Characterizing the genetic complexity of adaptation and trait evolution is a major emphasis of evolutionary biology and genetics. Incongruent findings from genetic studies have resulted in conceptual models ranging from a few large-effect loci to massively polygenic architectures. Here, we combine chromatin immunoprecipitation sequencing, Hi-C, RNA sequencing, and 40 whole-genome sequences from Heliconius butterflies to show that red color pattern diversification occurred via many genomic loci. We find that the red wing pattern master regulatory transcription factor Optix binds dozens of loci also under selection, which frequently form three-dimensional adaptive hubs with selection acting on multiple physically interacting genes. Many Optix-bound genes under selection are tied to pigmentation and wing development, and these loci collectively maintain separation between adaptive red color pattern phenotypes in natural populations. We propose a model of trait evolution where functional connections between loci may resolve much of the disparity between large-effect and polygenic evolutionary models.

表征适应和性状进化的遗传复杂性是进化生物学和遗传学的主要重点。遗传研究的不一致结果导致了从几个大效应位点到大规模多基因结构的概念模型。在这里，我们结合了染色质免疫沉淀测序、Hi-C、RNA 测序和来自Heliconius的 40 个全基因组序列蝴蝶表明红色图案多样化是通过许多基因组基因座发生的。我们发现红翼模式主调控转录因子 Optix 结合了几十个也在选择的基因座，这些基因座经常形成三维自适应中心，选择作用于多个物理相互作用的基因。许多选择下的 Optix 结合基因与色素沉着和翅膀发育有关，这些基因座共同保持自然种群中适应性红色图案表型之间的分离。我们提出了一个性状进化模型，其中基因座之间的功能联系可以解决大效应和多基因进化模型之间的大部分差异。