Haplotype-resolved telomere-to-telomere assembly and haplotype-aware annotation pipeline enable high-quality reannotation of three Citrus genomes.

Citrus species are economically and nutritionally vital, with their fruits cultivated globally. Despite the publication of multiple genomes for Citrus, high-quality assemblies that achieve both haplotype resolution and telomere-to-telomere (HR-T2T) continuity remain scarce-pummelo (Citrus maxima) being a notable example. Compounded by limitations in gene annotation quality, these gaps hinder functional genomic research and genomics-assisted breeding. Here, we report the first high-quality HR-T2T genome assembly of pummelo, generated using PacBio HiFi and Oxford Nanopore sequencing. The two haplotype assemblies presented contig N50 values of 38.58 and 32.57 Mb, completeness scores of 99.36% and 99.66%, and nucleotide accuracies of 99.99994% and 99.99997%, respectively. We developed HapGene, a haplotype-aware annotation pipeline that integrates short-read RNA-Seq and long-read Iso-Seq data to enable unbiased annotation. Benchmarking showed HapGene captured 3% to 10% of genes missed or misannotated by conventional pipelines and reduces false haplotype-specific genes by 4- to 5-fold. Leveraging 380 Gb of newly sequenced and 2792 Gb of public transcriptomic data, we comprehensively annotated protein-coding and non-coding genes across three major Citrus crops (sweet orange, pummelo, and mandarin). This effort revealed 18 757-21 083 alternative splicing events, 1725-1853 resistance gene analogues, and 2392-3757 long intergenic RNAs (lincRNAs). Genomic and transcriptomic characterization of lincRNAs indicated their functional innovation (many associated with stress responses) in Citrus. Additionally, we revealed around one-third of genes exhibited tissue-specific allelic differential expression. Our work provides a critical genomic resource and analytical tool to advance Citrus genomic research, thereby driving progress in functional and evolutionary genomics while laying a robust foundation for precise genomics-assisted breeding.