Table 1.

Enrichment analysis of predicted pathogenic rare variants shared in affected or unaffected relatives, after combining data across 15 Australian BD families

	PEVs (498 genes)		LNVs (495 genes)
Gene set (n genes)	O (E)	P-val (corr P-val)	O (E)	P-value (corr P-val)
PSD (1445)	59 (41.5)	0.002 (0.024)	40 (38.5)	0.423 (1)
FMRP targets (837)	44 (39.6)	0.218 (1)	44 (38.2)	0.143 (1)
De novo PSY (1627)	71 (66.2)	0.254 (1)	67 (68.7)	0.627 (1)
NMDARs (61)	2 (1.7)	0.527 (1)	3 (2)	0.321 (1)
ARC (27)	0 (0.5)	1 (1)	1 (0.7)	0.541 (1)
Mitochondrial (1124)	22 (24.3)	0.727 (1)	21 (22.6)	0.679 (1)

Hypergeometic p-value, examining hits per category unadjusted for gene length, genic intolerance or sequence coverage of reference genes; AdjP-value, examining hits per category after adjustment for gene length, genic intolerance and sequence coverage of reference genes; AdjP-values that exceed Bonferroni correction for 12 independent tests are indicated in bold

PEVs potentially etiologic variants, which were shared in ≥3 affected relatives ± one unaffected relative (Supplementary Table S1), LNVs likely neutral variants were shared in 1 to 3 unaffected ± an affected relative (Supplementary Table S1), O number of genes observed in this category, E number of genes expected in this category, PSD genes expressed in the postsynaptic density³³, De novo PSYCH de novo variants found in autism, schizophrenia and intellectual disability²⁰, NMDARs N-methyl-d-aspartate (NMDA) receptor gene set²⁰, ARC activity-regulated cytoskeleton-associated protein gene set²⁰, Mitochondrial autosomal genes encoding mitochondrial localized proteins³⁵