Skeletal muscle mass is central to mobility and metabolic health, yet it remains unclear whether muscle mass–related plasma proteomic signals align with free-living behaviours and physiology captured by wearable devices and smartphone-based logs (Taohealth). 

We used a two-step discovery–replication design. We first screened 2,919 plasma proteins quantified by Olink Normalized Protein eXpression (NPX) for association with bioimpedance-derived skeletal muscle mass index (SMI) in UK Biobank (UKB; n=43,434) using multivariable linear regression adjusted for age, sex, smoking, physical activity, and income, with Benjamini–Hochberg false discovery rate (FDR) control (q<0.05). We then defined replicated proteins as those showing concordant directions and q<0.05 in the Yamanashi Multi-omics Cohort (YMoC; n=162) in Japan. In YMoC, replicated proteins were further related to food-frequency questionnaire (FFQ) nutrients and to digital phenotypes from Taohealth logs (body weight, sleep timing, resistance training) and Fitbit-derived sleep and heart rate variability (HRV) metrics using Spearman correlations in YMoC. 

Seven proteins replicated across cohorts (IGSF3, CKB, IGFBP2, THBS4, CRYBB2, COMP, CTHRC1). In YMoC, IGSF3 was positively associated with α-carotene intake, whereas THBS4 was inversely associated with ethanol intake (FDR q<0.1). Taohealth logs showed that weight gain was positively associated with THBS4 and inversely associated with CKB and IGFBP2; later average bedtime was associated with lower COMP; and resistance training frequency was positively associated with CKB. Fitbit analyses showed inverse associations of changes in sleep efficiency with CKB and IGFBP2 (p<0.01), positive associations of changes in daily HRV with CKB, and an inverse association of changes in sleep duration with CRYBB2. 

These findings identify a cross-cohort muscle mass–related proteomic signature that links diet, sleep, autonomic function, and app-logged behaviours with extracellular matrix integrity (THBS4, COMP), growth-factor signaling (IGFBP2), energy metabolism (CKB), tissue remodeling (CTHRC1), and emerging stress/adhesion pathways (IGSF3, CRYBB2). Prospective and mechanistic studies are needed to clarify causal pathways and to evaluate potential applications in personalised strategies for skeletal muscle health.