Comparative analysis of 22 tea plant genomes reveals conserved ARR subgroups and orthogroup-level variation

Background Gene-family analyses based on a single reference genome cannot resolve how deeply conserved functional groups coexist with intraspecific variation in gene content and copy number. Results We identified and compared the broad response-regulator family, comprising type-A, type-B, type-C, and pseudo-response regulators collectively termed ARR here, across 22 chromosome-scale tea plant genomes, integrating family composition with phylogenetic, duplication, local genomic variation, and coding-sequence analyses. In total, 979 ARR genes were identified, with 36–60 members per genome. All eight ARR subgroups were retained in nearly every genome, whereas individual genomes contained only 36–49 of the 105 pangenome orthogroups (OGs). Core and soft-core OGs accounted for 17.1% of OGs but 40.1% of ARR members, while shell OGs accounted for 50.5% of OGs and 56.0% of members. Copy gains occurred in 25 OGs and added 41 ARR copies, 34 of which belonged to shell OGs. Duplication histories differed among subgroups: whole-genome duplicates predominated in RRA, PRR1, and PRR5, whereas proximal and tandem duplicates dominated RRB-III. Transposable-element coverage varied primarily by ARR subgroup and gene region rather than by OG retention class or tea type. We detected 759 distinct structural variants overlapping 45 reference ARR genes and their neighborhoods, with 52.0% of overlaps located in 10-kb flanking regions. Among 5,417 retained orthologous pairs, 86.8% had Ka/Ks < 1. This predominance persisted after excluding low-Ks pairs, whereas subgroup contrasts were sensitive to the Ks threshold. Public transcriptomes from 12 accessions linked 255 ARR genes to tissue profiles, revealing descriptive organ-associated patterns within the matched subset. Conclusions Thus, tea ARR diversity is organized around a conserved deep subgroup structure, whereas most variation is concentrated in shell OG membership, subgroup-specific duplication histories, and local structural contexts.