Table E1.

Ninety-five percent confidence intervals for model B parameters: π₁, ${\sigma }_b^2$, and ${\sigma }_0^2$ (see Table 1)

Phenotype	π 1	${\mathit{\sigma }}_b^2$	${\mathit{\sigma }}_0^2$
SCZ 2014	[5.03e−2, 5.54e−2]	[1.27e−6, 1.42e−6]	[1.073, 1.077]
BIP	[3.77e−2, 8.05e−2]	[8.26e−7, 1.52e−6]	[1.009, 1.012]
CD	[9.10e−4, 1.00e−3]	[4.61e−5, 5.36e−5]	[1.136, 1.138]
UC	[1.05e−3, 1.27e−3]	[3.13e−5, 4.09e−5]	[1.116, 1.118]
CAD	[1.73e−3, 2.03e−3]	[1.01e−5, 1.19e−5]	[0.919, 0.924]
AD Chr19	[3.11e−4, 5.57e−4]	[6.76e−5, 1.35e−4]	[1.083, 1.093]
AD NoC19	[7.44e−4, 1.35e−3]	[1.26e−5, 2.40e−5]	[1.039, 1.041]
ALS Chr9	[8.45e−3, 1.40e−2]	[5.51e−6, 9.17e−6]	[0.985, 0.991]
Edu	[1.39e−2, 1.48e−2]	[1.56e−6, 1.87e−6]	[0.937, 0.943]
IQ 2018	[1.24e−2, 1.31e−2]	[6.55e−7, 8.45e−7]	[1.168, 1.178]
Height 2010	[9.72e−4, 1.07e−3]	[3.74e−5, 4.42e−5]	[0.894, 0.900]
Height 2014	[1.11e−3, 1.18e−3]	[3.54e−5, 3.93e−5]	[1.563, 1.570]
Height 2018	[2.38e−3, 2.61e−3]	[7.83e−6, 9.50e−6]	[2.114, 2.124]
HDL	[2.37e−3, 2.71e−3]	[1.00e−5, 1.16e−5]	[0.906, 0.909]
LDL	[5.09e−3, 6.59e−3]	[2.83e−6, 3.80e−6]	[0.914, 0.917]
BMI GIANT 2015	[1.52e−3, 1.55e−3]	[2.12e−5, 2.17e−5]	[0.848, 0.857]
BMI GIANT-UKB 2018	[2.31e−3, 2.36e−3]	[1.58e−5, 1.61e−5]	[1.712, 1.720]
TC	[1.02e−3, 1.28e−3]	[1.44e−5, 1.95e−5]	[0.920, 0.923]

These confidence intervals are likely underestimates (too narrow), due to LD and a consequent over-counting of the number of degrees of freedom (independent z-scores).