RSparsePro Improves Genetic Fine-Mapping Despite Linkage Disequilibrium Mismatch

Described in an article published on 06 November 2026 in Nature Genetics, the system couples a probabilistic graphical model with an efficient variational inference algorithm. The framework directly addresses the challenge of linkage disequilibrium discordance, providing enhanced analytical reliability for genetic association studies that depend on external reference panels to pinpoint causal variants.

Fine-mapping methods based on summary statistics from genome-wide association studies and linkage disequilibrium information are widely used across genetics to identify potential causal variants. However, linkage disequilibrium mismatch between an external linkage disequilibrium reference panel and the genome-wide association study population is common in contemporary analyses.

To overcome this limitation, the investigators created RSparsePro to maintain statistical accuracy across datasets affected by reference mismatch. In simulation studies evaluating varying degrees of linkage disequilibrium mismatch, the authors systematically compared RSparsePro with existing fine-mapping methods. Across these simulated conditions, which included benchmark analyses modeling scenarios with three causal variants, RSparsePro identified credible sets with consistently higher power and coverage than existing methods, demonstrating substantial resilience when reference linkage disequilibrium patterns deviate from the study sample.

The researchers further evaluated RSparsePro using empirical biological data by fine-mapping complex trait risk loci and cis-protein quantitative trait loci. In one primary analysis, the team summarized fine-mapping results for 1,353 plasma proteins with a cis-protein quantitative trait locus.

Across these fine-mapping analyses of complex trait risk loci and cis-protein quantitative trait loci, RSparsePro identified credible sets characterized by consistently higher enrichment of variants with potential functional impacts compared to alternative tools. In addition, the credible sets generated by RSparsePro achieved consistently higher replication rates across different studies and across different linkage disequilibrium reference panels, confirming the reproducibility of the methodology when evaluated across diverse datasets.

Beyond protein quantitative trait loci, RSparsePro successfully identified biologically relevant variants in drug target genes and provided candidate regulatory insights. When applied to genome-wide significant loci from ancestry-specific genome-wide association studies of low-density lipoprotein cholesterol using the 1000 Genomes linkage disequilibrium reference panel, the method identified potential regulatory elements and implicated prospective regulatory mechanisms for low-density lipoprotein cholesterol, underscoring its utility for resolving complex trait genetics.

Subject of Research: Genetic fine-mapping under linkage disequilibrium mismatch

Article Title: Robust fine-mapping in the presence of linkage disequilibrium mismatch

Article References: Zhang, W., Lu, T., Sladek, R., Dupuis, J., & Lettre, G. (2026). Robust fine-mapping in the presence of linkage disequilibrium mismatch. Nature Genetics. https://doi.org/10.1038/s41588-026-02780-y

Image Credits: AI Generated

DOI: 10.1038/s41588-026-02780-y

Keywords: statistical genetics, fine-mapping, linkage disequilibrium, genome-wide association studies, RSparsePro, variational inference, quantitative trait loci, LDL cholesterol, Robust, presence, linkage, disequilibrium