Skip to content

easymode model library

This sections lists the features for which pretrained easymode models are currently available. We have split them up into two categories: 🐦 species and 🏞️ biomes. The distinction is purely descriptive.

The 🐦species include ribosomes, microtubules, vaults, and other well-defined macromolecular assemblies that you might consider averaging. They are what M also calls species.

The 🏞️ biomes category covers organelles and other cellular environments, such as mitochondria, the nuclear envelope, and the cytoplasm, as well as auxiliary classes such as 'void' (i.e. non-sample areas, non-vitreous ice, etc.). These models can be used to sample the context within which the species are embedded; or in other words, the biomes are where the species live.

You can always run easymode list to see the most up-to-date list of available models.

Model readiness classification

We classify the readiness of each model for routine use into three categories.

  1. Models marked with 🟢 are available for general use and have been validated in some way via subtomogram averaging.
  2. Models marked with 🔵 are available, and while we personally trust the networks work well, validation by subtomogram averaging is not 100% complete. For example, the npc (nuclear pore complex) model is online and seems accurate, but we aren't happy with the subtomogram averaging resolution we achieved in validation yet (40 Å - mostly a matter of STA processing).
  3. Models marked with 🚧 are a work in progress; some are already online, but validation by STA hasn't progressed very far yet (mostly for a lack of time) and we can't promise great results.