protein-qc

Automate protein design quality-control checks with standardized thresholds and composite scoring.

151|20|Updated Jan 15, 2026

Search Tags:#metrics #pLDDT #quality-control #thresholds #protein-design #ipTM #composite-scoring

Ready to run

npx skills add https://github.com/adaptyvbio/protein-design-skills --skill protein-qc

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: protein-qc
Source: https://github.com/adaptyvbio/protein-design-skills/tree/main/skills/protein-qc
Command: npx skills add https://github.com/adaptyvbio/protein-design-skills --skill protein-qc

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill automates quality-control checks for protein design campaigns, streamlining evaluation by applying standardized thresholds to ensure design viability and reduce failure risk.

Core Features & Use Cases

Tiered filtering to prune designs by structural confidence, interface quality, and sequence liabilities.
Composite ranking by combining multiple metrics (pLDDT, ipTM, PAE, SC, esm2_pll) into a single score.
Risk assessment with severity scoring to identify actionable liabilities and guide redesign.

Quick Start

Upload your designs CSV and run the QC pipeline to filter by pLDDT > 0.85, ipTM > 0.50, scRMSD < 2.0, then compute the composite score and output top candidates for experimental testing.

Frequently Asked Questions about protein-qc

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I automate protein design quality-control checks for high-throughput campaigns?▼

Protein design quality-control can be automated by applying standardized thresholds to evaluate design viability, filtering by structural confidence, interface quality, and sequence liabilities. This Skill processes your designs CSV to prune unsuitable candidates and reduce experimental failure risk.

What pLDDT and ipTM thresholds should I use to filter protein designs?▼

Standard protein design filtering thresholds include pLDDT > 0.85, ipTM > 0.50, and scRMSD < 2.0. These hard thresholds evaluate structural confidence and interface quality to ensure design viability before advancing candidates.

How do I compute a composite score for ranking protein designs?▼

You can compute a composite score for protein designs by combining multiple metrics such as pLDDT, ipTM, PAE, SC, and esm2_pll into a single ranking score. This composite scoring produces prioritized design sets for experimental testing.

Can I identify sequence liabilities in my protein designs automatically?▼

Yes, automated quality-control checks include risk assessment with severity scoring to identify actionable sequence liabilities across binding, expression, and structural assessments. This guides targeted redesign efforts for problematic candidates.

What is the best way to rank protein designs for experimental testing?▼

The best way to rank protein designs is tiered filtering followed by composite scoring. By applying hard thresholds for structural confidence and interface quality, then combining multiple metrics, you output top candidates optimized for experimental testing.

Does this protein QC pipeline work with standard CSV design outputs?▼

Yes, the protein QC pipeline accepts standard design CSV uploads to evaluate viability. It applies standardized thresholds across structural, binding, and expression assessments to support tiered filtering in high-throughput campaigns.