Massively parallel characterization and predictive modelling of neuronal regulatory variation

Massively parallel characterization reveals context-dependent and non-additive regulatory effects of closely spaced variant pairs

Precise control of gene expression relies in part on cis-regulatory elements (CREs), including enhancers. Genetic variation within enhancers can alter their regulatory activity and contribute to variation in gene expression, yet the effects of multiple nearby variants within the same enhancer remain poorly understood. To address this, we designed a massively parallel reporter assay (MPRA) to test the regulatory effect of 7,285 pairs of close-proximity single-nucleotide variants (SNVs) selected from blood-specific and broadly active enhancers. To enrich for functionality, we required at least one variant in each pair to be annotated as an eQTL in eQTLGen dataset. We measured the regulatory activity of all four haplotypes in K562 leukemia cells, where 57% of tested variant pairs had at least one derived haplotype that differed significantly in activity from the ancestral haplotype. The effects of individual variants frequently depended on the allelic background provided by the neighboring variant, with some variants showing opposite effects in different allelic backgrounds. Among a smaller, high-confidence subset selected for analysis of additivity, 59% (105/178) of variant pairs showed non-additive effects, and non-additive pairs were located closer together than additive pairs. Among pairs in which both single-derived haplotypes increased activity, the double-derived haplotype generally showed a smaller effect than expected under additivity. Together, our results demonstrate that nearby variants within the same enhancer can jointly shape regulatory activity and highlight the importance of considering local allelic context when interpreting the functional effects of regulatory variation.

Prediction of human missense variant effects from functional evidence

Regulatory Evolution and the Genetic Basis of Human Brain Expansion.

Scalable context-dependent single-cell eQTL mapping reveals disease-relevant regulatory variation beyond static models

Many disease-associated variants are thought to act through gene regulation, yet conventional eQTL mapping explains only a fraction of GWAS loci, potentially because regulatory effects vary across cellular states and environments. We present CASTIE, a scalable Poisson mixed-model framework that directly models sparse single-cell read counts and enables genome-wide testing of genotype-by-context interactions without pre-screening for static effects. Applying CASTIE to 1.2 million peripheral blood mononuclear cells from 982 OneK1K donors identified 3,155 context-dependent eQTL associations, including 2,022 eGenes without detectable static effects. These associations yielded 374 colocalizations across 94 traits, representing 270 unique loci, of which 197 were not recovered using the corresponding static eQTLs. The colocalizations linked trait associations to specific cellular contexts and genes including GCHFR, RNASET2 and ATP1A3. In adipose-derived mesenchymal stem cells exposed to metabolic stimulations, CASTIE increased eGene discovery by 36-92% across cell populations and identified stimulation-dependent regulatory effects at metabolic trait loci. Thus, modeling cellular context reveals disease-relevant regulatory variation beyond static eQTL mapping.

Sequence-to-function deep learning decodes human cis-regulatory evolution

Deciphering the regulatory consequences of sequence divergence across human evolution is essential to understanding the molecular basis of human-specific traits and disease. Although millions of derived alleles distinguish humans from great apes, only a small fraction are likely to influence human-specific traits. Previous studies have focused on regions of elevated sequence divergence, assuming that rapid evolution reflects functional adaptation, yet individual high-impact regulatory mutations evade such scans. Here, we apply sequence-to-function deep learning to predict chromatin accessibility across modern human, archaic hominin, and great ape personalized genomes, identifying lineage-specific cis-regulatory elements (linCREs) across diverse cellular contexts. Compared to conserved elements, linCREs are shorter, less pleiotropic, less conserved, and enriched in neurodevelopmental pathways. Many linCREs occur in regions with limited sequence divergence that acceleration-based approaches would overlook. We validate lineage-specific enhancer activity through luciferase reporter assays and demonstrate that a single motif-generating derived allele nominated by model interpretability tools drives a hominin-specific neurodevelopmental enhancer.

Unlocking cis-regulatory landscapes across 500 million years of evolution and disease mechanisms

Abstract Genomic DNA encodes regulatory information that determines where, when, and to what extent genes are expressed. Theoretically, we should be able to identify these transcriptional “instructions” by examining genomic DNA sequence alone, yet this has remained challenging. Here we present the Vertebrate Regulatory MOdule Detector (VRMOD), a method that accurately predicts gene regulatory sequences using only the query genomic sequences. We applied VRMOD to 309 Ensembl genomes, generating a compendium of high-resolution, genome-position-fixed cis-regulatory modules without parameter tuning. We performed extensive computational evaluation and experimental validation of VRMOD predictions. Notably, VRMOD predicted three sub-enhancers within the human hs52 enhancer at the FTO locus from the VISTA database, including one missed by existing methods. Using a chicken embryo system and 3D tissue imaging, we showed that each sub-enhancer exhibits restricted spatiotemporal activity within specific subsets of tissues where the full enhancer is active. We further demonstrated VRMOD’s utility for identifying evolutionarily non-conserved enhancers, annotating regulatory sequences in non-model organisms, and identifying candidate disease-causal variants. Collectively, VRMOD provides a universal coordinate reference system for regulatory sequences across 309 vertebrate genomes and enables genome-wide annotation of non-coding regulatory elements in any vertebrate species using genomic sequence alone.