Disease associations for rare immune genetic variants require statistical or rule-based support

Validating weak gene-disease associations requires statistical support from large case-control studies. In contrast, strong (e.g., monogenic) gene-disease associations may be established from relatively few affected individuals, when independent orthogonal evidence converges. Here, we examine four publications associating heterozygous variants in five genes-TNIP1, P2RY8, BLK, BANK1, and SH2B3-with autoimmune disease (predominantly systemic lupus erythematosus [SLE]). While we do not dispute the molecular findings of these studies, we do not accept the proposed disease associations, which lack case-control evidence and where high minor allele frequencies in non-disease populations, indicate that the genotypes are far too common to represent Mendelian disease alleles. Additionally, 8 of 18 alleles were searchable in UK Biobank, and none showed enrichment in SLE. Until further evidence emerges, caution is warranted in ascribing disease association to heterozygous genotypes in TNIP1, P2RY8, BLK, BANK1, and SH2B3, and linkage to patients in clinical pedigrees should not be taken as inferring a monoallelic inborn error of immunity.