Fig. 4: Distinct protein profiles of the short and long sleep groups.

a, Associations between short sleep (≤7 h, left) and long sleep (≥7 h, right) with proteins using linear regression. The regression coefficients (β) and corresponding two-tailed P values were derived using a t-test. The horizontal line indicates the Bonferroni-corrected significance threshold (two-tailed P < 0.01). The x axis represents β values and the y axis represents −log₁₀P values of the associations. b, Histogram of −log₁₀P values (two-tailed) of linear regression between short and long sleep groups with four panels of proteins. Red represents associations between sleep ≤7 h and proteins, and blue represents associations between sleep ≥7 h and proteins. c, Overlapping significant proteins between the middle-aged and older adult groups of short and long sleep groups. Red and green represent middle-aged and older adult groups, respectively, for the long sleep group, and blue and purple represent the middle-aged and older adult groups, respectively, for the short sleep group. d, Proteins with the most significant differences in their associations with short and long sleep groups within the middle-aged (left) and older adult (right) groups, respectively. The x axis represents the difference in the absolute correlation values (|r|) between the short and long sleep groups for protein associations using linear regression (|r_sleep≤7| − |r_sleep≥7|). Blue indicates proteins more strongly associated with the long sleep group, and red indicates proteins more strongly associated with the short sleep group. Abs. r value diff.: |r_sleep≤7| − |r_sleep≥7|.