Cord serum metabolic signatures associated with neonatal anthropometry and neurodevelopment through age two in twins

Conventional singleton studies struggle to disentangle environmental from genetic influences on phenotypes. This twin study (n = 129 pairs) leveraged within-pair comparisons to clarify these influences on early development. At the population level, we identified 130 cord serum metabolites associated with neonatal anthropometry and neurodevelopment up to 24 months. Monozygotic and dizygotic comparisons showed that metabolite associations with fetal growth measures (e.g., birth weight) were genetically influenced, whereas links to 6-month fine motor skills reflected non-shared intrauterine factors, especially unequal placental sharing. Twin pairs were categorized into high- and low-discordance groups based on the relative percentage difference of each phenotype. Specific metabolites (e.g., cholecalciferol, cytosine, and LPC 18:2) collectively predicted multiple neurodevelopmental discordance (100 bootstrap-validated mean area under the curve (AUC): 0.758-0.832). These findings demonstrate that metabolite-development associations are co-shaped by genetic predispositions and unique intrauterine factors, with differential exposures contributing to phenotypic variation even within twin pairs.