Syndirella 🤝 HIPPO

Setup

1. Fragalysis download of crystallographic data
2. Create blank hippo DB
3. Load aligned hits into DB
4. Register an extra pose for Ax0310a (Matteo's relaxed) (store pose_ID somewhere global)

Base compounds

1. Get inspiration 1 inspiration1 = animal.poses[...]

2. Get inspiration 2 inspiration2 = animal.poses[...]

3. Register the base compound

   1. tags = ['base']
   2. metadata = dict(warren_name=..., )
   3. base = animal.register_compound(smiles=smiles, tags=tags, metadata=metadata)

4. Place the compound with Fragmenstein

5. Register a pose

   1. tags = ['base', 'flat_placement']
   2. metadata = dict(ddG=..., mRMSD=...)
   3. pose = animal.register_pose(smiles=smiles, compound=base, tags=tags, metadata=metadata, inspirations=[inspiration1, inspiration2])

Elaborations & placement pipeline (parralel jobs)

1. same as before but stop at num_atom_diff == 15

2. Write a new CSV output with columns:

   • compound_ID (from hippo DB)
   • paths to minimised .mol (only if successful?)
   • reactant smiles
   • reaction names
   • ddG
   • RMSD
   • acceptable?
   • errors?

base	reaction 1	...	...	...	reaction 2	...	...	...	...	reaction 3	...	...	...	...	placement
	name	reactant1	reactant2	product	name	reactant1	reactant2	product	r_previous_product	name	reactant1	reactant2	product	r_previous_product	ddG	RMSD	acceptable	errors
smiles	Amidation	smiles	smiles	smiles	BOC deprotect	smiles	smiles	smiles	int	BOC deprotect	smiles	smiles	smiles	int			True/False	"Something went wrong"

Load in poses (single thread / job)

1. Parse through the new placement CSV's
2. Register poses if ddG, RMSD, acceptable, errors columns are all OK
3. For each registered pose add the following detail:
   • tags = ['base'] or ['elab']
   • metadata = dict(ddG=..., RMSD=...)