Table 3. Summary of the variables with evidence for interactions with FTO.

Variable	Estimate	95% CI	P value
	−0.19	[−0.34, −0.05]	1.0e−02
Activity score	−0.35	[−0.6, −0.1]	5.8e−03
	−0.23	[−0.36, −0.11]	3.1e−04
	−0.28	[−0.43, −0.13]	2.8e−04
Alcohol frequency	−0.12	[−0.38, 0.14]	3.6e−01
	−0.24	[−0.37, −0.11]	3.0e−04
	0.25	[0.11, 0.4]	7.0e−04
Diet score	0.43	[0.17, 0.69]	1.1e−03
	0.30	[0.17, 0.43]	5.0e−06
	0.13	[0.04, 0.22]	4.6e−03
Sleep squared	0.14	(−0.01, 0.3]	7.3e−02
	0.13	[0.06, 0.21]	8.0e−04
	0.23	[0.08, 0.38]	2.9e−03
Added salt	0.17	[−0.09, 0.44]	1.9e−01
	0.21	[0.08, 0.34]	1.2e−03

BMI, body mass index; CI, confidence interval; FTO, fat mass and obesity associated.

The table shows the estimated interaction effect with FTO expressed as the % change in BMI per copy of FTO and per s.d. of the variable. The first line for each variable gives the estimate in the British Sample, the second line gives the estimate in the Diverse Sample and the third line gives the combined estimate. ‘Sleep squared' refers to squared deviations from mean sleep duration.