Section 10 Cross-compartment analyses in maternal and infant blood

Intervention effects in the MISAME-III subset with maternal plasma, maternal and infant dried-blood microsample (VAMS) metabolomics, and maternal blood proteomics, and their comparison with milk. MISAME-III randomized prenatal and postnatal BEP in a 2 × 2 design: control (IFA/IFA), prenatal only (BEP/IFA), postnatal only (IFA/BEP), and both (BEP/BEP). Effects are covariate-adjusted TMLE estimates within each dataset (SD units). Because the platforms measure on different scales, compartments are compared only by the direction of their effects. Names of untargeted features are putative (mass-based, MSI level 3).

Visit codes: tri3 third trimester, incl enrollment, acco birth, pn12 1–2 months postpartum, pn34 3–4 months, pn56 5–6 months.

10.1 Effects in each blood compartment

The number of FDR-significant features by compartment, visit, contrast, and direction, and the putative names of those features. Per-feature results for all features are too large to include.

10.2 Concordance between compartments

Agreement in direction of BEP-responsive features between compartment pairs. Features were matched across platforms by mass within ionization mode, at ±25 ppm between the two measured masses; maternal and infant postnatal VAMS share one feature catalogue and are matched by identifier. Mass-only matches can pair different molecules of the same mass. The tables give agreement among significant features, rank-based measures that use all features (rank–rank hypergeometric overlap, weighted correlation, directional GSEA), and the sensitivity of the matching to its tolerance.

10.3 Milk and maternal-blood proteomes

Proteome effects in milk and maternal blood matched by UniProt accession: 302 comparisons of 215 proteins with the immunodepleted or untreated (naive) maternal-blood assay.

10.4 Acylcarnitine feature at m/z 286.202

Blood effects of the features at m/z 286.202 (putatively octenoylcarnitine, an isobaric-ambiguous annotation), then the mass and retention-time evidence for the annotation and the comparison with the acylcarnitines measured on the targeted panel.

10.5 Features increased in two or more compartments

BEP-up features increased in at least two of milk, maternal blood, and infant blood, linked by shared feature identifier (maternal and infant postnatal VAMS), identical putative name, or same-mode mass (±25 ppm), with their effect in each compartment and visit and whether they were detected in the BEP supplement (18 replicates on the same platform, ±25 ppm).

10.6 Pathway direction and chemical classes

Mummichog pathway enrichment of the blood features, which does not use the direction of effects, and binomial sign tests that give each pathway a direction from the signs of its mapped features’ effects. Then chemical-class over-representation among BEP-responsive blood features by direction of effect (Fisher exact test; Table S9 prints the rows with P < 0.05).