Getting started with HIPPO

To create a HIPPO database or interface with an existing one, create a HIPPO ‘animal’ object:

from hippo import HIPPO
animal = HIPPO(project_name, "path/to/db.sqlite")

Loading crystallographic hits from Fragalysis

1. Go Fragalysis and download a dataset. For public datasets you can use fragalysis.download_target()

Load the crystallographic data

animal.add_hits(
        target_name='A71EV2A',
        metadata_csv='/path/to/metadata.csv',
        aligned_directory='/path/to/aligned_files',
)

Attention

N.B. all poses loaded into a HIPPO database only have an absolute path stored to the original file - they are not copied! It is your responsibility to ensure that their original files remain accessible.

Navigating compounds and poses

The below sections explain how to work with Compound objects and sets thereof. For further details on the concept of Compound objects and others see Definitions and units.

Getting compounds/poses

Compounds can be accessed via the compounds property which wraps a CompoundTable:

all_compounds = animal.compounds

CompoundTables can be indexed using their database id (positive integer) or InChiKey (str):

c1 = animal.compounds[1]
c2 = animal.compounds["FHZALEJIENDROK-UHFFFAOYSA-N"]

You can select a subset of compounds using slices, tuples, or lists:

subset1 = animal.compounds[20:30]
subset2 = animal.compounds[1,2,3,4]
subset3 = animal.compounds[[1,2,3,4,5]]

Additionally you can get compounds by their tag:

hits = animal.compounds(tag='hits')

Equivalent methods exist for animal.poses (returns a PoseTable), animal.reactions (returns a ReactionTable), animal.interactions (returns a InteractionTable), and animal.tags returns a TagTable). See also the API Reference pages.

Inspecting a compound and its poses

Once you have a compound you can access database properties using its properties:

c = animal.compounds[1]

c.id # Database ID (int)
c.name # InChiKey
c.smiles # (flattened) smiles
c.mol # rdkit.Chem.Mol
c.tags # assigned tags
c.metadata # metadata dictionary

c.draw() # draw the molecule (and its base)

You can access a compounds poses, which have similar functionality

poses = c.poses

p = poses[0]

p.id # Database ID (int)
p.name # pose name
p.smiles # (stereo) smiles
p.mol # rdkit.Chem.Mol
p.tags # assigned tags
p.metadata # metadata dictionary

c.draw() # draw the molecule pose (3d)

See also the API reference for compounds and poses.

Graphing

Several convenient graphing methods are available. Try:

animal.plot_tag_statistics()
animal.plot_pose_property('CanonSites')

Interaction fingerprinting

Interactions fingerprinted as follows:

import mrich

for pose in mrich.track(animal.poses):
        pose.calculate_interactions()

N.B. mrich.track just gives you a nice progress bar

Interaction fingerprints can be visualised with a ‘punchcard’, per-residue histogram, or viewed individually for a Pose.

Interaction Punchcard

animal.plot_interaction_punchcard(poses=animal.poses(tag='hits'), subtitle='hits', group='pose_name')

Interactions by residue

animal.plot_residue_interactions(poses=animal.poses(tag='hits'), residue_number=123, chain='A', subtitle='hits')

Interactions of a single Pose

This will create an HTML file you can open in your browser:

import molparse as mp
pose = animal.poses[1]
fig = animal.plot_pose_interactions(pose=pose)
mp.write(f'{pose}_interactions.html', fig)

This method of writing to an HTML file works for all the above figures.