treeArches for learning cell-type hierarchies
treeArches automatically identifies novel disease states
I am excited to share our new work, we tackled: 1) how to learn a harmonized cell-type hierarchy/taxonomy across many studies with different annotations, 2) how to automatically identify novel cell states (e.g. disease) when mapping new query data. We use transfer learning and hierarchical classifiers to build and update the hierarchies when new datasets are added. If a population is identified to match reference populations, it will be added to the same level o.w either as a subpopulation or a new state.
See twitter thread here: https://twitter.com/Mohlotf/status/1545328933247373312
Link to the paper: https://www.biorxiv.org/content/10.1101/2022.07.07.499109v1
Link to the code and tutorials: https://scarches.readthedocs.io/en/latest/treeArches_pbmc.html
See advance guide how to use human lung cel atlas to identify disease states: https://scarches.readthedocs.io/en/latest/treeArches_identifying_new_ct.html