nganesh's picture
Update README.md
68920f3 verified
metadata
license: other
license_name: apoha-license
license_link: LICENSE
language:
  - en
tags:
  - antibody
  - developability
  - biophysical
  - monoclonal
  - prediction
  - biology
  - pharma
extra_gated_fields:
  Name:
    type: text
    optional: true
  Organisation:
    type: text
    optional: true
  By requesting access, I acknowledge that I read, understand, and agree to the terms of use and licensing (https://huggingface.co/datasets/apoha/antibody_jain_vibescore/blob/main/LICENSE) associated with this dataset:
    type: checkbox

Using an excitable substrate and only 10μg of antibodies, the Liquid Brain® performs single-shot classification of antibody developability risks, demonstrated through case studies matching the combined predictive capability of 12 biophysical assays across 135 clinical-stage antibodies. In another customer study, it identified outliers in 70 VHH - Fcs fragments with >80% accuracy using first-principle dimensionality reduction and unsupervised clustering.

Our dataset of risk scores for clinical antibodies is now publicly available here. Write to us at [email protected] to learn how Liquid Brain® is enabling a state-to-function approach for protein behavior prediction and how it can integrate seamlessly into your workflows to drive more reliable decision-making in antibody development.


This dataset contains the following information about the antibodies from Jain et. al :

  1. Assay values and flags
  2. Termination/Approval status
  3. Sequence based descriptors
  4. VIBE score derived from proprietary data generated by the Liquid Brain.

The dataset repo also contains results from a case study that was performed to validate the use of the VIBE score for predicting antibody developability (becomes available once access is granted).