Data Sheet 1_Causal inference study of plasma proteins and blood metabolites mediating the effect of obesity-related indicators on osteoporosis.zip

Background Osteoporosis and obesity are both major global public health problems. Observational studies have found that osteoporosis might be related to obesity. Mendelian randomization (MR) analysis could overcome the limitations of observational studies in assessing causal relationships. Objective This study aims to evaluate the causal potential relationship between obesity-related indicators and osteoporosis by using a two-sample MR analysis and to identify potential mediators. Method A total of 53 obesity-related indicators, 3,282 plasma protein lists, and 452 blood metabolite lists were downloaded from the public data set as instrumental variables, and the osteoporosis GWAS data of the MRC IEU Open GWAS database was used as the outcome indicators. Using two-sample univariate MR, multivariate MR, and intermediate MR, the causal relationship and mediating factors between obesity-related indicators and osteoporosis were identified. Results The IVW model results show that 31 obesity-related indicators may have a significant causal relationship with osteoporosis (P < 0.05), except for waist circumference (id: Ieu-a-71, OR = 1.00566); the remaining 30 indicators could reduce the risk of osteoporosis (OR: 0.983–0.996). A total of 25 plasma protein indicators may have a significant causal relationship with osteoporosis (P < 0.05), and 10 of them, such as ANKED46, KLRF1, and LPO, CA9 may have a protective effect on osteoporosis (OR: 0.996–0.999), while the other 15 such as ATP1B1, zinc finger protein 175, could increase the risk of osteoporosis (OR: 1.001–1.004). For blood metabolite indicators, except for alanine (id: Met a-469, OR: 1.071), the other six blood metabolite indicators including uridine and 1-linoleoylglycerophosphoethanolaminecan may have a protective effect on osteoporosis (P < 0.05, OR: 0.961–0.992). The direction of causal relationship of MR is all correct; the heterogeneity is all not significant and not affected by horizontal pleiotropy. Using multivariate and mediated MR analysis, it ...