Using 35 metabolic markers analyzed by mass spectrometry from newborn heel stick blood samples, this team was able to identify term and preterm newborns with more than 95% accuracy in a cohort of 700,000 California babies. They were able to assign a week of gestation within two weeks in approximately 90% of newborns.

With co-funding from BMGF Grand Challenges and UCSF PTBi-EA, validation work is being conducted in Tororo to test the algorithm on more than 800 newborns, as well as existing banked biospecimens.