Fig. 2: Autoantibody profile of the cohort with associated clinical features.

A Most detected autoantibodies were found in a small number of individuals only (range 1–345, median = 17, mean = 37 seropositive individuals per autoantibody). No antibody binding was detected for 68 antigens (not shown in figure). Gene names of antigens are indicated for the antigens targeted by the 5% most prevalent autoantibodies. B The autoantibody count, i.e., the total number of detected autoantibodies in each sample, deviates slightly from a normal distribution centered on 25 autoantibodies per individual (Shapiro–Wilk test, W = 0.993, p = 0.035, Fig. S3). We defined individuals with autoantibodies toward fewer than 16 or more than 36 autoantigens (corresponding to the 5th and 95th percentiles; n = 25, 22, respectively) as having a low or high autoantibody count, respectively. C People with high autoantibody count had higher odds of being female, having a family history of obesity, and having a family history of psychiatric disorder other than Schizophrenia. People with low autoantibody count had higher odds of any present or past subjective thought disorder, and a trend for higher odds of clozapine use. RIN3 Ras and Rab interactor 3, ZNF688 Zinc finger protein 688, FAXDC2 Fatty acid hydroxylase domain-containing protein 2, CCDC85C coiled-coil domain-containing protein 85C, PAPOLA* Poly(A) polymerase alpha, Poly(A) polymerase beta, NAPG, gamma-soluble NSF attachment protein, and SVIL supervillin, ARFGAP1 ADP-ribosylation factor GTPase-activating protein 1, PRR20A* Proline-rich protein 20A, 20B, 20C, 20D, 20E, MEX3C RNA-binding E3 ubiquitin–protein ligase MEX3C, ATF-3 cyclic AMP-dependent transcription factor ATF-3, PTPRS receptor-type tyrosine-protein phosphatase S, ACBD7 Acyl-CoA-binding domain-containing protein 7, IQSEC1 IQ motif and SEC7 domain-containing protein 1, HECTD2 probable E3 ubiquitin–protein ligase HECTD2, FAM184A protein FAM184A, SCZ schizophrenia. Missing data shown in gray.