CompanyKAISTPartnershipResearchWet Lab

Foldexa Partners with KAIST Bioengineering Lab

We are proud to announce a formal research partnership with KAIST's Department of Biological Sciences to validate Foldexa's computational pipelines against experimental wet-lab results — closing the loop between in silico design and physical reality.

SI

Sarzmuza Issabek

COO & Business Development

Jan 15, 2026 4 min read

Computational protein design is only as valuable as its correlation with experimental outcomes. Our partnership with KAIST will produce the ground truth data needed to validate and improve every metric in the Foldexa pipeline.

Partnership Overview

The Korea Advanced Institute of Science and Technology (KAIST) is one of Asia's leading research universities, with a Department of Biological Sciences that houses state-of-the-art protein expression, purification, and binding assay capabilities. Our partnership establishes a closed feedback loop: Foldexa generates computational designs; KAIST expresses and tests them; the experimental results flow back into Foldexa's scoring models.

Collaboration Goals

  • Express and purify 200+ DiffAb CDR designs across 5 antigen targets (2026 Q1–Q2)
  • Measure binding affinity (SPR and BLI) for all expressed designs
  • Correlate Foldexa Score tiers (S/A/B/C) with experimental binding affinity
  • Publish correlation data in a peer-reviewed journal (target: Q3 2026)
  • Use experimental data to fine-tune AlphaFold2 scoring weights in the Foldexa Score formula
  • Establish benchmarks for CDR loop RMSD predictions across loop length categories

Why This Matters for Users

Right now, Foldexa's rankings are based purely on structural confidence metrics. We believe these correlate strongly with binding function — but we do not yet have the experimental data to prove it at scale. This partnership will either validate our scoring system or reveal where it needs to improve. Either outcome is a win for our users.

Open Data Commitment

All experimental validation data produced through this partnership will be made publicly available upon journal publication. We are committed to open science and reproducible benchmarks for computational protein design.

This collaboration represents exactly what we need in the field: a computational platform that is serious about experimental validation. We look forward to providing the ground truth.

Prof. Jae-Woo Lee, Department of Biological Sciences, KAIST

Share
SI

Written by

Sarzmuza Issabek

COO & Business Development