Introduction: Elevated plasma triglycerides (TG) are associated with risk of cardiovascular disease and are modifiable through lifestyle interventions such as regular exercise. However, TG responses to regular exercise are characterized by significant inter-individual differences. Hypothesis: We hypothesized that baseline levels of circulating proteins and metabolites are associated with TG response to exercise and can predict exercise-induced changes in plasma TG traits. Methods: We measured circulating proteins (n=4979 proteins) and metabolites (n=300) in 650 Black and White adults of the HERITAGE Family Study who completed 20 weeks of exercise training and had complete data on TG traits. We investigated two TG-related traits that significantly improved with training: fasting TG and large TG-rich lipoprotein particle concentration (LTRLP). The association between baseline analyte values and exercise-induced changes in TG traits were examined using linear mixed models adjusted for age, sex, race, BMI, baseline trait value, and the random effect of family membership. Significance was determined as FDR<0.05. Molecular signatures of TG trait responses were generated via elastic net regression tuned using leave-one-out cross validation. Results: Regular exercise significantly reduced plasma TG (-1.03±1.3 mmol/L, p=0.01) and LTRLP (-0.24±2.2 nmol/L, p= 0.007), with their training-induced changes moderately correlated (r=0.4, p<0.0001). We identified largely distinct panels of baseline analytes associated with changes in TG and LTRLP ( Table 1 ). Similarly, elastic net regression yielded distinct models of 99 analytes for TG and 315 analytes for LTRLP responses to exercise with accuracy of RMSE=1.28 mmol/L for TG and 2.16 nmol/L for the LTRLP model. Conclusions: We identified panels of proteins and metabolites associated with exercise-induced changes in TG traits and demonstrate, within our data, circulating analytes hold promise for predicting the exercise responsiveness of plasma TG-related traits.