Retention time prediction

3DMolMS can be used to predict MS/MS-related properties, such as retention time (RT) and collision cross section (CCS). This guide shows how to train a model for RT prediction and CCS prediction, and how to transfer these models to your own RT and CCS dataset.

All models mentioned can be downloaded from release v1.2.0.

Setup

Please set up the environment as shown in the Source code setup page.

Step 1: Data preparation

Download the retention time dataset, METLIN. The structure of data directory is:

|- data
  |- origin
    |- SMRT_dataset.sdf

Step 2: Preprocessing

Use the following commands to preprocess the datasets. The settings of datasets are in ./molnetpack/config/preprocess_etkdgv3.yml.

python scripts/preprocess.py --task rt \
--data_config_path ./molnetpack/config/preprocess_etkdgv3.yml

Step 3: Training

Use the following commands to train the model. The settings of model and training are in ./molnetpack/config/molnet_rt.yml.

Using the command-line script:

Learning from scratch:

python scripts/train.py --task rt \
--train_data ./data/metlin_etkdgv3_train.pkl \
--test_data ./data/metlin_etkdgv3_test.pkl \
--checkpoint_path ./check_point/molnet_rt_etkdgv3.pt

If you’d like to train this model from the pre-trained model on MS/MS prediction, please download the pre-trained model from release v1.2.0.

Learning from MS/MS model:

python scripts/train.py --task rt \
--train_data ./data/metlin_etkdgv3_train.pkl \
--test_data ./data/metlin_etkdgv3_test.pkl \
--checkpoint_path ./check_point/molnet_rt_etkdgv3_tl.pt \
--transfer \
--resume_path ./check_point/molnet_qtof_etkdgv3.pt

Using the Python API:

import torch
from molnetpack import MolNet

device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
molnet_engine = MolNet(device, seed=42)

# Learning from scratch:
molnet_engine.train(
    task='rt',
    train_data='./data/metlin_etkdgv3_train.pkl',
    valid_data='./data/metlin_etkdgv3_test.pkl',
    checkpoint_path='./check_point/molnet_rt_etkdgv3.pt',
)

# Learning from MS/MS model (transfer learning):
molnet_engine.train(
    task='rt',
    train_data='./data/metlin_etkdgv3_train.pkl',
    valid_data='./data/metlin_etkdgv3_test.pkl',
    checkpoint_path='./check_point/molnet_rt_etkdgv3_tl.pt',
    resume_path='./check_point/molnet_qtof_etkdgv3.pt',
    transfer=True,
    use_scaler=True,
)