Published May 7, 2025
| Version 1.0
Software
Restricted
XPert: A Knowledge-Informed Dual-Branch Transformer Model for Predicting Drug-Induced Cellular Perturbation Effects
Authors/Creators
-
1.
Zhejiang University
Description
Introduction
Systematic mapping of chemical perturbation responses is revolutionizing polypharmacological drug discovery, yet remains constrained by experimental scalability. Here, we introduce XPert, a biologically informed Dual-Branch Transformer model that predicts gene-specific drug responses across dose-time conditions, outperforming VAE-based methods. It generalizes to unseen cells, transfers knowledge to clinical settings, and reveals mechanistic insights, offering a scalable solution for precision medicine and perturbation-based drug discovery.
Overview
The repository is organised as follows:
- `processed_data/` contains processed data files;
- `HG_data/` contains the data for training the knowledge heterogeneous graph and the pre-trained embeddings;
- `dataset/` contains the necessary files for creating the dataset;
- `models/` contains different modules of XPert;
- `configs/` contains all the config files adapted for different datasets and scenarios;
- `experiment/` contains log files and output files;
- `scripts/` contains the scripts for training, inference, and testing the model;
- `saved_model/` contains the trained and pretrained weights;
- `evaluation_metrics/` contains all the evaluation metrics mentioned in the paper;
- `reproducing/` contains the code for reproducing the analysis results and figures from the paper;
Files
Restricted
The record is publicly accessible, but files are restricted. <a href="https://zenodo.org/account/settings/login?next=https://zenodo.org/records/15357712">Log in</a> to check if you have access.
Additional details
Software
- Development Status
- Active