Hi @Stephen Kelly, so the current plan has 2 big steps. First (what we call v1): Update igenomes and add missing indexes and files so that we can finally use rnaseq again with igenomes, so basically a v0.4 of igenomes (https://github.com/ewels/AWS-iGenomes/#changelog). • Updating the manifest so we know what we have in igenomes https://github.com/ewels/AWS-iGenomes/pull/22 • Convert manifest to assets: https://github.com/nf-core/references/pull/52 • Build and publish new indexes in igenomes (Probably need to download a couple gtf for that) Second (what we call v2): reboot igenomes and have at the same time a way of building references and serve it globally, but also allow for people to modify the "recipe" and have their own local flavor if they see fit. Reproducibility has always been at the center of my mind, so each file produced will be fully versioned with meta data and info on how it was build (downloaded etc...)