This article is more than 5 years old. Information may no longer be current.

Machine learning approach identifies more than 400 genes tied to schizophrenia

Researchers have identified 413 genetic associations with schizophrenia across 13 brain regions by using a novel machine learning approach, according to a study published in Nature Genetics.

Laura M. Huckins, PhD, assistant professor of genetics and genomic sciences, and psychiatry, Icahn School of Medicine at Mount Sinai, and colleagues used transcriptomic imputation, a machine learning method that combines expression quantitative trait loci reference panels with large-scale genotype data found in genome-wide association studies (GWAS), to examine disease-associated gene expression changes in the brain.

The investigators used the largest expression quantitative trait loci panel for the dorsolateral prefrontal cortex (DLPFC) in the brain to generate a set of gene expression predictors and show their utility. To predict gene expression in schizophrenia GWAS data, they applied DLPFC predictors and 12 Genotype-Tissue Expression–derived neurological prediction models to 40,299 patients with schizophrenia and 65,264 matched controls.

Huckins and colleagues detected 413 genic associations across 13 brain regions in patients with schizophrenia. Using stepwise conditioning, they also identified 67 non-major histocompatibility complex genes and 36 significantly enriched pathways in the brain.

They also discovered that schizophrenia risk genes were expressed throughout development. Schizophrenia-associated genes were significantly coexpressed in both prenatal and postnatal development and in all four brain regions. Analysis indicated that the same genes do not contribute to this coexpression pattern throughout development, but instead that separate groups of genes contribute to early prenatal, late prenatal and postnatal clustering, according to the study.

“Our new predictor models gave us unprecedented power to study predicted gene expression in schizophrenia, and to identify new risk genes associated with the disease,” Huckins said in a press release. “By laying the groundwork for combining transcriptomic imputation and genome-wide association study findings, our hope is to not only elucidate gene development as it relates to schizophrenia, but also shape the future of research methods and design.”

In addition, different regions of the brain conferred different risks for schizophrenia, with most associations observed in the dorsolateral prefrontal cortex, according to the press release.

“As disease status may alter gene expression but not the germline profile, analyzing genetically regulated expression ensures that we identify only the causal direction of effect between gene expression and disease,” the researchers wrote in the full study. “Large, imputed transcriptomic datasets represent the first opportunity to study the role of subtle gene expression changes (and therefore modest effect sizes) in disease development.” – by Savannah Demko

Disclosure: The authors report no relevant financial disclosures.