NEW YORK – New research suggests rare complex structural variants are overrepresented in human-specific parts of the genome, including sequences coding for neural genes and sites linked to psychiatric conditions such as schizophrenia or bipolar disorder.
For a paper published online in Cell on Monday, a team from the US, South Korea, Bulgaria, and Lebanon presented a probabilistic, machine-learning approach called ARC-SV for uncovering complex structural variations in the human genome using sequences generated with a range of short-read sequencing platforms.
"ARC-SV integrates evidence for and against each candidate in a statistically rigorous, data-driven manner," senior and co-corresponding author Alexander Urban, a psychiatry and behavioral sciences, genetics, and maternal and child health researcher at Stanford University, and his colleagues wrote, noting that the machine learning method pointed to complex SVs as "major sources of genetic variation" in the human "superpopulations" profiled.
When the team used ARC-SV to tally genetic variation in genome sequences for 4,262 individuals from continental superpopulations in the Americas, Africa, the Middle East, Europe, Central Asia, East Asia, Southeast Asia, and Oceania, it saw an overrepresentation of rare, complex SVs in sequences linked to speedy and human-specific evolution.
In particular, the complex SVs tended to be concentrated at sequences coding for neural genes and other genes implicated in corticogenesis, prompting follow-up experiments focused on the functional consequences of complex SVs in the brain.
With that in mind, the researchers did droplet-based single-nuclei RNA sequencing and single-nuclei ATAC-seq on post-mortem tissues from the dorsolateral prefrontal cortex, dorsal anterior cingulate cortex, and hippocampus brain regions in 40 brains from the "Human Brain Collection Core" (HBCC) collection, a set that includes brain samples from 15 individuals with bipolar disorder, 15 schizophrenia patients, and 10 unaffected control individuals.
Along with apparent interactions with alleles implicated in past genome-wide association studies of psychiatric conditions, the team saw ties between complex SVs and altered expression and chromatin accessibility in brain tissues and cell types considered, the researchers explained.
In brain samples from individuals with psychiatric conditions such as bipolar disorder or schizophrenia, meanwhile, they saw lower-than-usual expression of brain region- or cell type-specific genes — results backed up by subsequent analyses of available single-nucleus RNA-seq data for the dorsolateral prefrontal cortex region in brain samples from 32 individuals with schizophrenia and 47 control individuals, assessed for the CommonMind Consortium.
The team found further support for this pattern when it used ARC-SV to analyze RNA-seq data generated on brain samples from 287 individuals from GTEx, mapping complex SVs in relation to expression quantitative trait loci.
"Our results revealed that [complex SVs]-associated genes and open chromatin are significantly more downregulated across various cell types and brain regions for individuals with neuropsychiatric disorders, thus establishing [complex SVs] as a previously unknown component of the complex genetic architecture of human psychiatric disorders."