Back

Proteomic signatures of BMI generalize across ancestries and reveal cardiometabolic heterogeneity beyond measured BMI

Muti, M.; Ilboudo, Y.; Aliyu, U.; Zhang, C.; Ware, L.; Chong, M.; Micklesfield, L. K.; Goedecke, J. H.; Ntozini, R.; Pare, G.; Cho, M. H.; Ramsay, M.; Morris, A. P.; Franks, P. W.; Kamiza, A. B.; Norris, S. A.; Albagha, O. M. E.; Moll, M.; Chikowore, T.

2026-09-10 endocrinology
10.64898/2026.09.05.26362276 medRxiv
Show abstract

Plasma proteomic scores for body mass index have been derived primarily in European populations, and it is unclear whether these scores are generalizable and useful in other ancestries. Here, we show that plasma proteomic BMI scores developed in participants of European, East Asian, South Asian and African ancestries from the UK Biobank (UKBB; N=50,621) and the South African Middle-Aged Soweto Cohort (MASC; N=948) are portable across ancestries, explain up to 48.7% of trait variance and significantly predict incident obesity. We identified eight protein biomarkers common across all scores, six of which showed causal associations with BMI in Mendelian randomization analyses and were associated with BMI across the Olink (UKBB, MASC) and SomaScan platforms (Qatar Biobank; N= 2410). Discrepancy analysis between the high predicted proteomic BMI and low measured BMI revealed metabolically unhealthy normal weight (MUNW) individuals, with higher visceral fat, elevated triglycerides, and low insulin sensitivity. Thus, plasma proteomic BMI scores may enhance the precision of cardiometabolic risk stratification across diverse populations

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.