and 5 months, fruits and vegetables contributed the most to both PC1 and PC2, while at later time points, animal-derived food items gained in importance (Figure 1B). Specifically, at 4 months, carrot, banana and pear had the largest impact on PC1 and PC2, whereas at 5 months cauliflower, pear, and potato had the highest summed loadings on the first two PCs. Furthermore, meat, beans and pasta became more influential at 6 months and fish, margarine and strawberry at 9 months. Finally, when the diet became more adult-like, butter, strawberry and peach had the largest impact at 11 months and butter, margarine and orange at 14 months (Figure 1B). Dietary classes exhibit variation across different ages, and these differences become more pronounced over time (Supplementary Figure 3).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eNext, we examined how the timing of introduction of individual food items differed between the three dietary classes. The introduction of new food items started early on for infants in Dietary Class (DC) 1 (median = 17 weeks of age, Interquartile Range (IQR) [16-18]) with the age at first introduction being significantly lower as compared to infants in DC2 (median = 20, IQR [17-24], p-value = 0.001) and DC3 (median = 18, IQR [17-23], p-value = 0.018) (Figure 1C, Supplementary Table 1B-C).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWhile few food items had been introduced by the age of 4 months, the proportion of infants already receiving broccoli (\u0026chi;2=10.273, p-value\u003cem\u003e\u0026nbsp;\u003c/em\u003e= 0.006), carrot (\u0026chi;2=10.167, p-value = 0.006) and pear (\u0026chi;2=6.984, p-value = 0.03) significantly differed between the dietary classes (Supplementary Tables 6-7). Across all cases, infants in DC1 most frequently received these food items \u0026mdash; for instance, broccoli (20% in DC1 vs. 0% in DC2 and 3% in DC3), carrot (39% vs. 15% and 12), and pear (20% vs. 0% and 9%) (Figure 1D). By 6 months of age, the proportion of infants receiving these and other fruits and vegetables, including apple, banana, beans and cauliflower, gradually increases. Again, the proportion of infants in DC1 was highest and in DC2 lowest for all of these food items (all p-value \u0026lt; 0.05, Supplementary Table 7). By 9 months of age, more than half of the infants in DC1 were already introduced to 75% (21/28) of the measured food items, in comparison to 54% (15/28) and 61% (17/28) of food items being received by at least half of the infants in DC2 and DC3, respectively (Supplementary Table 5A-B, Supplementary Figure 4). Interestingly, while most food items were introduced more frequently in DC1, certain items\u0026mdash;like melon and butter\u0026mdash;were given to a higher proportion of infants in DC3 despite some \u0026ndash; like margarine and porridge \u0026mdash;given in lower proportions to the same infants (Figure 1D, Supplementary Figure 5-6, Supplementary Tables 6-7).\u003c/p\u003e\n\u003cp\u003eIn line with the earlier introduction of many food items, the diversity of items being introduced, expressed as the FVS, and to a lesser extent the DDS, was also highest among infants in DC1, whereas infants in DC2 had the lowest FVS, particularly from 6 months onwards. In general, the FVS and DDS gradually increased over time in all DCs, with the largest shift between the ages of 6 and 9 months (pairwise t-test and Wilcoxon test, adjusted p-value (q) \u0026lt; 0.001 for each of the dietary classes) (Supplementary Figure 7, Supplementary Table 8).\u003c/p\u003e\n\u003cp\u003eThe food allergen diversity (FAD, milk, egg, wheat, fish, soy and nuts) was higher among infants in DC1 as compared to DC3 at 5 (pairwise t-test, p-value = 0.019), 6 (pairwise t-test, p-value = 0.041) and 14 (pairwise t-test, p-value = 0.045) months of age. Except for modest differences in the proportion of breastfed infants at 4 months of age (Chi-square test, \u0026chi;2=5.99, p-value = 0.05) and formula-fed infants at 5, 6 and 11 months, no associations were found between dietary classes and perinatal data, such as birth mode and place, gestational age, birth weight, sex, furry pet exposure and number of older siblings (Supplementary Figures 1-2).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eOverall, the age at introduction of food items and the diversity of food items being introduced contributed the most to the formation of the three dietary classes. DC1 had a significantly higher amount of fruits (apple, banana, kiwi, melon, orange, peach, pear, and strawberry) than DC2 at all ages, and higher than DC3 at 5 and 11 months of age (Supplementary Figure 8, Supplementary Tables 9A-9B).\u003c/p\u003e\n\u003cp\u003eTherefore, we descriptively named infants in DC1 as \u0026ldquo;\u003cstrong\u003eearly and diverse fruit introducers\u003c/strong\u003e\u0026rdquo;, DC2 as \u0026ldquo;\u003cstrong\u003epartial introducers with reduced fruit diversity\u0026rdquo;\u0026nbsp;\u003c/strong\u003eand DC3 as \u0026ldquo;\u003cstrong\u003edelayed introducers with frequent butter consumption\u0026rdquo;\u003c/strong\u003e.\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eDietary classes associate with infant microbiome development\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFecal samples, collected at 4, 5, 6, 9, 11 and 14 months of age, were available for metagenomic profiling in 105 out of the 112 infants (Supplementary Table 2A). At the age of 4 months, microbial diversity, measured by the Shannon index, was significantly higher among infants in DC1 as compared to DC2 (pairwise t-test, p-value\u003cem\u003e\u0026nbsp;\u003c/em\u003e= 0.011) and DC3 (pairwise t-test,\u003cem\u003e\u0026nbsp;\u003c/em\u003ep-value\u003cem\u003e\u0026nbsp;\u003c/em\u003e= 0.005). Moreover, these \u0026ldquo;early and diverse food introducers\u0026rdquo; also had the highest observed and estimated (Chao1) species richness indices at 4, 6 and 11 months (Figure 2A, Supplementary Table 10). No significant differences in microbial richness and diversity were observed between DC2 and DC3. Next, we examined the progression in the microbiome community structure and across dietary classes by conducting a Principal Coordinates Analysis (PCoA) on a species level Aitchison\u0026rsquo;s dissimilarity matrix (Supplementary Figure 9). Ordination along the first principal coordinate significantly differed between DC1 and DC2 (pairwise t-test, adjusted p-value\u003cem\u003e\u0026nbsp;\u003c/em\u003e= 0.027), and DC1 and DC3 (pairwise t-test, q = 0.005) at 4 months, and also between DC1 and DC3 (pairwise t-test, q = 0.002) at 5, and at 6 months (pairwise t-test, q = 0.004) (Figure 2B, Supplementary Table 11). With PCo1 scores explaining 13.8% of variance, thus capturing high variability, these results suggest distinct microbial community structures of infants across dietary classes particularly between 4 and 6 months of life. Permutational analysis of variance confirmed that dietary classes were significantly associated with the microbial community structure at 5 months of age even upon adjustment for potential confounding factors, including breastfeeding status (PERMANOVA, p-value\u003cem\u003e\u0026nbsp;\u003c/em\u003e= 0.008, Supplementary Table 12). Overall, infants in DC1 had the richest and most diverse GM, with a microbial community structure distinct from those of the other two dietary classes.\u003c/p\u003e\n\u003cp\u003eRMM and MAZ were used to compare the GM development trajectories between the dietary classes and age. The age-related GM development was similar across the dietary classes, except for a slightly more mature microbiota among infants in DC1 as compared to DC3 at 4 months of age (pairwise Wilcoxon test, adjusted p-value\u003cem\u003e\u0026nbsp;\u003c/em\u003e= 0.05, Figure 2C, Supplementary Table 13).\u003c/p\u003e\n\u003cp\u003eTo identify compositional states (enterotypes) of the infants, and to compare the transitions between these states in different dietary classes, DMM clustering was performed. DMM clustering resulted in 6 enterotypes with different predominant bacterial species (Supplementary Figure 10, Supplementary Table 14) and significantly increasing species richness and diversity across clusters (Supplementary Figure 11, Supplementary Tables 15A-B). Enterotype (E) 1 to 5 were mainly driven by the abundance of various \u003cem\u003eBifidobacterium\u003c/em\u003e species, where E1 and E2 were dominated by \u003cem\u003eBifidobacterium longum\u003c/em\u003e, and \u003cem\u003eBifidobacterium bifidum,\u0026nbsp;\u003c/em\u003erespectively, while a more balanced distribution of multiple \u003cem\u003eBifidobacterium\u003c/em\u003e species, including \u003cem\u003eBifidobacterium breve and B. longum\u003c/em\u003e (E3) and \u003cem\u003eB. longum\u0026nbsp;\u003c/em\u003e(E4), was observed for the other enterotypes. \u003cem\u003eFaecalibacterium prausnitzii\u003c/em\u003e abundance increased across enterotypes and was significantly higher in E6 as compared to all of the other enterotypes \u0026nbsp;(both q \u0026lt; 0.001). E6 was also characterized by an increased abundance of \u003cem\u003ePrevotella copri\u003c/em\u003e and \u003cem\u003eBacteroides uniformis\u003c/em\u003e as compared to all other enterotypes, (q \u0026le; 0.05 for all) (Supplementary Tables 16A-B). With age, infants\u0026rsquo; microbiota transitioned from the first four enterotypes to enterotypes five and six. The most profound transition is noticeable between 6 and 9 months for every dietary class (Figure 2D). Infants in DC2 had all transitioned to E6 at 14 months, while some infants in DC1 and almost half of the infants in DC3 remained in E5 until the age of 14 months. Interestingly, in DC2, the highest rate of transition from E2 to E6 occurred between 6 and 9 months of age when the highest amount of new food items were introduced. The same rapid transition was visible for DC3 at the same ages, whereas infants in DC1 had a more gradual and smoother transition between the DMM clusters, likely due to the early introduction of a wide variety of solid foods in comparison with the other two classes.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eDietary classes drive the abundance of specific microbial species\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo further investigate these age-dependent associations of dietary classes with the relative abundance of specific taxa, differential abundance analysis was performed, including the interaction between dietary classes and age, and adjusting for relevant confounding factors such as birth weight, sex, birth mode, birthplace, duration of breastfeeding, and the age of the first introduction of solids. Age was the strongest driver of significant differences in the abundance of many bacterial species (Figure 3) with a significant decrease in many \u003cem\u003eBifidobacterium\u0026nbsp;\u003c/em\u003especies\u003cem\u003e, Veillonella parvula\u0026nbsp;\u003c/em\u003eand \u003cem\u003eEnterobacterales\u003c/em\u003e (\u003cem\u003eEscherichia coli\u003c/em\u003e, \u003cem\u003eKlebsiella\u0026nbsp;pneumoniae\u003c/em\u003e) and a strong increase in members of the order \u003cem\u003eClostridia\u003c/em\u003e, including \u003cem\u003eRoseburia intestinalis\u003c/em\u003e, \u003cem\u003eRoseburia faecis\u003c/em\u003e, \u003cem\u003eRoseburia hominis,\u003c/em\u003e \u003cem\u003eF. prausnitzii\u003c/em\u003e, \u003cem\u003eBlautia wexlerae, Fusicatenibacter saccharivorans, Flavonifractor plautii\u0026nbsp;\u003c/em\u003eand \u003cem\u003eEubacterium rectale\u0026nbsp;\u003c/em\u003e(LinDA, all q \u0026lt; 0.05 at 14 months, Supplementary Table 17). Additionally, many previously reported associations (6, 59-61), including a lower abundance of \u003cem\u003eBacteroides\u0026nbsp;\u003c/em\u003eamong infants born by\u003cem\u003e\u0026nbsp;\u003c/em\u003eC-section and a higher abundance of \u003cem\u003eBifidobacterium\u003c/em\u003e species with increasing duration of breastfeeding, were confirmed in our study (Figure 3, Supplementary Table 17). When examining the differential abundance of bacterial species among infants in the different dietary classes, most associations appeared to be age-dependent (i.e., an age*dietary class interaction) in DC3 in comparison to DC1. The exception was the age-independent decreased abundance of \u003cem\u003eB. longum\u003c/em\u003e (LinDA, q = 0.02, Supplementary Table 17) among infants in DC3 as compared to infants in DC2.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eMost age-dependent differences in bacterial species abundance were observed at ages 9 to 14 months and predominantly in DC3 as compared to DC1. Here, the abundance of\u003cem\u003e\u0026nbsp;Bacteroides fragilis,\u0026nbsp;\u003c/em\u003e\u003cem\u003eBifidobacterium adolescentis\u003c/em\u003e, \u003cem\u003eB. bifidum\u003c/em\u003e, \u003cem\u003eB. breve\u003c/em\u003e, \u003cem\u003eBifidobacterium kashiwanohense\u003c/em\u003e, \u003cem\u003eB. longum\u003c/em\u003e, \u003cem\u003eCollinsella aerofaciens\u003c/em\u003e, \u003cem\u003eE. coli\u003c/em\u003e , \u003cem\u003eF. plautii and V. parvula\u0026nbsp;\u003c/em\u003eat 11 months and \u003cem\u003eB. fragilis\u003c/em\u003e and \u003cem\u003eC. aerofaciens\u003c/em\u003e at 9 months was significantly enriched (LinDA, all q \u0026lt; 0.2) in DC3 in comparison to DC1, whereas \u003cem\u003eBacteroides ovatus\u003c/em\u003e was significantly enriched (q = 0.126) in DC2 in comparison to DC1 at 14 months (Figure 3, Supplementary Table 17).\u003c/p\u003e\n\u003cp\u003eAltogether, our results demonstrate that the impact of feeding patterns on bacterial species abundance is most pronounced from the age of 9 months onwards and, in line with the enterotype transition trajectories, reveal a stunted shift towards keystone members of the adult gut microbiome among infants in DC3.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eDietary classes are linked to the microbiome\u0026rsquo;s fiber-degrading capacity\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNext, we inspected the capability of the infants\u0026rsquo; microbiomes to degrade dietary fibers (DF) within and between the dietary classes. After inferring fiber degradation profiles (IFDPs), t-SNE dimensionality reduction was performed. This showed that the variation in IFDP\u0026rsquo;s was mainly driven by the infants\u0026rsquo; age (Supplementary Figure 12). Overall, the infant\u0026rsquo;s GM fiber-degrading capacity was highest for fibers such as xanthan, xyloglucan, chitin, beta-glucan, cellulose and xylan across all time points (Supplementary Figure 13, Supplementary Table 18). Notably, the degradation capacity increased with age for gellan (Dunn\u0026rsquo;s test, 6-9 mon, q = 0.011; 9-11 mon, q = 0.001) and laminaran (9-11 mon, q = 0.001) (Supplementary Tables 19A-B). Galactoglucomannan (q = 0.040), galactomannan (q = 0.040) and dextran (q = 0.024) had a peak in degradation capacity between 6 and 9 months (Supplementary Tables 19B).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eSubsequently, we explored the connection between dietary classes and microbial fiber degradation potential. The T-statistic heatmap (Figure 4A, Supplementary Table 20) highlights that the microbiome of infants in DC1 had amongst others statistically significantly higher degradation capabilities of arabinoxylan, rhamnogalacturonan and arabinogalactan at ages 4, 5 and 6 months and also higher degradation capacity of xylan, pectin and arabinan at 5 and 6 months of age. The microbiome of infants in DC3 was characterized by a higher capacity to degrade chitin, laminaran, xanthan and gellan at various ages (Figure 4A, Supplementary Table 20).\u003c/p\u003e\n\u003cp\u003eTo identify the bacterial species driving these differences in fiber-degrading capacity of the infant microbiome in across dietary classes, we first analyzed longitudinal correlations between IFDPs and bacterial species abundances. The analysis revealed two significant positive correlation patterns between fibers rhamnogalacturonan, arabinogalactan, pectin, arabinan, arabinoxylan, xylan, xyloglucan, mannan, carrageenan, dextran, and galactomannan and bacteria \u003cem\u003eB. uniformis\u003c/em\u003e, \u003cem\u003eB. ovatus\u003c/em\u003e, \u003cem\u003eParabacteroides\u003c/em\u003e \u003cem\u003edistasonis\u003c/em\u003e, \u003cem\u003eBacteroides\u003c/em\u003e\u003cem\u003e\u0026nbsp;vulgatus\u003c/em\u003e, \u003cem\u003eC. aerofaciens\u003c/em\u003e, \u003cem\u003eBacteroides dorei\u003c/em\u003e, \u003cem\u003eB. fragilis\u003c/em\u003e, \u003cem\u003eP. copri\u003c/em\u003e, \u003cem\u003eAlistipes\u003c/em\u003e\u003cem\u003e\u0026nbsp;finegoldii\u003c/em\u003e, \u003cem\u003eRuminococcus gnavus,\u003c/em\u003e and. \u003cem\u003eF. plautii\u003c/em\u003e at 4-6 months. And between fibers laminaran, xanthan, gellan, resistant starch, inulin, and chitin and bacteria \u003cem\u003eR. gnavus\u003c/em\u003e, \u003cem\u003eV. parvula,\u003c/em\u003e and \u003cem\u003eE.coli\u003c/em\u003e at 9-14 months (Supplementary Figure 14, Supplementary Table 21).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eNext, we investigated degradation capacity of the bacterial communities. For that we selected species with a statistically significant (q \u0026lt; 0.05) positive correlations (Spearman rho \u0026gt; 0.3) with a given fiber at least at two time points. Then, we summed the compositionally transformed reads for all selected bacteria capable of potentially degrading that specific fiber (Supplementary Table 22). Out of 15 fiber degradation profiles, 10 profiles correlated with more than one bacterial species based upon our selection criteria. Afterwards, we examined potential differences in the abundance of inferred fiber degrading bacteria between dietary classes at each time point (Supplementary Table 23A). At the age of 4 months, the total abundance of bacteria associated with arabinoxylan degradation differed between DC1-DC2 (Dunn\u0026rsquo;s test, q = 0.024), with the degradation of rhamnogalacturonan DC1-DC3 (Dunn\u0026rsquo;s test, q = 0.044) and xylan DC1-DC2 (Dunn\u0026rsquo;s test, q = 0.046) and DC1-DC3 (Dunn\u0026rsquo;s test, q = 0.046) (Figure 4B, Supplementary Table 23B). At 5 months the abundance of bacteria associated with rhamnogalacturonan degradation differed between DC2-DC3 (Dunn\u0026rsquo;s test, q = 0.049), with dextran, galactoglucomannan, galactomannan between DC1-DC3 (Dunn\u0026rsquo;s test, q = 0.027 for all). Finally, at 6 months with xylan degradation between DC1-DC3 (Dunn\u0026rsquo;s test, q = 0.036).\u003c/p\u003e\n\u003cp\u003eGiven the observed differences in fiber degradation capacity among infants in the three dietary classes, we next explored whether these shifts might translate into altered metabolic potential. We focused on xylan-degrading microbes at 4 months of age because xylan, although more abundant in grains, broccoli, carrots, and cauliflower, is also present in apples (62), pears (63) and strawberries (64) \u0026mdash; foods introduced significantly earlier and more frequently in DC1 than in DC2 and DC3. Indeed, xylan-degrading capacity at 4 months was notably enriched in DC1 (Figure 4B), and correlated strongly (q \u0026lt; 0.001) with \u003cem\u003eF. plautii\u003c/em\u003e (Supplementary Table 21), reflecting this species\u0026rsquo; metabolic potential to utilize xylan at this early age.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTiming and diversity of complementary feeding can modulate butyrate production capacity\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo further assess whether these microbiome differences might influence SCFA production, we compared the gene abundances for key enzymes involved in butyrate, acetate, and propionate biosynthesis at 4 months (Supplementary Tables 24A-B). HUMAnN detected genes belonging to 49 KO identifiers related to the metabolism of propionate (n = 25), acetate (n = 19), and butyrate (n = 10). Among these, five identifiers were associated with two SCFAs. Strikingly, the \u003cem\u003ebuk\u003c/em\u003e gene (encoding butyrate kinase; K00929) was significantly enriched in DC1 relative to both DC2 and DC3 (Figure 4C), indicating an enhanced potential for butyrate formation in infants exposed to earlier and more diverse complementary foods. Several additional butyrate-related genes, including \u003cem\u003ecroR\u003c/em\u003e (3-hydroxybutyryl-CoA dehydratase; K17865), \u003cem\u003eabfD\u003c/em\u003e (4-hydroxybutyryl-CoA dehydratase, K14534), \u003cem\u003ehbd\u003c/em\u003e (3-hydroxybutyryl-CoA dehydrogenase; K00074), mcmA1 (methylmalonyl-CoA mutase, N-terminal domain; K01848), were similarly elevated in DC1 (Figure 4C).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTaxon distribution analyses pointed to \u003cem\u003eF. plautii\u003c/em\u003e as the principal carrier of these genes at 4 months (Figure 4D), followed by \u003cem\u003eAcidaminococcus intestini\u003c/em\u003e, consistent with their documented role in butyrate fermentation (65-67). Intriguingly, however, \u003cem\u003eF. plautii\u003c/em\u003e peaked later in DC3 at 11 months (Figure 3, Supplementary Table 17), while in DC1, \u003cem\u003eF. prausnitzii\u003c/em\u003e rapidly expanded from 9 to 14 months and dominated butyrate production pathways at older ages. This dynamic suggests that, although \u003cem\u003eF. plautii\u003c/em\u003e initially contributes substantially to butyrate-forming potentials in infants who consume xylan-rich vegetables early, it is eventually supplanted by \u003cem\u003eF. prausnitzii\u003c/em\u003e as the diet diversifies further, particularly in DC1. Even though both acetate and propionate are commonly converted upon fiber degradation in the intestine, no substantial differences were observed between the three dietary classes (Supplementary Table 24A-B).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eOverall, our findings highlight how the timing and diversity of complementary food introduction can modulate infant gut fiber metabolism and, by extension, the capacity for butyrate production, potentially influencing immune and metabolic maturation in early life.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eThis study aimed to explore the influence of feeding patterns on the development of the infant\u0026rsquo;s GM community, taking into account confounding factors, and to examine its association with the degradation capacity of various dietary fibers into e.g. SCFA\u0026rsquo;s.\u003c/p\u003e \u003cp\u003eWe identified three DCs based on the timing and types of solid foods introduced. Infants in DC1 received solid foods earlier and had a more diverse intake of fruits; therefore, this group was labeled \u003cb\u003e'early and diverse fruit introducers\u003c/b\u003e\u0026rsquo;, subsequently DC2 was labeled \u0026ldquo;\u003cb\u003epartial introducers with reduced fruit diversity\u003c/b\u003e\u0026rdquo; and DC3 \u0026ldquo;\u003cb\u003edelayed introducers with frequent butter consumption\u003c/b\u003e\u0026rdquo;, due to higher consumption of butter, cheese, and milk that do not contain dietary fibers.\u003c/p\u003e \u003cp\u003eEarly introducers had higher early-life microbial diversity and richness, and their microbiome structure between 4 and 6 months of age differed significantly from the delayed introducers. The same trend was reported in a Danish cohort (\u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e68\u003c/span\u003e), that observed lower alpha and beta diversity in infants receiving solids later on in life. However, in the present study the effect of the early introduction of the solids diminished over time, and by 14 months, it was no longer distinguishable in the GM alpha or beta diversity.\u003c/p\u003e \u003cp\u003eEarly introducers also had more mature microbiome than delayed introducers. Although the exact timing of how and when this is essential for the healthy development of infants is still unclear, previous studies have linked immature GM to negative health outcomes, e.g. asthma development later in life (\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e)(\u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e69\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eInterestingly, our findings contrast with findings from other studies (\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e70\u003c/span\u003e, \u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e71\u003c/span\u003e) by indicating that the type and timing of the introduction of solid foods, rather than the cessation of breast milk, primarily drives the development of the infant's GM, which was also confirmed by Laursen et al. (\u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e68\u003c/span\u003e). Unlike those earlier studies\u0026mdash;some of which used more limited dietary questionnaires or aggregated complementary feeding data, and primarily relied on 16S rRNA sequencing\u0026mdash;our approach integrated comprehensive longitudinal metagenomic profiling with detailed dietary diversity information to capture the nuanced impacts of specific food items on gut microbiome maturation. Despite the age being the main dominant driver of microbiome shifts, delayed introducers showed delayed microbial maturation, resulting in higher levels of early-life bacteria at later age compared to early introducers (\u003cspan citationid=\"CR72\" class=\"CitationRef\"\u003e72\u003c/span\u003e). The only non-age-driven difference in species abundance was \u003cem\u003eB. longum\u003c/em\u003e, higher in partial introducers as compared to delayed introducers. Thus, our study contributes valuable evidence for refining current guidelines on the introduction of solid food.\u003c/p\u003e \u003cp\u003eDue to earlier introduction of the solids, earlier introducers had a smoother transition between the enterotypes. This finding supports previous research indicating that the development of infants' GM is influenced by the introduction of solid foods (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e, \u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e68\u003c/span\u003e, \u003cspan citationid=\"CR73\" class=\"CitationRef\"\u003e73\u003c/span\u003e). Nevertheless, by age 14 months, only partial introducers were fully transitioned to enterotype 6, characterized by the highest abundance of bacteria \u003cem\u003eF. prausnitzii\u003c/em\u003e, a bacterium which typically increases with age, notably after 6 months (\u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e74\u003c/span\u003e, \u003cspan citationid=\"CR75\" class=\"CitationRef\"\u003e75\u003c/span\u003e). In contrast, delayed introducers exhibited lower abundance of this bacterium compared to other classes. It\u0026rsquo;s worth noting that \u003cem\u003eF. prausnitzii\u003c/em\u003e depletion is associated with increase in inflammatory responses, and related with atopic dermatitis in infants (\u003cspan citationid=\"CR76\" class=\"CitationRef\"\u003e76\u003c/span\u003e) and with inflammatory bowel disease in adults (\u003cspan citationid=\"CR77\" class=\"CitationRef\"\u003e77\u003c/span\u003e, \u003cspan citationid=\"CR78\" class=\"CitationRef\"\u003e78\u003c/span\u003e). The feeding patterns observed in delayed introducers resulted in decreased abundance of this bacterium, suggesting a potential risk for AD development in the future (\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e). However, reverse causation cannot be excluded during this time window- given that atopic dermatitis commonly manifests within the first 1\u0026ndash;2 years of life (\u003cspan citationid=\"CR79\" class=\"CitationRef\"\u003e79\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eChanges in gut community structure triggered by different types of diet, will lead towards different community functionality, in our study manifested in DF degradation capacity. Earlier introducers had better capacity in degrading several DF (arabinan containing fibers, pectin, and xylan) at earlier ages, while delayed introducer had higher capacity at various ages for xanthan, gellan, laminaran (\u003cspan citationid=\"CR80\" class=\"CitationRef\"\u003e80\u003c/span\u003e) and chitin despite the main source of these fibers like seaweed, kombu, and gluten-free or vegan alternatives for gelatine are unlikely to be given to the infants. The latter could possibly be explained by the fact that structurally similar polysaccharides often require the same enzymatic machinery for degradation, allowing these microbes to respond to familiar substrates, e.g. chitin and cellulose (\u003cspan citationid=\"CR81\" class=\"CitationRef\"\u003e81\u003c/span\u003e). Moreover, some of the more exotic fibers \u0026mdash; such as gellan and xanthan are widely used as food additive (\u003cspan citationid=\"CR82\" class=\"CitationRef\"\u003e82\u003c/span\u003e, \u003cspan citationid=\"CR83\" class=\"CitationRef\"\u003e83\u003c/span\u003e). Chitin, in comparison, is a component of fungal cell walls (\u003cspan citationid=\"CR84\" class=\"CitationRef\"\u003e84\u003c/span\u003e), and infant exposure to airborne fungal spores or soil microbes may also shape the GM\u0026rsquo;s fiber-degrading capabilities. In our study, chitin degradation capability was positive correlated with \u003cem\u003eE. coli\u003c/em\u003e and \u003cem\u003eR. gnavus\u003c/em\u003e in contrast to a previous finding (\u003cspan citationid=\"CR85\" class=\"CitationRef\"\u003e85\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eAdditionally, early introducers had greater xylan-degrading capacity at 4 months compared to the other DCs, which was strongly linked to \u003cem\u003eF. plautii\u003c/em\u003e. \u003cem\u003eF. plautii\u003c/em\u003e increased with age, and was the highest in delayed introducers at the later ages. This bacterium is known to suppress Th2 immune responses in mice (\u003cspan citationid=\"CR86\" class=\"CitationRef\"\u003e86\u003c/span\u003e), indicating possible anti-inflammatory properties. Moreover, early introducers had significant enrichment of the \u003cem\u003ebuk\u003c/em\u003e gene, indicating higher butyrate production potential as early as at 4 months. Our study demonstrates that introducing diverse fiber-containing foods as early as 4\u0026ndash;5 months leads to the development of a versatile GM community capable of producing butyrate, which supports the findings of Differding et al. (\u003cspan citationid=\"CR87\" class=\"CitationRef\"\u003e87\u003c/span\u003e). Interestingly, \u003cem\u003eF. plautii\u003c/em\u003e is capable of butyrogenic lysine degradation (\u003cspan citationid=\"CR88\" class=\"CitationRef\"\u003e88\u003c/span\u003e), and this pathway is more prominent at 4 months in early and diverse introducers. Thus, our study revealed that a diet rich in fruits and vegetables promotes \u003cem\u003eF. plautii\u003c/em\u003e, which, in turn, potentially enhances butyrate production. This is in alignment with findings in a synthetic gut bacterial community (\u003cspan citationid=\"CR89\" class=\"CitationRef\"\u003e89\u003c/span\u003e). Overall, this indicates that early and diverse complementary feeding enhances SCFA-producing potential, particularly for butyrate, potentially supporting immune and metabolic development in infancy.\u003c/p\u003e \u003cp\u003eMoving forward, it is essential to investigate infant gut metabolites to understand how theoretical dietary fiber degradation capacity profiles correspond to gut metabolome profiles. Furthermore, it is important to consider its immunological response in the gut, e.g. Nwaru et al. (\u003cspan citationid=\"CR90\" class=\"CitationRef\"\u003e90\u003c/span\u003e) demonstrated that introduction of solid foods already at 3 months may protect high-risk infants against atopic sensitization.\u003c/p\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003eStrengths and limitations\u003c/h2\u003e \u003cp\u003eA major strength of this study lies in its integration of dense, longitudinal dietary data with high-resolution metagenomic sequencing across a critical developmental window (4\u0026ndash;14 months). This allowed us to go beyond taxonomic composition and directly assess functional potential, such as fiber-degrading and butyrate-producing capacity, which are rarely examined in such detail during early infancy.\u003c/p\u003e \u003cp\u003eMoreover, the richness of the dietary data\u0026mdash;capturing both timing and diversity of specific food introductions\u0026mdash;enabled us to define biologically meaningful dietary patterns with strong associations to microbial maturation. Our methodological approach, combining logistic PCA with dynamic time warping and clustering, provided nuanced insight into real-life dietary trajectories and their microbial consequences. While our approach offers novel insights, some limitations warrant consideration. Quantification of food intake was not feasible as validated instruments are lacking for this age group as it is difficult to estimate due to frequent spitting up and regurgitation, which usually resolve around 12 months of age (\u003cspan citationid=\"CR91\" class=\"CitationRef\"\u003e91\u003c/span\u003e). As such, our dietary analysis focused on the introduction (yes or no) of individual food items.\u003c/p\u003e \u003cp\u003eIn our clustering approach, we performed separate binary principal component analyses (PCA) at each time-point, and then used the resulting PC scores to compare dietary trajectories using dynamic time warping. A limitation of this method is that the principal component axes inferred from each time-point may represent somewhat different aspects of the underlying dietary information. Consequently, the PC scores are not guaranteed to be on a common scale or orientation, and their use as a longitudinal time series may introduce minor distortions in trajectory comparison that could affect downstream clustering.\u003c/p\u003e \u003cp\u003eDespite this potential limitation, we observed biologically plausible and interpretable groupings, suggesting that our approach captured meaningful variation in complementary feeding patterns. Future studies may consider comparing alternative approaches, such as performing PCA on pooled data across time-points to obtain time-invariant component structures. Nonetheless, we believe our analysis offers valuable insights into early dietary development and provides a basis for further exploration.\u003c/p\u003e \u003c/div\u003e"},{"header":"Conclusion","content":"\u003cp\u003eTaken together, these results underscore how early complementary feeding decisions, even as early as 4 months, can shape the infant gut\u0026rsquo;s metabolic capacity. Introducing xylan-containing grains, vegetables and fruits at this age appears to foster butyrate-producing taxa, including \u003cem\u003eF. plautii\u003c/em\u003e. Over time, \u003cem\u003eF. prausnitzii\u003c/em\u003e eventually predominates in the microbiome of children who were fed rapidly diversifying diets. These shifts in fiber-degrading and butyrate-producing capabilities may significantly impact immune development and long-term metabolic health, highlighting the need for further research into optimal early-life dietary strategies to support microbiome maturation.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cdiv class=\"DefinitionList\"\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eBH\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eBenjamini-Hochberg\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eDC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eDietary class\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eDDS\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eDietary Diversity Score\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eDF\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eDietary Fibers\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eDMM\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eDirichlet Multinomial Mixture\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eDTW\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eDynamic Time Warping\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eFAD\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eFood Allergen Diversity\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eFVS\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eFood Variety Score\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eGM\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eGut microbiota\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eIFDP\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eInferred Fiber Degradation Profiles\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eIQR\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eInterquartile Range\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eLinDA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eLINear models for Differential Abundance analysis\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eMAZ\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eMicrobiota-for-Age Z-score\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eMDS\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eMultidimensional Scaling\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ePC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ePrincipal Components\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ePCA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ePrincipal Component Analysis\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ePCoA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003ePrincipal Coordinates Analysis\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eRMM\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eRelative microbiota maturity\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eSCFA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eShort Chain Fatty Acids\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003et-SNE\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003et-Distributed Stochastic Neighbour Embedding\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWritten informed consent was obtained from both parents/legal caregivers prior to enrolment in the study. This research confirmed to the principles of the Helsinki Declaration. Ethical approval was obtained by the Medical Ethical Committee of Maastricht University Medical Center (study number: METC-15-4-237 ).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTrimmed and quality filtered reads, with removed human reads can be found at European Nucleotide Archive under project number PRJEB89491.\u0026nbsp;The metadata are not publicly available due to the potentially identifiable nature of the data and privacy concerns by study participants, but are available upon reasonable request from the corresponding authors (NvB and MM).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNo potential conflict of interest was reported by the author(s).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe LucKi Gut study was funded by a grant from The Netherlands Organization for Health Research and Development (ZonMw) through the European Union Joint Programming Initiative—A Healthy Diet for a Healthy Life (received by J.P. and M.M.; project #: 529051010).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors' contributions\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eConceptualization and design: L.B., J.P., N.v.B. and M.M.; Study coordination: L.B.; Data collection: L.B., M.M., and E.D.; Methodology: E.D., M.S., L.B., J.P., N.v.B. and M.M.; Formal analysis: E.D., M.S., G.L., D.B., G.G., C.D.; Data curation: M.M., and E.D.; writing—original draft preparation: E.D., M.S., N.v.B. and J.P.; Writing—review and editing, all; supervision: N.v.B. and J.P.; project administration: C.D., M.M., and E.D.; funding acquisition: M.M. and J.P. All authors have read and agreed to the published version of the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors would like to sincerely thank all families for their participation during this study, as well as the participating midwifery practices in the Limburg area of the Netherlands. We also would like to thank members of the John Penders lab\u0026nbsp;for their help with this study. ChatGPT (OpenAI) was used to assist with language editing and improving grammar in this manuscript. The authors take full responsibility for the content and interpretation of the results.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors' information (optional)\u003c/strong\u003e\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003ePenders J, Thijs C, Vink C, Stelma FF, Snijders B, Kummeling I, et al. Factors influencing the composition of the intestinal microbiota in early infancy. Pediatrics. 2006;118(2):511-21.\u003c/li\u003e\n\u003cli\u003ePenders J, Gerhold K, Thijs C, Zimmermann K, Wahn U, Lau S, et al. New insights into the hygiene hypothesis in allergic diseases: mediation of sibling and birth mode effects by the gut microbiota. Gut microbes. 2014;5(2):239-44.\u003c/li\u003e\n\u003cli\u003eJokela R, Korpela K, Jian C, Dikareva E, Nikkonen A, Saisto T, et al. Quantitative insights into effects of intrapartum antibiotics and birth mode on infant gut microbiota in relation to well-being during the first year of life. Gut Microbes. 2022;14(1):2095775.\u003c/li\u003e\n\u003cli\u003eKorpela K. Impact of delivery mode on infant gut microbiota. Annals of Nutrition and Metabolism. 2021;77(Suppl. 3):11-9.\u003c/li\u003e\n\u003cli\u003eKorpela K, Helve O, Kolho K-L, Saisto T, Skogberg K, Dikareva E, et al. Maternal fecal microbiota transplantation in cesarean-born infants rapidly restores normal gut microbial development: a proof-of-concept study. Cell. 2020;183(2):324-34. e5.\u003c/li\u003e\n\u003cli\u003eStewart CJ, Ajami NJ, O\u0026rsquo;Brien JL, Hutchinson DS, Smith DP, Wong MC, et al. Temporal development of the gut microbiome in early childhood from the TEDDY study. Nature. 2018;562(7728):583-8.\u003c/li\u003e\n\u003cli\u003eHomann C-M, Rossel CA, Dizzell S, Bervoets L, Simioni J, Li J, et al. Infants\u0026rsquo; first solid foods: impact on gut microbiota development in two intercontinental cohorts. Nutrients. 2021;13(8):2639.\u003c/li\u003e\n\u003cli\u003eParajuli A, Gr\u0026ouml;nroos M, Siter N, Puhakka R, Vari HK, Roslund MI, et al. Urbanization reduces transfer of diverse environmental microbiota indoors. Frontiers in microbiology. 2018;9:84.\u003c/li\u003e\n\u003cli\u003ePanzer AR, Sitarik AR, Fadrosh D, Havstad SL, Jones K, Davidson B, et al. The impact of prenatal dog keeping on infant gut microbiota development. Clinical \u0026amp; Experimental Allergy. 2023.\u003c/li\u003e\n\u003cli\u003eLiu W, Sun Z, Ma C, Zhang J, Ma C, Zhao Y, et al. Exposure to soil environments during earlier life stages is distinguishable in the gut microbiome of adult mice. Gut Microbes. 2021;13(1):1830699.\u003c/li\u003e\n\u003cli\u003eOlivares M, Neef A, Castillejo G, De Palma G, Varea V, Capilla A, et al. The HLA-DQ2 genotype selects for early intestinal microbiota composition in infants at high risk of developing coeliac disease. Gut. 2015;64(3):406-17.\u003c/li\u003e\n\u003cli\u003eCatassi G, Aloi M, Giorgio V, Gasbarrini A, Cammarota G, Ianiro G. The Role of Diet and Nutritional Interventions for the Infant Gut Microbiome. Nutrients. 2024;16(3):400.\u003c/li\u003e\n\u003cli\u003eWilson AS, Koller KR, Ramaboli MC, Nesengani LT, Ocvirk S, Chen C, et al. Diet and the human gut microbiome: an international review. Digestive diseases and sciences. 2020;65:723-40.\u003c/li\u003e\n\u003cli\u003eVoreades N, Kozil A, Weir TL. Diet and the development of the human intestinal microbiome. Frontiers in microbiology. 2014;5:494.\u003c/li\u003e\n\u003cli\u003eAvershina E, Lundg\u0026aring;rd K, Sekelja M, Dotterud C, Storr\u0026oslash; O, \u0026Oslash;ien T, et al. Transition from infant‐to adult‐like gut microbiota. Environmental microbiology. 2016;18(7):2226-36.\u003c/li\u003e\n\u003cli\u003eJian C, Carp\u0026eacute;n N, Helve O, de Vos WM, Korpela K, Salonen A. Early-life gut microbiota and its connection to metabolic health in children: Perspective on ecological drivers and need for quantitative approach. EBioMedicine. 2021;69.\u003c/li\u003e\n\u003cli\u003eYang I, Corwin EJ, Brennan PA, Jordan S, Murphy JR, Dunlop A. The infant microbiome: implications for infant health and neurocognitive development. Nursing research. 2016;65(1):76-88.\u003c/li\u003e\n\u003cli\u003eMeng Q, Luo Z, Cao C, Sun S, Ma Q, Li Z, et al. Weaning alters intestinal gene expression involved in nutrient metabolism by shaping gut microbiota in pigs. Frontiers in microbiology. 2020;11:694.\u003c/li\u003e\n\u003cli\u003eAl Nabhani Z, Dulauroy S, Marques R, Cousu C, Al Bounny S, D\u0026eacute;jardin F, et al. A weaning reaction to microbiota is required for resistance to immunopathologies in the adult. Immunity. 2019;50(5):1276-88. e5.\u003c/li\u003e\n\u003cli\u003eBail\u0026eacute;n M, Bressa C, Mart\u0026iacute;nez-L\u0026oacute;pez S, Gonz\u0026aacute;lez-Soltero R, Montalvo Lominchar MG, San Juan C, et al. Microbiota features associated with a high-fat/low-fiber diet in healthy adults. Frontiers in nutrition. 2020;7:583608.\u003c/li\u003e\n\u003cli\u003eFritsch J, Garces L, Quintero MA, Pignac-Kobinger J, Santander AM, Fern\u0026aacute;ndez I, et al. Low-fat, high-fiber diet reduces markers of inflammation and dysbiosis and improves quality of life in patients with ulcerative colitis. Clinical Gastroenterology and Hepatology. 2021;19(6):1189-99. e30.\u003c/li\u003e\n\u003cli\u003eBarber C, Mego M, Sabater C, Vallejo F, Bendezu RA, Masihy M, et al. Differential effects of western and mediterranean-type diets on gut microbiota: A metagenomics and metabolomics approach. Nutrients. 2021;13(8):2638.\u003c/li\u003e\n\u003cli\u003eBowyer RC, Jackson MA, Pallister T, Skinner J, Spector TD, Welch AA, et al. Use of dietary indices to control for diet in human gut microbiota studies. Microbiome. 2018;6:1-11.\u003c/li\u003e\n\u003cli\u003eDel Chierico F, Vernocchi P, Dallapiccola B, Putignani L. Mediterranean diet and health: food effects on gut microbiota and disease control. International journal of molecular sciences. 2014;15(7):11678-99.\u003c/li\u003e\n\u003cli\u003eGarc\u0026iacute;a-Vega \u0026Aacute;S, Corrales-Agudelo V, Reyes A, Escobar JS. Diet quality, food groups and nutrients associated with the gut microbiota in a nonwestern population. Nutrients. 2020;12(10):2938.\u003c/li\u003e\n\u003cli\u003eLalli MK, Salo TE, Hakola L, Knip M, Virtanen SM, Vatanen T. Associations between dietary fibers and gut microbiome composition in the EDIA longitudinal infant cohort. The American Journal of Clinical Nutrition. 2025;121(1):83-99.\u003c/li\u003e\n\u003cli\u003eGalazzo G, van Best N, Bervoets L, Dapaah IO, Savelkoul PH, Hornef MW, et al. Development of the microbiota and associations with birth mode, diet, and atopic disorders in a longitudinal analysis of stool samples, collected from infancy through early childhood. Gastroenterology. 2020;158(6):1584-96.\u003c/li\u003e\n\u003cli\u003eOkada H, Kuhn C, Feillet H, Bach J-F. The \u0026lsquo;hygiene hypothesis\u0026rsquo; for autoimmune and allergic diseases: an update. Clinical \u0026amp; Experimental Immunology. 2010;160(1):1-9.\u003c/li\u003e\n\u003cli\u003eAhrens AP, Hy\u0026ouml;tyl\u0026auml;inen T, Petrone JR, Igelstr\u0026ouml;m K, George CD, Garrett TJ, et al. Infant microbes and metabolites point to childhood neurodevelopmental disorders. Cell. 2024;187(8):1853-73. e15.\u003c/li\u003e\n\u003cli\u003eFang C, Zhong H, Lin Y, Chen B, Han M, Ren H, et al. Assessment of the cPAS-based BGISEQ-500 platform for metagenomic sequencing. Gigascience. 2018;7(3):gix133.\u003c/li\u003e\n\u003cli\u003eM\u0026ouml;lder F, Jablonski KP, Letcher B, Hall MB, Tomkins-Tinch CH, Sochat V, et al. Sustainable data analysis with Snakemake. F1000Research. 2021;10.\u003c/li\u003e\n\u003cli\u003eChen S, Zhou Y, Chen Y, Gu J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics. 2018;34(17):i884-i90.\u003c/li\u003e\n\u003cli\u003eLangmead B, Salzberg SL. Fast gapped-read alignment with Bowtie 2. Nature methods. 2012;9(4):357-9.\u003c/li\u003e\n\u003cli\u003eDanecek P, Bonfield JK, Liddle J, Marshall J, Ohan V, Pollard MO, et al. Twelve years of SAMtools and BCFtools. Gigascience. 2021;10(2):giab008.\u003c/li\u003e\n\u003cli\u003eBeghini F, McIver LJ, Blanco-M\u0026iacute;guez A, Dubois L, Asnicar F, Maharjan S, et al. Integrating taxonomic, functional, and strain-level profiling of diverse microbial communities with bioBakery 3. elife. 2021;10:e65088.\u003c/li\u003e\n\u003cli\u003eLouis P, Flint HJ. Formation of propionate and butyrate by the human colonic microbiota. Environmental microbiology. 2017;19(1):29-41.\u003c/li\u003e\n\u003cli\u003eReichardt N, Duncan SH, Young P, Belenguer A, McWilliam Leitch C, Scott KP, et al. Phylogenetic distribution of three pathways for propionate production within the human gut microbiota. The ISME journal. 2014;8(6):1323-35.\u003c/li\u003e\n\u003cli\u003eSingh V, Lee G, Son H, Koh H, Kim ES, Unno T, et al. Butyrate producers,\u0026ldquo;The Sentinel of Gut\u0026rdquo;: Their intestinal significance with and beyond butyrate, and prospective use as microbial therapeutics. Frontiers in microbiology. 2023;13:1103836.\u003c/li\u003e\n\u003cli\u003eZhang B, Lingga C, Bowman C, Hackmann TJ. A new pathway for forming acetate and synthesizing ATP during fermentation in bacteria. Applied and environmental microbiology. 2021;87(14):e02959-20.\u003c/li\u003e\n\u003cli\u003eTeam RDC. R: A language and environment for statistical computing. (No Title). 2010.\u003c/li\u003e\n\u003cli\u003eLandgraf AJ, Lee Y. Dimensionality reduction for binary data through the projection of natural parameters. Journal of Multivariate Analysis. 2020;180:104668.\u003c/li\u003e\n\u003cli\u003eGiorgino T. Computing and visualizing dynamic time warping alignments in R: the dtw package. Journal of statistical Software. 2009;31:1-24.\u003c/li\u003e\n\u003cli\u003eGalili T. dendextend: an R package for visualizing, adjusting and comparing trees of hierarchical clustering. Bioinformatics. 2015;31(22):3718-20.\u003c/li\u003e\n\u003cli\u003eWickham H, Wickham H. Data analysis: Springer; 2016.\u003c/li\u003e\n\u003cli\u003ePowell C. CGPfunctions: powell miscellaneous functions for teaching and learning statistics. R package version 06. 2020;3.\u003c/li\u003e\n\u003cli\u003eRathnayake KM, Madushani P, Silva K. Use of dietary diversity score as a proxy indicator of nutrient adequacy of rural elderly people in Sri Lanka. BMC research notes. 2012;5:1-6.\u003c/li\u003e\n\u003cli\u003eLee BR, Jung H-I, Kim SK, Kwon M, Kim H, Jung M, et al. Dietary Diversity during early infancy increases microbial diversity and prevents egg allergy in high-risk infants. Immune Network. 2022;22(2).\u003c/li\u003e\n\u003cli\u003eJari Oksanen F, Friendly M, Kindt R, Legendre P, McGlinn D, Minchin PR, et al. Vegan: community ecology package. R package version. 2018;2(6).\u003c/li\u003e\n\u003cli\u003eBarnett DJ, Arts IC, Penders J. microViz: an R package for microbiome data visualization and statistics. Journal of Open Source Software. 2021;6(63):3201.\u003c/li\u003e\n\u003cli\u003eGu Z, Eils R, Schlesner M. Complex heatmaps reveal patterns and correlations in multidimensional genomic data. Bioinformatics. 2016;32(18):2847-9.\u003c/li\u003e\n\u003cli\u003eHolmes I, Harris K, Quince C. Dirichlet multinomial mixtures: generative models for microbial metagenomics. PloS one. 2012;7(2):e30126.\u003c/li\u003e\n\u003cli\u003eMorgan M. DirichletMultinomial: Dirichlet-Multinomial Mixture Model Machine Learning for Microbiome Data. 2021. p. R package.\u003c/li\u003e\n\u003cli\u003eZhou H, He K, Chen J, Zhang X. LinDA: linear models for differential abundance analysis of microbiome compositional data. Genome biology. 2022;23(1):95.\u003c/li\u003e\n\u003cli\u003eSubramanian S, Huq S, Yatsunenko T, Haque R, Mahfuz M, Alam MA, et al. Persistent gut microbiota immaturity in malnourished Bangladeshi children. Nature. 2014;510(7505):417-21.\u003c/li\u003e\n\u003cli\u003eLiaw A, Wiener M. Classification and regression by randomForest. R news. 2002;2(3):18-22.\u003c/li\u003e\n\u003cli\u003eCohen Y, Borenstein E. The microbiome\u0026rsquo;s fiber degradation profile and its relationship with the host diet. BMC biology. 2022;20(1):266.\u003c/li\u003e\n\u003cli\u003eBuchfink B, Reuter K, Drost H-G. Sensitive protein alignments at tree-of-life scale using DIAMOND. Nature methods. 2021;18(4):366-8.\u003c/li\u003e\n\u003cli\u003eBenjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal statistical society: series B (Methodological). 1995;57(1):289-300.\u003c/li\u003e\n\u003cli\u003eShao Y, Forster SC, Tsaliki E, Vervier K, Strang A, Simpson N, et al. Stunted microbiota and opportunistic pathogen colonization in caesarean-section birth. Nature. 2019;574(7776):117-21.\u003c/li\u003e\n\u003cli\u003eReyman M, van Houten MA, van Baarle D, Bosch AA, Man WH, Chu MLJ, et al. Impact of delivery mode-associated gut microbiota dynamics on health in the first year of life. Nature communications. 2019;10(1):4997.\u003c/li\u003e\n\u003cli\u003eMatharu D, Ponsero AJ, Dikareva E, Korpela K, Kolho K-L, de Vos WM, et al. Bacteroides abundance drives birth mode dependent infant gut microbiota developmental trajectories. Frontiers in Microbiology. 2022;13:953475.\u003c/li\u003e\n\u003cli\u003eBorujeni NE, Alavijeh MK, Denayer JF, Karimi K. A novel integrated biorefinery approach for apple pomace valorization with significant socioeconomic benefits. Renewable Energy. 2023;208:275-86.\u003c/li\u003e\n\u003cli\u003eJermyn M, Isherwood F. Changes in the cell wall of the pear during ripening. Biochemical Journal. 1956;64(1):123.\u003c/li\u003e\n\u003cli\u003eHotchkiss AT, Chau HK, Strahan GD, Nu\u0026ntilde;ez A, Harron A, Simon S, et al. Structural characterization of strawberry pomace. Heliyon. 2024;10(9).\u003c/li\u003e\n\u003cli\u003eJumas-Bilak E, Carlier J-P, Jean-Pierre H, Mory F, Teyssier C, Gay B, et al. Acidaminococcus intestini sp. nov., isolated from human clinical samples. International journal of systematic and evolutionary microbiology. 2007;57(10):2314-9.\u003c/li\u003e\n\u003cli\u003eCardoso BB, Amorim C, Franco-Duarte R, Alves JI, Barbosa SG, Silv\u0026eacute;rio SC, et al. Epilactose as a promising butyrate-promoter prebiotic via microbiota modulation. Life. 2024;14(5):643.\u003c/li\u003e\n\u003cli\u003eCui J, Ramesh G, Wu M, Jensen ET, Crago O, Bertoni AG, et al. Butyrate-producing bacteria and insulin homeostasis: the microbiome and insulin longitudinal evaluation study (MILES). Diabetes. 2022;71(11):2438-46.\u003c/li\u003e\n\u003cli\u003eLaursen MF, Andersen LB, Michaelsen KF, M\u0026oslash;lgaard C, Trolle E, Bahl MI, et al. Infant gut microbiota development is driven by transition to family foods independent of maternal obesity. Msphere. 2016;1(1):10.1128/msphere. 00069-15.\u003c/li\u003e\n\u003cli\u003eStokholm J, Blaser MJ, Thorsen J, Rasmussen MA, Waage J, Vinding RK, et al. Maturation of the gut microbiome and risk of asthma in childhood. Nature communications. 2018;9(1):141.\u003c/li\u003e\n\u003cli\u003eB\u0026auml;ckhed F, Roswall J, Peng Y, Feng Q, Jia H, Kovatcheva-Datchary P, et al. Dynamics and stabilization of the human gut microbiome during the first year of life. Cell host \u0026amp; microbe. 2015;17(5):690-703.\u003c/li\u003e\n\u003cli\u003eBergstr\u0026ouml;m A, Skov TH, Bahl MI, Roager HM, Christensen LB, Ejlerskov KT, et al. Establishment of intestinal microbiota during early life: a longitudinal, explorative study of a large cohort of Danish infants. Applied and environmental microbiology. 2014;80(9):2889-900.\u003c/li\u003e\n\u003cli\u003eLaursen MF, Sakanaka M, von Burg N, M\u0026ouml;rbe U, Andersen D, Moll JM, et al. Bifidobacterium species associated with breastfeeding produce aromatic lactic acids in the infant gut. Nature microbiology. 2021;6(11):1367-82.\u003c/li\u003e\n\u003cli\u003eLaursen MF. Gut microbiota development: influence of diet from infancy to toddlerhood. Annals of Nutrition and Metabolism. 2021;77(Suppl. 3):21-34.\u003c/li\u003e\n\u003cli\u003eRobertson RC, Manges AR, Finlay BB, Prendergast AJ. The human microbiome and child growth\u0026ndash;first 1000 days and beyond. Trends in microbiology. 2019;27(2):131-47.\u003c/li\u003e\n\u003cli\u003eKoga Y, Tokunaga S, Nagano J, Sato F, Konishi K, Tochio T, et al. Age-associated effect of kestose on Faecalibacterium prausnitzii and symptoms in the atopic dermatitis infants. Pediatric Research. 2016;80(6):844-51.\u003c/li\u003e\n\u003cli\u003eCandela M, Rampelli S, Turroni S, Severgnini M, Consolandi C, De Bellis G, et al. Unbalance of intestinal microbiota in atopic children. BMC microbiology. 2012;12:1-9.\u003c/li\u003e\n\u003cli\u003eSokol H, Seksik P, Furet J, Firmesse O, Nion-Larmurier I, Beaugerie L, et al. Low counts of Faecalibacterium prausnitzii in colitis microbiota. Inflammatory bowel diseases. 2009;15(8):1183-9.\u003c/li\u003e\n\u003cli\u003eMachiels K, Joossens M, Sabino J, De Preter V, Arijs I, Eeckhaut V, et al. A decrease of the butyrate-producing species Roseburia hominis and Faecalibacterium prausnitzii defines dysbiosis in patients with ulcerative colitis. Gut. 2014;63(8):1275-83.\u003c/li\u003e\n\u003cli\u003eIlli S, von Mutius E, Lau S, Nickel R, Gr\u0026uuml;ber C, Niggemann B, et al. The natural course of atopic dermatitis from birth to age 7 years and the association with asthma. Journal of Allergy and Clinical Immunology. 2004;113(5):925-31.\u003c/li\u003e\n\u003cli\u003eSharma PP, Baskaran V. Polysaccharide (laminaran and fucoidan), fucoxanthin and lipids as functional components from brown algae (Padina tetrastromatica) modulates adipogenesis and thermogenesis in diet-induced obesity in C57BL6 mice. Algal research. 2021;54:102187.\u003c/li\u003e\n\u003cli\u003eLatańska I, Kolesińska B, Draczyński Z, Sujka W. The use of chitin and chitosan in manufacturing dressing materials. Progress on Chemistry and Application of Chitin and its Derivatives. 2020(25):16-36.\u003c/li\u003e\n\u003cli\u003eDo MH, Lee HHL, Lee J-E, Park M, Oh M-J, Lee H-B, et al. Gellan gum prevents non-alcoholic fatty liver disease by modulating the gut microbiota and metabolites. Food Chemistry. 2023;400:134038.\u003c/li\u003e\n\u003cli\u003eOstrowski MP, La Rosa SL, Kunath BJ, Robertson A, Pereira G, Hagen LH, et al. Mechanistic insights into consumption of the food additive xanthan gum by the human gut microbiota. Nature microbiology. 2022;7(4):556-69.\u003c/li\u003e\n\u003cli\u003eLopez-Santamarina A, Mondragon AdC, Lamas A, Miranda JM, Franco CM, Cepeda A. Animal-origin prebiotics based on chitin: An alternative for the future? a critical review. Foods. 2020;9(6):782.\u003c/li\u003e\n\u003cli\u003eRodriguez J, Neyrinck AM, Zhang Z, Seethaler B, Nazare J-A, Robles S\u0026aacute;nchez C, et al. Metabolite profiling reveals the interaction of chitin-glucan with the gut microbiota. Gut Microbes. 2020;12(1):1810530.\u003c/li\u003e\n\u003cli\u003eOgita T, Yamamoto Y, Mikami A, Shigemori S, Sato T, Shimosato T. Oral administration of Flavonifractor plautii strongly suppresses Th2 immune responses in mice. Frontiers in immunology. 2020;11:379.\u003c/li\u003e\n\u003cli\u003eDifferding MK, Benjamin-Neelon SE, Hoyo C, \u0026Oslash;stbye T, Mueller NT. Timing of complementary feeding is associated with gut microbiota diversity and composition and short chain fatty acid concentrations over the first year of life. BMC microbiology. 2020;20:1-13.\u003c/li\u003e\n\u003cli\u003eShetty SA, Kuipers B, Atashgahi S, Aalvink S, Smidt H, de Vos WM. Inter-species metabolic interactions in an in-vitro minimal human gut microbiome of core bacteria. npj Biofilms and Microbiomes. 2022;8(1):21.\u003c/li\u003e\n\u003cli\u003eWeiss AS, Niedermeier LS, von Strempel A, Burrichter AG, Ring D, Meng C, et al. Nutritional and host environments determine community ecology and keystone species in a synthetic gut bacterial community. Nature Communications. 2023;14(1):4780.\u003c/li\u003e\n\u003cli\u003eNwaru BI, Takkinen HM, Niemel\u0026auml; O, Kaila M, Erkkola M, Ahonen S, et al. Introduction of complementary foods in infancy and atopic sensitization at the age of 5 years: timing and food diversity in a Finnish birth cohort. Allergy. 2013;68(4):507-16.\u003c/li\u003e\n\u003cli\u003eHegar B, Dewanti NR, Kadim M, Alatas S, Firmansyah A, Vandenplas Y. Natural evolution of regurgitation in healthy infants. Acta Paediatrica. 2009;98(7):1189-93.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"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":"Complimentary food, introduction of solids, dietary fiber, xylan, butyrate, Flavonifractor plautii","lastPublishedDoi":"10.21203/rs.3.rs-6673441/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6673441/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground:\u003c/strong\u003e The infant gut microbiota (GM) matures rapidly during early life with complementary feeding marking a pivotal dietary shift that can shape long-term health trajectories. Detailed insight into how timing and composition of solid food introduction influences the composition and metabolic potential of the infant microbiome remains, however, limited. This study aimed to evaluate the impact of complementary feeding dynamics on the infant GM.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods:\u003c/strong\u003e We conducted longitudinal whole metagenomic sequencing of fecal samples collected at \u0026nbsp;4, 5, 6, 9, 11, and 14 months of age from 112 Dutch infants in the LucKi Gut cohort. Based on dietary questionnaires, \u0026nbsp;infants were grouped into three distinct dietary classes of complementary feeding.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults:\u003c/strong\u003e Infants introduced earlier to a wider variety of solid foods exhibited more diverse and mature gut microbiota already at 4 months, with increased abundance of butyrate-producing taxa such as \u003cem\u003eFlavonifractor plautii\u003c/em\u003e. Their microbiomes also showed enhanced capacity to degrade dietary fibers like xylan and rhamnogalacturonan, suggesting accelerated development of metabolic functionality. Functional profiling revealed early enrichment in genes involved in butyrate synthesis, pointing to a link between early feeding diversity and SCFA-producing potential.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions:\u003c/strong\u003e Our findings highlight that early and diverse complementary feeding fosters a functionally mature microbiota with enhanced fiber degradation and butyrate production capacity. These microbial trajectories may influence immune and metabolic development, underscoring the importance of timely dietary diversification in infancy.\u003c/p\u003e","manuscriptTitle":"Dietary impact on infants' gut microbiota and its capacity in SCFA metabolism: a longitudinal Dutch cohort study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-06-09 05:11:00","doi":"10.21203/rs.3.rs-6673441/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":"c86e47c1-817e-4be8-aa86-16e36482f1dd","owner":[],"postedDate":"June 9th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-01-25T12:24:25+00:00","versionOfRecord":[],"versionCreatedAt":"2025-06-09 05:11:00","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6673441","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6673441","identity":"rs-6673441","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}