<i>GABRB2</i> association in PANSS score-based SCZ subgroups.

The Beijing SCZ cohort (‘All’) was partitioned into ‘High score’ and ‘Low score’ subgroups by applying the K-Means clustering function to the PANSS scores for positive symptoms. The p-values were obtained from case-control association analysis using the likelihood ratio test. Significant p values (p<0.05) that passed the global permutation test are shown in bold font. The odds ratio (OR) and its 95% confidence level (CI) were based on allele frequencies.