How to Generate Molecule Properties & Visualize the Results

Learn how to use OpenAD in a terminal or Jupyter Notebook to load, organize, visualize and triage your molecule sets.

In this tutorial we'll cover:

1. Collate molecules from a number of SDF, CSV and SMI files
2. Calculate the ESOL property for your molecules ( see note below   )
3. Visually inspect and triage the results
4. Export the top candidates into a new file

Note

This tutorial will focus exclusively on managing and inspecting your molecules. If you want to learn how to use the OpenAD model service to calculate the ESOL property for your molecules, we will post a separate tutorial for that in the future.

Install OpenAD

Find more detailed installation instructions here.

Bash

pip install openad

Note

To run this tutorial from a Jupyter Notebook, make sure to enable OpenAD magic commands and prepend every command with %openad.

Launch OpenAD in the Terminal

Bash

openad

Using Jupyter Notebook

If you plan to use Jupyter notebook, you’ll first have to port our virtual environment to Jupyter.

Bash

python -m ipykernel install --user --name=ad-venv

After selecting the ‘ad-venv’ virtual environment in the Jupyter UI, you can now access OpenAD with the use of “magic commands”.

This means you can run any OpenAD command by prepending it with %openad.

Bash

# Command in the CLI
list files

# Command in Jupyter Notebook
%openad list files

Using your Molecule Working Set

Your molecule working set (MWS) is an in-memory list of small molecules that is the subject of your calculations and manipulations.

Loading Individual Molecules

Let’s start with opening your (empty) MWS.

Bash

show molecules

Now we’ll add a few individual molecules to your MWS using various identifiers. By default, they will be enriched with data from PubChem.

Bash

add mol CC(C)Cc1ccccc1 force
add mol InChI=1S/C7H16O/c1-5-7(4,8)6(2)3/h6,8H,5H2,1-4H3 force
add mol Heptachlor force
add mol '3,3-dimethyl-2-butanol' force

Next you can quickly list your molecules…

Bash

list molecules

…or inspect them in the GUI.

Bash

show molecules

Loading Molecules in Batch

You can load many molecules at once, from different file formats.

Download the sample files below and then copy them into your workspace. Update the file path if needed.

• sample_molecules1.sdf
• sample_molecules2.csv
• sample_molecules3.smi

Bash

import from '~/Downloads/sample_molecules1.sdf' to 'sample_molecules1.sdf'
import from '~/Downloads/sample_molecules2.csv' to 'sample_molecules2.csv'
import from '~/Downloads/sample_molecules3.smi' to 'sample_molecules3.smi'

Once the files are in your workspace, you can easily load them into your MWS.

Bash

load mols from file 'sample_molecules1.sdf'
load mols from file 'sample_molecules2.csv' append
load mols from file 'sample_molecules3.smi' append

Now when you inspect your MWS, you’ll see you have 34 molecules loaded.

Bash

show molecules

Manipulating Your Molecules

This is the stage where you’d use OpenAD or any other external tool to calculate certain properties or perform any other kind of manipulation.

If you’re using a Jupyter Notebook, we can load some example data into your MWS using a dataframe. Let’s assume we have calculated the estimated solubility of our molecules and we have stored the result into a CSV file.

Download the sample file below. Please note that this is random data, not actual ESOL values.

sample_molecules_esol.csv

Now let’s turn this CSV into a dataframe. Paste the code below into a Jupyter Notebook cell and update the file path if necessary.

Python

import pandas as pd
file_path = os.path.join('~/Downloads', 'sample_molecules_esol.csv')
esol_df = pd.read_csv(file_path)
esol_df

We can now add the data from this dataframe to your list of molecules.

Please note that your dataframe should have lowercase columns "smiles", "property" and "result" for this command to work.

Python

%openad merge mols data from dataframe esol_df

If you now inspect any individual molecules that’s present in your MWS, you’ll see it will have the ESOL property listed.

Python

%openad show mol CC(C)Cc1ccccc1

Performing Basic Triage

Now we have the ESOL value for each molecule, we can use the GUI to sort our list by ESOL value and trim it down to the 5 best performing molecules.

Bash

open molecules

1. First turn on the ESOL property in the “Identifiers & Properties” box on top.
2. Sort your molecules by their ESOL value.
3. Select the top-five molecules by clicking on the first one and shift-clicking on the last one.
4. Choose Actions > keep selected to delete all other molecules.
5. Choose Actions > save as and choose the format you want to save them as.

For further processing within OpenAD, we recommend storing your molecules into our own molset.json format. For export to other applications, you can choose between CSV, SDF or SMI.