Rare disease genomics in an era of human pangenomics and telomere-to-telomere genome references

Despite considerable efforts investigating the genetic aetiology of rare diseases in the past decades, approximately 50% of cases remain without a genetic diagnosis. Many missing diagnoses can be attributed to the limitations of short-read sequencing (SRS), compounded by (mis)-alignment to incomplete and inaccurate reference genomes such as GRCh37/38. SRS cannot resolve many regions that are challenging to map, including large contiguous tandem repeats, segmental duplications (SDs), sites of complex structural variants (SV), or highly diverged population-specific loci. Long-read sequencing (LRS) technologies have delivered the first complete human genome assembly, T2T-CHM13. Compared to GRCh38, T2T-CHM13 resolves the remaining 8% of the genome, corrects structural errors and improves both SRS- and LRS-based read mapping and variant discovery. LRS has also facilitated the generation of high-quality, haplotype-resolved assemblies from globally diverse cohorts, enabling the construction of pangenome references for multiple ancestral groups. By representing more human genomic variation, a pangenome reference can improve mapping and variant calling accuracy. These new genome resources represent alternative reference paradigms that have the potential to uncover pathogenic variants underlying unsolved rare genetic diseases. Here, we examine the limitations of GRCh38 for rare disease variant discovery and explore how emerging resources like T2T-CHM13 and pangenomes can improve accuracy. We highlight key studies that have leveraged these references to improve diagnostic outcomes and discuss the potential for broader adoption. Finally, we consider the current barriers to research and clinical implementation and outline available resources and tools to expedite the transition to these new reference models.