protein-design-skills

🥈Silver

Claude Code skills for designing and engineering proteins. Researchers and biotech teams use these skills to automate protein design tasks, optimizing sequences for stability, function, and therapeutic applications. Integrates with molecular modeling tools and lab workflows.

12570Updated 2w ago

Intermediate30min to implementautomation

Saves ~360 min per use

Quick InstallView Source

git clone https://github.com/adaptyvbio/protein-design-skills.git

Works with:

Claude

Overview

About This Skill

How to Use

["1. **Define Your Requirements**: Fill in the [PLACEHOLDERS] in the prompt template with your specific goals (e.g., target protein, stability criteria, solubility requirements). Use tools like [UniProt](https://www.uniprot.org/) or [PDB](https://www.rcsb.org/) to gather structural/sequence data.","2. **Choose Your Tools**: Select the molecular modeling tools (e.g., [Rosetta](https://www.rosettacommons.org/), [AlphaFold2](https://deepmind.com/research/case-studies/alphafold), [CamSol](http://www-mvsoftware.ch.cam.ac.uk/index.php/camsol)) and design algorithms (e.g., RosettaDesign, ProteinMPNN) for your workflow. Ensure they are compatible with your computational resources.","3. **Run the Design Pipeline**: Execute the prompt in your AI tool (e.g., Claude Code) to generate candidate sequences. Adjust parameters like [DESIGN_ALGORITHM] or [SCORING_METRIC] based on your priorities (e.g., speed vs. accuracy).","4. **Validate and Iterate**: Use the output to rank candidates and generate reports. Cross-validate predictions with experimental data (e.g., DSF for Tm, SPR for affinity) and refine the design iteratively. Tools like [PyMOL](https://pymol.org/2/) or [ChimeraX](https://www.cgl.ucsf.edu/chimerax/) can help visualize structures.","5. **Integrate with Workflows**: Automate repetitive tasks (e.g., sequence generation, scoring) using scripts or workflow managers like [Snakemake](https://snakemake.readthedocs.io/) or [Nextflow](https://www.nextflow.io/). Share results with collaborators via platforms like [Benchling](https://benchling.com/) or [DNAnexus](https://www.dnanexus.com/)."]

Setup & Installation

Quick Install

No install command available. Check the GitHub repository for manual installation instructions.

Alternative Install (Git Clone)

git clone https://github.com/adaptyvbio/protein-design-skills

Requirements

Claude Code or compatible AI agent
Works with: Claude

Quick Start Guide

Install the Skill

Copy the install command above and run it in your terminal.

Open Your AI Agent

Launch Claude Code, Cursor, or your preferred AI coding agent.

Try It Out

Use the prompt template or examples below to test the skill.

Customize

Adapt the skill to your specific use case and workflow.

Usage Examples

Prompt Template

Design a [FUNCTION] protein with [STABILITY_CRITERIA] and [SOLUBILITY_REQUIREMENTS]. Start by analyzing the [TARGET_SEQUENCE] or [STRUCTURAL_TEMPLATE] using [MOLECULAR_MODELING_TOOL]. Generate 10-20 candidate sequences with [DESIGN_ALGORITHM], then rank them by [SCORING_METRIC]. For the top 3 candidates, provide a detailed report including: (1) predicted structure, (2) stability metrics (ΔG, melting temperature), (3) potential binding sites, and (4) experimental validation suggestions. Use [SPECIFIC_PARAMETERS] for optimization. [ADDITIONAL_CONSTRAINTS].

Example Output

### Protein Design Report: Engineered Antibody Fragment for SARS-CoV-2 Spike Protein

**Objective**: Design a high-affinity, thermostable single-domain antibody (VHH) targeting the SARS-CoV-2 receptor-binding domain (RBD) with a melting temperature (Tm) > 75°C and solubility > 50 mg/mL in PBS.

**Design Process**:
1. **Template Selection**: Used the crystal structure of PDB ID 6W41 (SARS-CoV-2 RBD) as a template. The initial VHH scaffold was derived from Lama glama (PDB: 5OM3), which has a Tm of 68°C and moderate solubility.
2. **Sequence Optimization**: Applied RosettaDesign with the following constraints:
   - **Stability**: Enforced disulfide bonds (Cys22-Cys92) and proline substitutions (Pro44, Pro85) to stabilize the CDR loops.
   - **Affinity**: Focused mutations on CDR1 (residues 27-32) and CDR3 (residues 99-110) to target the RBD epitope (K417, N440, Y453).
   - **Solubility**: Introduced surface-exposed polar residues (e.g., Asn28, Asp31) and removed hydrophobic patches.
3. **Candidate Generation**: Generated 15 sequences using Rosetta’s FastDesign protocol. Top candidates were scored using:
   - **Rosetta ddG**: Predicted ΔΔG for stability (target: < -2.0 REU).
   - **Solubility Score**: Calculated via CamSol (target: > 0.5).
   - **Binding Energy**: Estimated via Rosetta FlexPepDock (target: < -50 REU).

**Top 3 Candidates**:

| Candidate | Sequence (CDR1/CDR3) | ddG (kcal/mol) | Tm (°C) | Solubility (mg/mL) | Binding Affinity (nM) | Key Mutations |
|-----------|----------------------|----------------|---------|--------------------|-----------------------|---------------|
| VHH-001   | **GYTFTNYW**/**GQGTTLTVSS** | -2.8 | 79.2 | 62 | 12.5 | S31N, Y49F, T100S |
| VHH-002   | **GYTFTNYW**/**GQGTTLTVRS** | -2.5 | 77.8 | 58 | 8.3 | S31N, Y49F, T100R |
| VHH-003   | **GYTFTNYW**/**GQGTTLTVGS** | -2.3 | 76.5 | 55 | 15.2 | S31N, Y49F, T100G |

**Structural Validation**:
- **VHH-001** was modeled with AlphaFold2, showing a well-defined CDR3 loop forming a β-hairpin that docks into the RBD groove (see attached PDB file). The disulfide bond (Cys22-Cys92) stabilizes the framework, and the mutations S31N and T100S enhance solubility without disrupting the binding interface.
- **Binding Mode**: The CDR3 loop interacts with residues K417, N440, and Y453 via hydrogen bonds (S31N to K417) and hydrophobic contacts (F49 with Y453).

**Experimental Validation Plan**:
1. **Cloning**: Synthesize codon-optimized genes for VHH-001, VHH-002, and VHH-003 with a C-terminal His6 tag for purification.
2. **Expression**: Test expression in *E. coli* BL21(DE3) and *Pichia pastoris* (for glycosylation compatibility).
3. **Purification**: Use Ni-NTA affinity chromatography followed by size-exclusion chromatography (SEC) to assess monodispersity.
4. **Thermal Shift Assay**: Measure Tm via differential scanning fluorimetry (DSF) using SYPRO Orange.
5. **Binding Kinetics**: Perform surface plasmon resonance (SPR) with immobilized RBD to confirm affinity.
6. **Functional Assay**: Test neutralization of pseudotyped SARS-CoV-2 virus in Vero E6 cells.

**Next Steps**:
- Prioritize VHH-001 for further optimization (e.g., CDR3 loop truncation to reduce flexibility).
- Explore disulfide shuffling in CDR1 to improve affinity without compromising stability.
- Collaborate with a structural biology team to resolve the crystal structure of VHH-001 bound to RBD.

**Attachments**:
- PDB files for VHH-001 (modeled structure) and RBD (PDB: 6W41).
- Rosetta score files for all 15 candidates.
- CamSol solubility predictions.

**Recommendation**: Proceed with VHH-001 for scale-up and preclinical testing due to its optimal balance of stability, solubility, and affinity.