PROJECT SUMMARY
Significance: Human milk (HM) has evolved over millions of years to nourish and protect human infants - yet
we know surprisingly little about its composition, variation, and function. Traditionally, HM research has focused
on individual HM components, yet HM is a complex biological system comprising thousands of components
that interact and function in combination. Moreover, while HM composition is known to be affected by maternal,
infant, and environmental factors, these are poorly understood and rarely examined simultaneously. To address
these gaps, our team is championing a multi-omics systems biology approach to study HM as a “system within
a system”, reflecting that milk itself is a system embedded within the “mother-milk-infant” triad.
Approach: This grant will leverage and unite two established HM research platforms to investigate HM and its
determinants and health impacts among 1600 mother-infant dyads using a novel multi-omic approach. The
International Milk Composition (IMiC) Consortium is a network of HM researchers and data scientists with an
established infrastructure for multi-omic HM research. CHILD is an ongoing national pregnancy cohort of 3600
children born in 2009-12. Our team has already analyzed 1600 CHILD HM samples for 19 oligosaccharides, 28
fatty acids, and hundreds of bacteria. We will now enhance the rich CHILD dataset with new multi-omic HM
analyses (20 nutrients, 15 non-nutritive bioactive proteins and thousands of metabolites) and apply
unsupervised machine learning methods to identify discrete ‘lactotypes’ (Aim 1). Next, we will leverage the rich
CHILD data to identify maternal, infant and environmental factors associated with lactotype membership and/or
individual HM components (Aim 2). Finally, we will use machine learning methods to understand how HM
composition influences microbiome development, growth, wheezing and allergies during infancy and childhood
(Aim 3).
Innovation: Integrating the CHILD and IMiC platforms will facilitate unprecedented research on HM as a
system-within-a-system and generate the world’s largest and most deeply-phenotyped mother-milk-infant
dataset (n=1600 triads with multi-omic milk profiles and rich longitudinal maternal and infant metadata). This
project will unite expert HM scientists, renowned pediatric researchers and data scientists at the forefront of
multi-omic methods development, placing the interdisciplinary MuMi team in an unrivaled position to make
novel discoveries in this space and revolutionize the way HM is studied and understood.
