Skip to main content
. 2020 Jul 6;12:101. doi: 10.1186/s13148-020-00890-w

Table 2.

Results of MANCOVA of OXTR DNA methylation between patients with OCD and healthy controls for men and women

Men Women
OCDa (n = 112) HCa (n = 56) OCDa (n = 39) HCa (n = 52)
F pb ηp2c F pb ηp2c
Statistics: Wilks λ = 0.822, F(4, 161) = 8.693, p < 0.001 Statistics: Wilks λ = 0.736, F(4, 84) = 7.525, p < 0.001
OXTR1 (exon2) CpG1 (-959) 40.94 ± 5.17 45.67 ± 4.88 20.012 < 0.001 0.109 42.94 ± 5.13 47.70 ± 3.89 21.784 < 0.001 0.200
CpG2 (-934) 46.33 ± 5.21 50.44 ± 4.52 22.873 < 0.001 0.122 49.11 ± 5.75 52.54 ± 3.14 10.631 0.002 0.109
CpG3 (-924) 58.08 ± 5.25 59.62 ± 3.82 1.835 0.177 0.011 60.36 ± 5.86 60.74 ± 3.75 0.032 0.858 < 0.001
OXTR2 (exon3) Mean value of CpG1-8 13.81 ± 5.46 13.34 ± 5.66 0.233 0.630 0.001 14.80 ± 4.49 14.14 ± 4.57 0.020 0.888 < 0.001

MANCOVA demonstrated significant between-group differences after controlling for age and education

MANCOVA multivariate analysis of covariance, OCD obsessive-compulsive disorder, HC healthy control

aMean ± standard deviation of raw data

bStatistical significance was set at p < 0.0125 after Bonferroni correction for the 4 CpG sites

cEffect size was calculated using partial eta squared and interpreted according to the rule of Miles and Shevlin (2001)

Medium effect (> 0.06)

Large effect (> 0.14)