Thread by @NeonGolden27

Daniel Weinberger: been arguing for years that to understand schizophrenia we need to think about how brains are built and how different trajectories of post natal development affect the brain #BAP2018

Weinberger: schizophrenia not a disease in a classical medical sense but a state of brain formed through development #BAP2018

Weinberger: Schizophrenia runs in families. Roughly 1% of cases have loss of function mutations which is major risk factor, about 2% cases have CNVs and 97% cases cumulation of common variants in particular combination. #BAP2018

Weinberger: How do genes translate to schizophrenia (+other developmental disorder) risk? If affecting how brain is built, should be preferentially expressed during fetal life. Confirmed to be plausible by brain tissue study comparing to neuro degenerative illness #BAP2018

Weinberger: Can look at gene "junctions" which span two exons. Some genes, when look at gene level, seems developmentally static but have different patterns of expression at junctions across development (isoform shift). PGC2 risk loci enriched for isoform switch genes. #BAP2018

Weinberger: 50k differentially expressed region's (DERs) over development and aging, 40% non exonic regions, previously thought non expressing, majority highest expression in fetal life. These DERs enriched in schizophrenia GWAS positive regions. #BAP2018

Weinberger: GWAS risk loci don't tell you specific genes, like a street address but without information on the floor and room and exact location. Need to understand biology. #BAP2018

Weinberger: Most SCZ risk loci are non coding. Affect gene function in numerous ways but all read out in the transcriptome so RNA sequencing critical #BAP2018

Weinberger: If we want to understand genes, need to model molecular role in tissue, not do knock out studies which only model very rare situation. #BAP2018

Weinberger: 230k CpG sites associated with developmental stages. PGC2 risk regions enriched for prenatal development, but not age of onset.
Schizophrenia associated DNAm changes strongly enriched for fetal post natal changes but depleted for changes near age of onset.#BAP2018

(similar to findings from from @eilis_hannon and @PsyEpigenetics)

Weinberger: Suggests SCZ epigenetic marks from early life survive throughout life but marks from early adulthood do not (maybe too unpredictable or unstable) #BAP2018

Weinberger: Three SCZ things we know. Polygenic risk, role of early environment (DZ twins concordance twice of non twin siblings, share uterus), unexplained higher incidence in males. #BAP2018

Weinberger: 100 years of research find complicated pregnancies a risk factor for SCZ . Does combining genomic risk and information on intrauterine environment + obstetric complications increase prediction of scz risk? Yes, non additive interaction. #BAP2018

Weinberger: 7 fold increase in SCZ risk if high PRS score and obstetric complications. Higher PRS scores in SCZ accounted for by obstetric complications. Replicated in numerous samples. No association between genes and complications. #BAP2018

Weinberger: Has to be serious obstetric complication to interact with genetic risk for SCZ. Via placenta? Most routinely under studied organ. Schizophrenia PRS risk loci highly expressed in placenta. Especially in placenta from complicated pregnancies. #BAP2018

Weinberger: placental-implicated genes at orthogonal biology. Oxidative stress, protein folding. May prevent placental support for fetal development. Animal studies indicate initiates immune response. Placenta basically an alien organ + barrier, constantly under stress. #BAP2018

Weinberger: Even more up regulation of these genes if placenta from male off spring, so placental factors may be part of explaining of higher male incidence of SCZ. #BAP2018

Weinberger: Summing up, most significant SCZ genetic risk variants partly converge on developmental trajectory sensitive to pre and perinatal adversity via placental disruption. Preserving prenatal health may be primary form of SCZ prevention. #BAP2018