Train graph neural networks on molecular and biomedical datasets for drug discovery.
npx skills add https://github.com/rubensliv/k-dense-ai --skill torchdrug-rubensliv
Please help me install this Agent Skill. Skill: torchdrug Source: https://github.com/rubensliv/k-dense-ai/tree/main/scientific-skills/torchdrug Command: npx skills add https://github.com/rubensliv/k-dense-ai --skill torchdrug-rubensliv
TorchDrug enables researchers to build graph-based deep learning workflows for molecules, proteins, and biomedical graphs, accelerating discovery by turning complex biological data into actionable predictions.
Train a GIN model on BBBP to predict blood-brain barrier penetration.
High-intent search queries and answers about installing and using this skill.
Graph neural networks for molecular property prediction are enabled by training configurable models like GIN on molecular datasets to predict properties such as blood-brain barrier penetration. You can leverage rich dataset pipelines and training workflows tailored for chemistry.
Yes, protein modeling workflows support seamless integration with RDKit, AlphaFold/ESM, and PyTorch Lightning. This streamlines experiments by combining graph-based deep learning with established structural biology and machine learning frameworks.
Building a dataset pipeline for biomedical knowledge graph reasoning utilizes configurable datasets and tasks tailored for biology. The system supports YAML frontmatter for discovery and optional references directories to structure the workflow efficiently.
Graph-based deep learning applies directly to molecular generation tasks within chemistry workflows. It supports configurable models and training workflows to generate molecular structures based on biological data inputs.
PyTorch experience is helpful since the system integrates with PyTorch Lightning to streamline training workflows. However, configurable models and YAML-based setups lower the entry barrier for predicting blood-brain barrier penetration.
Graph neural networks excel at drug discovery because they directly model molecular, protein, and biomedical graph structures. This captures complex biological relationships that standard deep learning approaches struggle to represent.
sanand0Predict molecular properties with graph neural networks using PyTorch.
jimmc414Predict molecular properties from SMILES strings using a PyTorch-based toolkit.
LeonChaoXTrain PyTorch-native graph neural networks on TorchDrug datasets for molecular and protein tasks.
Yezez9Develop custom GNN architectures for molecular and protein data analysis with PyTorch.
