Biotechnology · Intermediate

A Computational Library Design Protocol for Rapid Improvement of Protein Stability: FRESCO

Written by SignalsLab

Early edition: these lessons are being reviewed.

Content

Summary

FRESCO is a computational library design protocol that rapidly improves protein thermostability by 20–35 °C with minimal experimental screening, typically requiring fewer than 200 variants. It identifies stabilizing point mutations and disulfide bonds through in silico screening using FoldX, Rosetta, and YASARA, generating a high-quality mutant library for targeted experimental verification.

Intuition

Traditional protein engineering often yields modest stability gains of only 2–15 °C and requires extensive screening. FRESCO overcomes this by using computational tools to predict stabilizing mutations in both flexible and rigid regions of a protein. This approach is particularly critical for large proteins, which tend to aggregate irreversibly after partial unfolding, unlike small proteins that unfold reversibly. By prescreening mutations in silico, FRESCO creates a 'smart library' where most variants are predicted to be stable, allowing researchers to verify improvements quickly with fewer than 200 experimental variants.

Derivation

  1. Identify the limitation of traditional methods which yield only 2–15 °C stability gains.
  2. Define the target improvement range for FRESCO as 20–35 °C.
  3. Distinguish between small proteins (reversible, single-step unfolding) and large proteins (irreversible aggregation after partial unfolding).
  4. Select computational tools: FoldX and Rosetta for point mutations, YASARA for disulfide bonds and MD simulations.
  5. Scan the entire protein structure to find mutations in both flexible and rigid (low B-factor) regions.
  6. Optimize speed by inspecting each mutant only once, rather than filtering out chemically unreasonable mutations before MD.
  7. Generate a high-quality library with >10% chance of enhanced stability.
  8. Perform experimental verification on fewer than 200 variants.

Worked example

Problem

A researcher wants to stabilize limonene epoxide hydrolase (LEH), a large protein that aggregates after partial unfolding. They need to improve its thermostability significantly without screening thousands of variants.

  1. Run FRESCO protocol using FoldX and Rosetta to predict stabilizing point mutations.
  2. Use YASARA to design disulfide bonds and perform molecular dynamics simulations.
  3. Scan the entire protein structure, including rigid regions, to identify stabilizing spots.
  4. Generate a mutant library based on these predictions.
  5. Experimentally screen fewer than 200 variants to verify stability improvements.

Answer The researcher successfully stabilizes LEH by achieving a thermostability increase of 20–35 °C through the targeted screening of a computationally designed library, avoiding the need for extensive high-throughput screening.

Common mistakes

  • Confusing the unfolding behavior of small proteins (reversible, single-step) with large proteins (irreversible aggregation after partial unfolding).
  • Assuming that mutations outside early unfolding regions in large proteins have a significant effect on stability, when they may be negligible.
  • Believing that FRESCO requires extensive high-throughput screening rounds, when it is designed to minimize this need by prescreening in silico.
  • Thinking that chemically unreasonable mutations are filtered out before MD simulations in the current protocol, whereas they are now inspected once for speed.

Animation

Animation coming soon

The animated version of this lesson is being prepared for the web. In the Signals AI app, lessons are drawn step by step.

Code

This is a text lesson: it has no animation code yet.

ccBYNC© 2026 SignalsLab. Lesson text shared as is under CC BY-NC 4.0.