Modern tomatoes have narrow genetic diversity limiting their improvement potential. We present a tomato pan-genome constructed using genome sequences of 725 phylogenetically and geographically representative accessions, revealing 4,873 genes absent from the reference genome. Presence/absence variation analyses reveal substantial gene loss and intense negative selection of genes and promoters during tomato domestication and improvement. Lost or negatively selected genes are enriched for important traits, especially disease resistance. We identify a rare allele in the TomLoxC promoter selected against during domestication. Quantitative trait locus mapping and analysis of transgenic plants reveal a role for TomLoxC in apocarotenoid production, which contributes to desirable tomato flavor. In orange-stage fruit, accessions harboring both the rare and common TomLoxC alleles (heterozygotes) have higher TomLoxC expression than those homozygous for either and are resurgent in modern tomatoes. The tomato pan-genome adds depth and completeness to the reference genome, and is useful for future biological discovery and breeding.

现代番茄的遗传多样性狭窄，限制了其改良潜力。我们介绍了一个番茄全基因组，使用725个系统发育和地理代表性种质的基因组序列构建而成，揭示了参考基因组中缺少的4,873个基因。存在/不存在变异分析表明，在番茄驯化和改良过程中，大量的基因缺失以及大量的基因和启动子阴性选择。丢失或阴性选择的基因可以丰富重要的性状，尤其是抗病性。我们在TomLoxC中鉴定出罕见的等位基因在驯化过程中选择反对的启动子。定量性状基因座作图和转基因植物分析揭示了TomLoxC在类胡萝卜素生产中的作用，这有助于产生理想的番茄风味。在橙色阶段的水果中，同时含有稀有和常见TomLoxC等位基因（杂合子）的种质的TomLoxC表达要高于纯合子，并且在现代番茄中可以复活。番茄全基因组为参考基因组增加了深度和完整性，可用于将来的生物学发现和育种。