Expanding Conditions for Newborn Screening—Findings from an Untargeted Metabolomics and Lipidomics Discovery Study

Advances in metabolomics and genomics present new opportunities to expand newborn screening. Traditional screening approaches have relied on condition-specific biomarkers; however, untargeted metabolomic profiling may enable broader detection of genetic conditions. In this study, we developed and evaluated a novel metabolomic screening approach based on defining a reference profile representing a healthy newborn population. Retrospective dried blood spot samples from 2948 newborns were analysed using untargeted metabolomics and lipidomics, generating more than 30,000 features. A discovery cohort of over 2000 healthy newborns and newborns with genetic conditions was used to identify 1115 features that best discriminated between groups. These features formed the basis of a screening algorithm, which was subsequently evaluated in an independent validation cohort containing genetic conditions not represented in the discovery cohort. Across the discovery and validation cohorts, 180 cases representing 113 unique genetic conditions were assessed, including metabolic, haematological, neurological, and syndromic conditions. In the independent validation cohort, the algorithm achieved an area under the receiver operating characteristic curve of 0.799, with a specificity of over 99%. These findings demonstrate the potential of untargeted metabolomic profiling to identify newborns at increased risk of genetic disease and support future exploration in a multi-omic newborn screening design.