Proteome-wide interaction study of fatty acids and mortality in the UK Biobank
British Journal of Nutrition, 2026
Qiao Z., Wang X., Tao C., Liao S., Lu J., Yuan Y., Xu Q., Fan Y., Wang X., Lu C.
| Disease area | Application area | Sample type | Products |
|---|---|---|---|
Metabolic Diseases | Pathophysiology | Plasma | Olink Explore 3072/384 |
Abstract
The heterogeneity in associations between circulating fatty acids (FA) and mortality remained underexplored. Proteomics can profile the human physiological status. This study aimed to estimate interactions between FA and proteins in relation to mortality. We randomly divided 30 190 UK Biobank participants into train and test datasets. Multivariable Cox regression was utilised to assess the associations between FA and all-cause mortality and to explore proteome-wide interactions of FA in relation to mortality. Subgroup analyses were conducted to examine heterogeneity across varied protein levels. We also explored interactions between proteins and FA in relation to cause-specific mortality. We documented 3345 deaths during 13·9 years of follow-up. MUFA to total fatty acids percentage (MUFA-pct), non-LA n -6 pct, n -6/ n -3 ratio and SFA to total fatty acids percentage were positively associated with all-cause mortality, while PUFA to total fatty acids percentage, docosahexaenoic acid to total fatty acids percentage, LA to total fatty acids percentage and n -3-pct were negatively associated. We identified several robust interactions of proteins with MUFA-pct ( n 3), n -3-pct ( n 4) and n -6/ n -3 ratio ( n 2). In subgroup analyses, individuals with high-level TNFRSF1B, MMP10 and CRHBP had higher all-cause mortality risks associated with MUFA-pct, while protective associations between n -3-pct and all-cause mortality were stronger among individuals with high-level TSPAN8, urokinase-type plasminogen activator (PLAU), ITGA5 and CEACAM1. Moreover, participants with high-level TSPAN8 and PLAU had higher risks of all-cause mortality with n -6/ n -3 ratio. For cause-specific mortality, interaction and subgroup results were largely consistent with those of all-cause mortality. Our findings can provide new insights into heterogeneity in FA-mortality associations and highlight potential protein targets for personalised interventions across individuals with different physiological status.