Uncovering Functional Sequence Gaps in Human Reference Genomes using African Pan Genome Contig Sequences

Predominantly used human reference genomes, including GRCh38 and the gapless T2T-CHM13 references, remain limited in their representation of African genomic diversity. We analyzed African Pan- Genome (APG) contig sequences representing 296.5 Mb of African-ancestry-specific sequence not represented in GRCh38 to assess their representation and functional potential in newer long-read assemblies. Alignments to T2T-CHM13 and the 47 Human Pangenome Reference Consortium (HPRC) linear assemblies positioned 40% and 83% of APG contigs, respectively, with high identity and coverage. Most T2T-CHM13 placements corresponded to sequences absent from GRCh38 (94.5%) and were enriched in centromeric and satellite repeats (94.2%). Functional overlap included annotated genes (2.6%) and CpG islands (3.6%), with enrichment in immune, synaptic and intracellular signaling pathways. HPRC alignments revealed ancestry-associated patterns, with African and Admixed American genomes showing the highest numbers of unique contig placements and shared alignments. A subset of 742 APG contigs showed weak or no mapping to any of the reference assemblies. These contigs were not enriched for repeat elements, and [~]58% showed predicted gene content. These findings highlight persistent gaps in even the most complete reference genomes and underscore the importance of incorporating ancestry-enriched sequences into future genome frameworks to reduce reference bias and advance equitable discovery in genomics.