This PNAS paper is the result of a very fruitful collaboration with Frank Chan and Marek Kucka from the Max Planck Institute in Tübingen, Germany, who developed a new linked-read sequencing technology. Their method allows extremely cheap sequencing and provides high-quality data that even allows to assess chromosomal rearrangements for very low costs. Our joint paper introduces this new method and applies it to a fascinating case of butterflies in Ecuador. The study of these butterflies was initiated by Chris Jiggins (U Cambridge) and Patricio Salazar (U Edinburgh). Two Heliconius butterfly species that mimick each other (look alike) each have a highland and a lowland subspecies that differ in wing coloration. At mid-elevation, in both species these two subspecies meet and interbreed giving rise to a new hybrid subspecies that combines the colour pattern elements of the highland and the lowland subspecies. In this study we show which genes underlie the wing colour differences between the subspecies and reveal amazing parallelism between the two species. In both species, there are four relevant genes that show highly similar patterns of differentiation between the three subspecies. We show that strong selection acts on these genes and that the different subspecies are maintained by positive frequency dependent selection. This form of selection is imposed by birds learning to recognize the colour patterns as warning signal of their toxicity. Butterflies with the locally most common combination of colour patterns have the advantage that birds are most likely to already know that they should not eat them.