Predicting

Contents

Predicting#

[1]:

# Install chemprop from GitHub if running in Google Colab
import os

if os.getenv("COLAB_RELEASE_TAG"):
    try:
        import chemprop
    except ImportError:
        !git clone https://github.com/chemprop/chemprop.git
        %cd chemprop
        !pip install .
        %cd examples

Import packages#

[2]:

import pandas as pd
import numpy as np
import torch
from lightning import pytorch as pl
from pathlib import Path

from chemprop import data, featurizers, models

Change model input here#

[3]:

chemprop_dir = Path.cwd().parent
checkpoint_path = chemprop_dir / "tests" / "data" / "example_model_v2_regression_mol.ckpt" # path to the checkpoint file.
# If the checkpoint file is generated using the training notebook, it will be in the `checkpoints` folder with name similar to `checkpoints/epoch=19-step=180.ckpt`.

Load model#

[4]:

mpnn = models.MPNN.load_from_checkpoint(checkpoint_path)
mpnn

[4]:

MPNN(
  (message_passing): BondMessagePassing(
    (W_i): Linear(in_features=86, out_features=300, bias=False)
    (W_h): Linear(in_features=300, out_features=300, bias=False)
    (W_o): Linear(in_features=372, out_features=300, bias=True)
    (dropout): Dropout(p=0.0, inplace=False)
    (tau): ReLU()
    (V_d_transform): Identity()
    (graph_transform): GraphTransform(
      (V_transform): Identity()
      (E_transform): Identity()
    )
  )
  (agg): MeanAggregation()
  (bn): BatchNorm1d(300, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
  (predictor): RegressionFFN(
    (ffn): MLP(
      (0): Sequential(
        (0): Linear(in_features=300, out_features=300, bias=True)
      )
      (1): Sequential(
        (0): ReLU()
        (1): Dropout(p=0.0, inplace=False)
        (2): Linear(in_features=300, out_features=1, bias=True)
      )
    )
    (criterion): MSE(task_weights=[[1.0]])
    (output_transform): UnscaleTransform()
  )
  (X_d_transform): Identity()
  (metrics): ModuleList(
    (0-1): 2 x MSE(task_weights=[[1.0]])
  )
)

Change predict input here#

[5]:

chemprop_dir = Path.cwd().parent
test_path = chemprop_dir / "tests" / "data" / "regression" / "mol" / "mol.csv"
smiles_column = 'smiles'

Load test smiles#

[6]:

df_test = pd.read_csv(test_path)
df_test

[6]:

	smiles	lipo
0	Cn1c(CN2CCN(CC2)c3ccc(Cl)cc3)nc4ccccc14	3.54
1	COc1cc(OC)c(cc1NC(=O)CSCC(=O)O)S(=O)(=O)N2C(C)...	-1.18
2	COC(=O)[C@@H](N1CCc2sccc2C1)c3ccccc3Cl	3.69
3	OC[C@H](O)CN1C(=O)C(Cc2ccccc12)NC(=O)c3cc4cc(C...	3.37
4	Cc1cccc(C[C@H](NC(=O)c2cc(nn2C)C(C)(C)C)C(=O)N...	3.10
...	...	...
95	CC(C)N(CCCNC(=O)Nc1ccc(cc1)C(C)(C)C)C[C@H]2O[C...	2.20
96	CCN(CC)CCCCNc1ncc2CN(C(=O)N(Cc3cccc(NC(=O)C=C)...	2.04
97	CCSc1c(Cc2ccccc2C(F)(F)F)sc3N(CC(C)C)C(=O)N(C)...	4.49
98	COc1ccc(Cc2c(N)n[nH]c2N)cc1	0.20
99	CCN(CCN(C)C)S(=O)(=O)c1ccc(cc1)c2cnc(N)c(n2)C(...	2.00

100 rows × 2 columns

Get smiles#

[7]:

smis = df_test[smiles_column]
smis

[7]:

0               Cn1c(CN2CCN(CC2)c3ccc(Cl)cc3)nc4ccccc14
1     COc1cc(OC)c(cc1NC(=O)CSCC(=O)O)S(=O)(=O)N2C(C)...
2                COC(=O)[C@@H](N1CCc2sccc2C1)c3ccccc3Cl
3     OC[C@H](O)CN1C(=O)C(Cc2ccccc12)NC(=O)c3cc4cc(C...
4     Cc1cccc(C[C@H](NC(=O)c2cc(nn2C)C(C)(C)C)C(=O)N...
                            ...
95    CC(C)N(CCCNC(=O)Nc1ccc(cc1)C(C)(C)C)C[C@H]2O[C...
96    CCN(CC)CCCCNc1ncc2CN(C(=O)N(Cc3cccc(NC(=O)C=C)...
97    CCSc1c(Cc2ccccc2C(F)(F)F)sc3N(CC(C)C)C(=O)N(C)...
98                          COc1ccc(Cc2c(N)n[nH]c2N)cc1
99    CCN(CCN(C)C)S(=O)(=O)c1ccc(cc1)c2cnc(N)c(n2)C(...
Name: smiles, Length: 100, dtype: object

Get molecule datapoints#

[8]:

test_data = [data.MoleculeDatapoint.from_smi(smi) for smi in smis]

Get molecule dataset#

[9]:

featurizer = featurizers.SimpleMoleculeMolGraphFeaturizer()
test_dset = data.MoleculeDataset(test_data, featurizer=featurizer)
test_loader = data.build_dataloader(test_dset, shuffle=False)

Set up trainer#

[10]:

with torch.inference_mode():
    trainer = pl.Trainer(
        logger=None,
        enable_progress_bar=True,
        accelerator="cpu",
        devices=1
    )
    test_preds = trainer.predict(mpnn, test_loader)

💡 Tip: For seamless cloud uploads and versioning, try installing [litmodels](https://pypi.org/project/litmodels/) to enable LitModelCheckpoint, which syncs automatically with the Lightning model registry.
GPU available: False, used: False
TPU available: False, using: 0 TPU cores

/home/knathan/anaconda3/envs/chemprop/lib/python3.11/site-packages/lightning/pytorch/trainer/connectors/data_connector.py:434: The 'predict_dataloader' does not have many workers which may be a bottleneck. Consider increasing the value of the `num_workers` argument` to `num_workers=11` in the `DataLoader` to improve performance.

[11]:

test_preds = np.concatenate(test_preds, axis=0)
df_test['pred'] = test_preds
df_test

[11]:

	smiles	lipo	pred
0	Cn1c(CN2CCN(CC2)c3ccc(Cl)cc3)nc4ccccc14	3.54	2.253542
1	COc1cc(OC)c(cc1NC(=O)CSCC(=O)O)S(=O)(=O)N2C(C)...	-1.18	2.235016
2	COC(=O)[C@@H](N1CCc2sccc2C1)c3ccccc3Cl	3.69	2.245891
3	OC[C@H](O)CN1C(=O)C(Cc2ccccc12)NC(=O)c3cc4cc(C...	3.37	2.249847
4	Cc1cccc(C[C@H](NC(=O)c2cc(nn2C)C(C)(C)C)C(=O)N...	3.10	2.228097
...	...	...	...
95	CC(C)N(CCCNC(=O)Nc1ccc(cc1)C(C)(C)C)C[C@H]2O[C...	2.20	2.233408
96	CCN(CC)CCCCNc1ncc2CN(C(=O)N(Cc3cccc(NC(=O)C=C)...	2.04	2.236931
97	CCSc1c(Cc2ccccc2C(F)(F)F)sc3N(CC(C)C)C(=O)N(C)...	4.49	2.237789
98	COc1ccc(Cc2c(N)n[nH]c2N)cc1	0.20	2.252625
99	CCN(CCN(C)C)S(=O)(=O)c1ccc(cc1)c2cnc(N)c(n2)C(...	2.00	2.235702

100 rows × 3 columns