The notebook folder contains old, not supported examples. We should replace them with some explanatory examples on the line of the protocol paper.
- Add a new non-competing dataset (i.e. the data is not directly related to any existing CC space, e.g. M1.001)
- Add a new competing dataset (i.e. the raw data input is the same as other dataset difference is in the processing, e.g. D1.002)
- Property prediction: supervised learning inspired by MoleculeNet TDIcommons or other
- 2D projections of small compound collections, Neighbor identification, enrichment analysis?
- Signature reversion of Alzheimer's-related traits?
- Signature mimicking of bio-drugs?