UniCure: Multi-modal Datasets and Weights for Personalized Cancer Therapy Prediction

UniCure is a multi-modal framework integrating omics and chemical foundation models to predict transcriptomic drug responses. This repository contains the pre-processed datasets, configuration files, and pre-trained model weights required to reproduce the results. The archive is 12.4 GB and was last updated on 2026-04-23 by Zexi Chen.

Use Cases

• Predicting transcriptomic drug responses based on integrated omics and chemical data.
• Reproducing model training and evaluation results for the UniCure framework.
• Fine-tuning pre-trained multi-modal models on new cancer therapy datasets.
• Benchmarking new drug response prediction methods against the provided LINCS L1000, SciPlex, and PTC datasets.

Strengths

• Includes pre-trained model weights (e.g., `Unicure_best_model.pth`) for immediate inference.
• Contains pre-processed datasets from multiple sources (LINCS L1000, SciPlex, PTC).
• Archive size is 12.4 GB, indicating a substantial collection of data and model artifacts.
• Dataset splits (Training, Validation, Test) are provided across different random seeds.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data may reflect the specific biases inherent to the source datasets (LINCS L1000, SciPlex, PTC).

Provenance

Source

Associated publication: Chen et al., 'UniCure: A Multi-modal Model for Predicting Personalized Cancer Therapy Response', Cancer Cell (2026).

Collection Method

Data likely pre-processed from public sources including LINCS L1000, SciPlex, and PTC datasets.

Freshness

Last updated 2026-04-23 05:02:46; freshness should be verified.