Design protein backbones and sequences conditioned on ligands, DNA, or RNA.
npx skills add https://github.com/LiorZ/protein-design-skills --skill disco-liorz
Please help me install this Agent Skill. Skill: disco Source: https://github.com/LiorZ/protein-design-skills/tree/main/skills/disco Command: npx skills add https://github.com/LiorZ/protein-design-skills --skill disco-liorz
DISCO enables end-to-end co-design of a protein's sequence and backbone structure, allowing conditioning on non-protein entities such as ligands, DNA, or RNA to streamline design workflows.
Run the DISCO runner with a ligand-conditioned input JSON to generate de novo backbone designs and sequences.
High-intent search queries and answers about installing and using this skill.
DISCO enables joint sequence-structure co-design by using diffusion-based generation to create protein backbones and sequences conditioned on non-protein partners. It accepts ligand, DNA, or RNA inputs to generate structured PDBs and sequence files.
Yes, DISCO supports ligand, DNA, and RNA conditioning for designing binding proteins. It uses a Hydra-based CLI to process input JSON and generate protein backbones and sequences specifically conditioned on these non-protein entities.
To run large-scale protein design campaigns, use the DISCO Hydra-based CLI with scalable options and explicit overrides. Provide valid ligand-conditioned input JSON to execute large-scale screens and output structured PDB and sequence files.
Yes, valid input JSON is required to run the DISCO CLI for protein backbone and sequence generation. The input schema defines the conditioning parameters, and the CLI outputs structured PDBs and sequence files under a specified dump directory.
Outputs from DISCO co-design runs include structured PDBs and sequence files saved under a dump directory. These outputs are specifically formatted for downstream refolding, quality control, and ranking pipelines to validate the generated protein designs.
Avoid using DISCO if your workflow lacks the DISCO CLI or valid input JSON, as these are strictly required. Additionally, it only supports conditioning on ligands, DNA, and RNA, so it may not apply if your targets require other types of molecular conditioning.
LiorZPredict biomolecular structures and affinities from YAML inputs via CLI diffusion workflow.
InternScienceValidate protein sequences and return per-sequence boolean flags and messages.
robinbarvaagPerform computational molecular biology tasks with Python for sequence analysis and file parsing.
zongtingweiGenerate flow-based protein backbones with fold conditioning for sequence design.
