Deep learning tools and models for MALDI-TOF mass spectra analysis.
Package features:
- Reading and preprocessing functions for MALDI-TOF MS spectra.
- Model definitions to process SMILES strings with state-of-the-art techniques (for feature-based AMR prediction).
- Model definitions to pre-train state-of-the-art Transformer networks on MALDI-TOF MS data
- Model definitions and scripts to train AMR models on the DRIAMS database.
- Model definitions and scripts to train species identification models.
maldi-nn
is distributed on PyPI.
pip install maldi-nn
You may need to install PyTorch before running this command in order to ensure the right CUDA kernels for your system are installed
This package contains all code and scripts to reproduce: "An antimicrobial drug recommender system using MALDI-TOF MS and dual-branch neural networks", and "Pre-trained Maldi Transformers improve MALDI-TOF MS-based prediction". All information regarding reproducing our results can be found in the reproduce folder README
- Implementations of many MALDI reading and processing functions were based on the R package MaldiQuant.
- Topological Peak Filtering was taken from the Topf package.
Antimicrobial drug recommenders:
@article{dewaele2023antimicrobial,
title={An antimicrobial drug recommender system using MALDI-TOF MS and dual-branch neural networks},
author={De Waele, Gaetan and Menschaert, Gerben and Waegeman, Willem},
journal={bioRxiv},
pages={2023--09},
year={2023},
publisher={Cold Spring Harbor Laboratory}
}
Maldi Transformers:
@article{dewaele2024pre,
title={Pre-trained Maldi Transformers improve MALDI-TOF MS-based prediction},
author={De Waele, Gaetan and Menschaert, Gerben and Vandamme, Peter and Waegeman, Willem},
journal={bioRxiv},
pages={2024--01},
year={2024},
publisher={Cold Spring Harbor Laboratory}
}