Hidden DNA Repeats in the Brain May Shape Splicing and Disease Risk

Deep inside the human genome, stretches of DNA that repeat like broken records have long been dismissed as genetic noise. A new study published in Genome Biology suggests they are anything but. Researchers led by Yang Li, Jonathan Margoliash, Alon Goren, and Melissa Gymrek at the University of California San Diego have carried out one of the most comprehensive analyses to date of how short tandem repeats, or STRs, influence the way genes are spliced in the human brain. Their findings, drawn from 336 samples of human dorsolateral prefrontal cortex, reveal tens of thousands of repeat regions whose length variations are statistically linked to alternative splicing events, and they connect several of these repeat variants to devastating brain disorders including Alzheimer’s disease and schizophrenia.

Alternative splicing is one of the elegant tricks that complex organisms use to squeeze enormous diversity out of a limited set of genes. After a gene is transcribed into messenger RNA, the cell’s molecular machinery cuts and rejoins the RNA transcript in different patterns, producing multiple protein variants from a single gene. In the human brain, this process is particularly active and particularly important. When splicing goes wrong, the results can be catastrophic. Disrupted splicing has been implicated in a wide range of heritable phenotypes, including numerous psychiatric and neurological disorders, which is why geneticists have spent years hunting for the DNA variants that tip splicing off balance.

Until now, most of that hunting has focused on single nucleotide polymorphisms, the single-letter changes in DNA that genome-wide association studies can detect with ease, and on short insertions and deletions. STRs, which consist of short sequences of two to six DNA letters repeated over and over in tandem, have been harder to study. Standard sequencing pipelines struggle to read through repetitive DNA accurately, and the number of repeats at a given locus, known as copy number, can vary dramatically between individuals. Despite growing evidence that STR variation influences splicing, large-scale studies of STR-mediated splicing in the brain have been limited by small sample sizes of available RNA sequencing data and by the exclusion of certain classes of repeats from analysis.

The new study overcomes these obstacles by tapping a rich resource: deep RNA sequencing and SNP array data from 336 human dorsolateral prefrontal cortex samples collected by the Human Brain Collection Core, a repository maintained by the National Institute of Mental Health. The dorsolateral prefrontal cortex is a region central to executive function, working memory, and decision-making, and it has been a focal point for post-mortem studies of psychiatric illness. By combining genotype information at repeat loci with detailed measurements of splicing events across the transcriptome, the team could ask a simple but powerful question: when the number of repeats at a particular STR changes, does the splicing of nearby genes change with it?

The answer, overwhelmingly, was yes. The researchers identified 51,343 unique STRs for which copy number is significantly associated with one or more alternative splicing events of nearby genes. The team dubbed these loci spliceSTRs. The sheer scale of the result is striking: it suggests that tandem repeats are not occasional curiosities in the splicing landscape but a pervasive layer of regulation that previous studies, focused on SNPs and short indels, have largely missed. Each spliceSTR represents a case where an individual’s repeat length at a specific locus correlates with how a neighboring gene’s RNA transcript is assembled, potentially altering the protein products that gene ultimately produces.

Correlation, however, is not causation, and the authors were careful to separate statistical association from likely mechanism. They applied three orthogonal fine-mapping strategies, computational approaches that weigh competing genetic variants in a region to determine which are most likely to be the true causal drivers of an observed effect. This rigorous triage narrowed the field to 1,313 high-confidence fine-mapped spliceSTRs. Fine-mapping is essential in densely variable genomic regions, where many nearby variants travel together through generations and can masquerade as the causal signal. By triangulating across multiple methods, the researchers built a prioritized list of repeat loci that are strong candidates for directly modulating splicing rather than merely hitchhiking alongside the true culprit.

The study also probed the molecular machinery that might translate repeat length into splicing changes. RNA-binding proteins, or RBPs, are the workhorses of splicing regulation, latching onto specific sequence motifs in RNA transcripts to guide the splicing machinery’s decisions. The researchers found strong associations between the copy number of certain repeat units and the binding of specific RBPs. In other words, when an STR expands or contracts, it can create or destroy binding sites for these regulatory proteins, changing how the surrounding transcript is processed. This points to a general relationship between STR polymorphism and RBP-mediated splicing events, a unifying mechanism that could explain how tens of thousands of repeat loci exert their influence on the brain’s transcriptome.

Perhaps the most clinically compelling results came from co-localization analyses, which ask whether the genetic signals associated with a disease in large genome-wide association studies point to the same causal variants as the molecular signals identified in a tissue-specific study. The team used summary statistics from GWAS for six brain-related disorders and found multiple signals that may be driven by fine-mapped spliceSTRs. Among them was a previously identified GT-repeat that acts as a spliceSTR for the gene PLEKHA1 and has been associated with Alzheimer’s disease, now confirmed through this independent framework. More excitingly, the analysis uncovered a newly identified AGG-repeat spliceSTR in the gene SEPTIN3 that co-localizes with the genetic signal for schizophrenia, offering a fresh candidate for how that disorder’s risk architecture might act through altered RNA processing.

The implications reach beyond the specific genes involved. Splicing disruption has been implicated in a growing list of heritable conditions, and if repeat variation contributes to that disruption at this scale, then a substantial fraction of the genetic basis of brain disorders may be hiding in stretches of DNA that conventional analyses skip over. The findings suggest that future genetic studies of psychiatric and neurological disease should incorporate STR genotyping rather than treating repeats as unreadable territory. They also raise the possibility that some of the so-called missing heritability of conditions like schizophrenia, the gap between known genetic risk factors and observed patterns of inheritance, could be tucked away in variable repeat copy number.

The study, published open access on 15 October 2026 in Genome Biology, was funded in part by the National Human Genome Research Institute and the California Institute for Regenerative Medicine, and relied on data accessed through the database of Genotypes and Phenotypes under approved data access protocols. As a secondary analysis of pre-existing molecular data, it involved no new human participants or samples. For a field that has spent decades cataloguing the genome letter by letter, the message is humbling: some of the most consequential variation may not be a matter of which letters are present, but of how many times they repeat. In the prefrontal cortex, those repetitions appear to help write the rules of RNA assembly, and when the count goes wrong, the mind may pay the price.

Subject of Research: The role of short tandem repeat variation in regulating alternative splicing in the human prefrontal cortex and its links to brain disorders

Article Title: The contribution of short tandem repeats to splicing variation in the human prefrontal cortex

Article References: The contribution of short tandem repeats to splicing variation in the human prefrontal cortex. (n.d.). https://doi.org/10.1186/s13059-026-04296-4

Image Credits: AI Generated

DOI: 10.1186/s13059-026-04296-4

Keywords: short tandem repeats, alternative splicing, prefrontal cortex, schizophrenia, Alzheimer's disease, RNA-binding proteins, fine-mapping, colocalization analysis, splicing quantitative trait loci, human brain, genomics, Genome Biology