Hiring in life sciences? Share your open positions with our professional community. Read more Close

Advertisement

Machine learning in human milk research: current applications, challenges, and future perspectives.

Created on 19 Nov 2026

Authors

Yuankui Wang, Yuankai Wang, Lu Wang, Fanlong Wang, Krisztina Takács, Leila Queiroz Zepka, Saied A Aboushanab, Tian Cai, Kewei Chen

Published in

Food research international (Ottawa, Ont.). Volume 246. Pages 121090. Dec 31, 2026. Epub Oct 31, 2026.

Abstract

Machine learning (ML) can help analyze high-dimensional and heterogeneous data and model nonlinear relationships in human milk research, but its value depends on study design, validation, and biological interpretation. This structured narrative review identifies three recurrent challenges involving high dimensionality relative to independent sample size, multilevel biological and sampling variation, and technical heterogeneity across analytical modalities. It relates these challenges to relevant ML and chemometric methods and assesses data resources for benchmarking. Evidence is examined across compositional analysis, milk bank quality control and safety assurance, and infant formula optimization. Applications include compositional profiling, functional peptide screening, spectroscopic quality assessment, and ingredient design. A four-level framework classifies reported validation designs to support the interpretation of model performance, with biological or experimental confirmation assessed separately. The strength and transferability of evidence remain uneven across these domains. Reported performance is difficult to compare because evaluation often relies on resampling or same-study held-out testing, independent units are not consistently separated, and separate-source independent testing remains uncommon. Multimodal integration remains limited. Progress requires data splitting by donor, batch, or other independent unit, transparent reporting, validation matched to intended use, and appropriate biological confirmation. Component-function modeling, portable sensing, federated learning, and digital twins may support translation, but their value depends on data quality and rigorous evaluation rather than technological novelty alone.

PMID:
43051838
Bibliographic data and abstract were imported from PubMed on 19 Nov 2026.

Read full publication at:
Please sign in to see all details.

Advertisement

Stats

  • Community rating n/a 0 votes
  • Reviewers' rating n/a 0 votes
  • Your rating

1-terrible, 9-excellent. How would you rate this publication? Sign in in to submit your rating.

  • Recommendations n/a n/a positive of 0 vote(s)
  • Views 3
  • Comments 0

Recommended by

  • No recommendations yet.

Post a comment

You need to be signed in to post comments. You can sign in here.

Comments

There are no comments yet.

Advertisement