NLRSeek: A reannotation-based pipeline for mining missing NLR genes in sequenced genomes

Nucleotide-binding leucine-rich repeat (NLR) proteins function as receptors and signaling factors in plant immune systems. Identifying the genes encoding NLR proteins in genomic sequences provides crucial information for breeding disease-resistant crops. However, NLR proteins are frequently misannotated during automated proteome prediction and downstream identification tools that rely on proteomic data struggle to recover these missing NLRs. To address this problem, we developed NLRSeek (https://github.com/Wang-Mengda/NLRSeek), a genome reannotation-based pipeline for NLR identification. This workflow integrates de novo detection of NLR loci at the genome level with targeted genome reannotation, systematically reconciling these results with existing annotations to produce a comprehensive set of NLR predictions. Our pipeline identified a larger number of NLRs than other NLR annotation tools: even in the well-annotated model plant Arabidopsis thaliana, NLRSeek identified a previously unannotated NLR gene whose expression and translation were confirmed by transcriptome and ribosome-profiling data. The NLRSeek pipeline showed particularly strong performance for non-model species with incomplete annotations. For example, in the yam species Dioscorea zingiberensis, Dioscorea tokoro, and Dioscorea dumetorum, NLRSeek identified 33.8 ​%-127.5 ​% more NLR genes than conventional methods. Importantly, 45.1 ​% of the newly annotated NLRs exhibited detectable expression, suggesting that they are true genes that were previously overlooked. Analysis of the newly identified sequences revealed that NLRs have undergone expansion in D. zingiberensis through tandem duplication, an insight that was not attainable using previous NLR annotation tools. Our novel NLR identification pipeline may reveal untapped genetic resources for engineering disease-resistant crops.