NLR DNA copy-number expansion outpaces accumulation of predicted coding-intact copies

Whether nucleotide-binding leucine-rich repeat (NLR) DNA copies and predicted coding-intact copies accumulate proportionally within homologous neighbourhoods remains unresolved. We compared these counts across 173 anchor-defined neighbourhoods in 11 Capsicum genomes. After accounting for neighbourhood and genome effects, predicted coding-intact copies increased less than proportionally with DNA copies. DNA-copy doubling corresponded to 1.48-fold and 1.84-fold increases in expected coding counts under strict and inclusive definitions, respectively. The strict association persisted across sequence and sampling controls, with highly expanded arrays and heterogeneous coding recovery affecting interpretation. An independently defined panel of 17 Arabidopsis thaliana accessions reproduced sublinear scaling using curated gene counts. Matched pepper RNA-seq showed a transcript-support gradient across predicted coding states. At DH06, a near-megabase segmental duplication in C. pubescens contained intact and disrupted family pairs. Copy-specific HiFi reads placed the same F396 coding lesions in both physical blocks. Together, these results reveal a recurrent relationship between NLR array size and coding composition in the surveyed plant systems. Interpreting DNA expansion together with coding state provides a more informative basis for comparing NLR repertoires and prioritizing immune-gene candidates.