Table 1.
Relationships of core body temperature amplitude with metabolite rhythmicity across potential definitions of ‘circadian’ metabolites.
| Rhythmicity cut-offa | Model outcome: Number of rhythmic metabolites | Model outcome: Intra-individual metabolite period variability | Metabolites classified as rhythmic due to type 1 error (%)b | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Estimate (SE) | F (1,15) | p-value | R2 | Estimate (SE) | F (1,15) | p-value | R2 | ||
| ≥ 1/23 | 89.0 (393) | 0.0514 | 0.824 | 0.0034 | − 2.67 (2.75) | 0.948 | 0.346 | 0.059 | 69.23 |
| ≥ 2/23 | 91.5 (386) | 0.0561 | 0.816 | 0.0037 | − 2.67 (2.74) | 0.946 | 0.346 | 0.059 | 32.04 |
| ≥ 3/23 | 101 (358) | 0.0789 | 0.783 | 0.0052 | − 2.72 (2.72) | 0.997 | 0.334 | 0.062 | 10.52 |
| ≥ 4/23 | 101 (286) | 0.125 | 0.729 | 0.0082 | − 2.80 (2.72) | 1.07 | 0.318 | 0.066 | 2.60 |
| ≥ 5/23 | 84.0 (204) | 0.169 | 0.687 | 0.011 | − 2.85 (2.66) | 1.14 | 0.302 | 0.071 | 0.50 |
| ≥ 6/23 | 69.5 (122) | 0.326 | 0.576 | 0.021 | − 3.04 (2.52) | 1.46 | 0.246 | 0.088 | 0.08 |
| ≥ 7/23 | 66.7 (63.8) | 1.09 | 0.313 | 0.068 | − 3.22 (2.47) | 1.69 | 0.213 | 0.10 | 0.01 |
| ≥ 8/23 | 52.3 (37.0) | 2.00 | 0.178 | 0.12 | − 3.53 (2.32) | 2.32 | 0.149 | 0.13 | 0.00 |
| ≥ 9/23 | 50.7 (26.7) | 3.61 | 0.0769 | 0.19 | − 4.27 (2.10) | 4.13 | 0.0601 | 0.22 | 0.00 |
| ≥ 10/23 | 41.2 (22.1) | 3.46 | 0.0825 | 0.19 | − 4.54 (1.71) | 7.05 | 0.0180 | 0.32 | 0.00 |
| ≥ 11/23 | 33.2 (20.4) | 2.64 | 0.125 | 0.15 | − 4.43 (1.41) | 9.91 | 0.00664 | 0.40 | 0.00 |
| ≥ 12/23 | 35.7 (16.1) | 4.92 | 0.0423 | 0.25 | − 4.23 (1.45) | 8.57 | 0.0104 | 0.36 | 0.00 |
| ≥ 13/23 | 30.1 (15.0) | 4.05 | 0.0625 | 0.21 | − 4.75 (1.49) | 10.2 | 0.00610 | 0.40 | 0.00 |
| ≥ 14/23 | 33.7 (11.3) | 8.83 | 0.00952 | 0.37 | − 4.67 (1.45) | 10.4 | 0.00571 | 0.41 | 0.00 |
| ≥ 15/23 | 32.7 (11.0) | 8.90 | 0.00927 | 0.37 | − 5.57 (1.52) | 13.4 | 0.00233 | 0.47 | 0.00 |
| ≥ 16/23 | 30.0 (10.0) | 8.89 | 0.00932 | 0.37 | − 5.31 (1.65) | 10.3 | 0.00579 | 0.41 | 0.00 |
| ≥ 17/23 | 19.4 (7.04) | 7.60 | 0.0147 | 0.34 | − 6.24 (2.15) | 8.43 | 0.0109 | 0.36 | 0.00 |
| ≥ 18/23 | 16.4 (6.13) | 7.18 | 0.0171 | 0.32 | − 6.75 (2.40) | 7.93 | 0.0130 | 0.35 | 0.00 |
| ≥ 19/23 | 13.2 (4.65) | 8.02 | 0.0126 | 0.35 | − 6.02 (2.32) | 6.77 | 0.0201 | 0.31 | 0.00 |
| ≥ 20/23 | 9.38 (3.49) | 7.24 | 0.0167 | 0.33 | − 4.49 (2.48) | 3.29 | 0.0896 | 0.18 | 0.00 |
| ≥ 21/23 | 7.02 (2.74) | 6.56 | 0.0217 | 0.30 | − 4.76 (2.53) | 3.55 | 0.0791 | 0.19 | 0.00 |
| ≥ 22/23 | 4.54 (1.21) | 14.2 | 0.00188 | 0.49 | − 0.32 (2.82) | 0.0129 | 0.911 | 0.00086 | 0.00 |
Data are estimates, statistical significance, and effect sizes of linear regression models predicting number of rhythmic metabolites, and intra-individual metabolite period variability, from core body temperature amplitude.
aRhythmicity cut-offs were used to define sets of ‘circadian’ metabolites. Each metabolite was required to exhibit significant circadian rhythmicity in ≥ N/23 participants for inclusion as a ‘circadian’ metabolite.
bExpected type 1 errors were simulated across 1000 sets of 929 metabolites*23 participants, and average percentage of metabolites defined as ‘circadian’ due to expected type 1 error alone were calculated for each rhythmicity cut-off.