Table 1.
Description of the sleep genome-wide association studies (GWAS) included in the two-sample Mendelian randomization analyses
| Trait definition (units) |
Participants, n(cases/controls
for binary traits) |
Loci identified, n, in GWAS | F statistic | |
|---|---|---|---|---|
| Self-report measures | ||||
| Chronotype20 | Whether a person identifies as being a ‘morning person’ or an ‘evening person’ (ordered categorical variable of definitely a morning person, more a morning than an evening person, do not know, more an evening than morning person and definitely an evening person)a | 449 734 | 351 | 33.1 |
| Sleep duration21 | Average number of hours slept in 24 h, including naps (continuous variable, hours) | 446 118 | 78 | 39.6 |
| Short sleep duration21 | Person has an average of 6 h or less per night vs 7-8 h per 24 h (binary variable of yes/no) | 411 934 (106 192/305 742) | 27 | 25.6 |
| Long sleep duration21 | Person has an average of 9 h or more per night vs 7-8 h per 24 h (binary variable of yes/no) | 339 926 (34 184/305 742) | 8 | 30.6 |
| Frequent insomnia22 | Person has trouble falling asleep at night or wakes up in the middle of the night (binary variable of usually vs never/rarely) | 453 379 (131 480/321 899) | 48 | 41.7 |
| Excessive daytime sleepiness23 | Person dozes off or falls asleep during the day without meaning to (ordered categorical variable of never or rarely, sometimes, often and all the time) | 452 071 | 37 | 42.3 |
| Daytime napping | Person naps during the day (ordered categorical variable of never, sometimes, usually) | 452 633 | 112 | 46.1 |
| Accelerometer measures | ||||
| L5 timing24 | Timing of the least active 5 h of the day (continuous variable of hours elapsed since previous midnight; provides indication of phase of most restful hours with later times indexing greater tendency towards ‘eveningness’) | 85 205 | 6 | 55.3 |
| Sleep duration24 | Average number of hours of nocturnal sleep per night (continuous variable, hours) | 84 810 | 11 | 52.1 |
| Sleep fragmentation24 | The average number of nocturnal sleep episodes separated by at least 5 min of wakefulness per night (continuous variable, number of episodes) | 84 810 | 21 | 38.6 |
a
Note that in the original chronotype GWAS, categories were ordered from more ‘eveningness’ to more ‘morningness’. In this analysis, to ensure that the ordinal chronotype variable correlated positively with the accelerometer-measured measure of L5 timing, single nucleotide polymorphism (SNP)-exposure coefficients for chronotype were reordered from more ‘morningness’ to more ‘eveningness’ (where ‘definitely a morning person’ is the reference category).