IMiC intervention effects on human milk: online resource
Overview
This online resource accompanies Dailey-Chwalibóg, Mertens et al. (2026), “Nutritional interventions’ impacts on human milk: three trials in low-resource settings”, from the International Milk Composition (IMiC) Consortium. It holds the full results tables, interactive views of the estimates, and analyses that are not in the article.
| Section | Contents |
|---|---|
| 1 | Baseline characteristics by trial and arm |
| 2 | All milk components analyzed; their distributions by trial, visit, and arm |
| 3 | Effects by maternal BMI |
| 4 | Effects on the microbiome, untargeted proteome, and untargeted metabolome |
| 5 | Component metadata; effects on each targeted milk component, with an interactive explorer |
| 6 | Effects on nutrient deficiency |
| 7 | Effects on change between visits |
| 8 | Unadjusted and native-unit effects (Table S1), treatment-arm prediction, principal-component scores, infant growth |
| 9 | Pathway and enrichment analyses (Tables S2–S7), with an interactive explorer |
| 10 | Maternal and infant blood in MISAME-III (Tables S8–S11) |
Estimates and plots
Effects are average treatment effects (ATEs) estimated by targeted maximum likelihood estimation (TMLE) with adjustment for baseline covariates, in standard deviations of each component’s study- and visit-specific z-score unless a table gives native units. Q is the Benjamini–Hochberg FDR-adjusted P value, computed within outcome group, study, and visit.
In volcano plots, points with Q ≥ 0.05 are black (squares if P < 0.05), and points with Q < 0.05 take the color and symbol of their component category; the gray line marks P = 0.05 and the green line the FDR threshold of each panel. In forest plots of randomized arms, each arm has its own color and the symbol marks significance: open circle P ≥ 0.05, open triangle P < 0.05, filled circle Q < 0.05. Hover over a point for its values. Tables can be searched, sorted, and filtered by column.
Data files
The complete versions of the supplementary tables printed in part, as CSV files:
table_s1_primary_secondary_native_units.csv: Table S1, one row per trial, visit, and outcome: per-arm sample size and raw distribution side by side, with the native-unit ATE, 95% CI, P, and Q (627 rows; Section 8.2.2)targeted_metabolites_native_units.csv: native-unit arm means and effects for the 581 targeted metabolites, one row per trial, visit, and metabolite (3,911 rows; Section 8.2.1)table_s3_tertiary_msea_full.csv: Table S3, targeted-metabolome MSEA, all rows (463 rows; Section 9.2)table_s6_mummichog_full.csv: Table S6, untargeted-metabolome Mummichog enrichment, pathways with P < 0.05 (505 rows; Section 9.4)table_s7_proteomics_go_full.csv: Table S7, proteome GO Biological Process enrichment, terms with P < 0.05 and Storey q < 0.20 (182 rows; Section 9.5)
How to cite
Cite the article, and cite this online resource by the DOI of its archive, 10.5281/zenodo.22104233, which also holds the analysis code.
- Source for this site: https://github.com/amertens/IMiC-intervention-effects-supplement
- Analysis code: https://github.com/amertens/IMiC-intervention-effects-public
- Trial registrations (ClinicalTrials.gov): IMiC NCT05119166; MISAME-III NCT03533712; Mumta-LW NCT03564652; ELICIT NCT03268902