AI for Safer Antimicrobials

Accelerating AMP Discovery with Madam Therapeutics

The Challenge

Antimicrobial resistance is one of the biggest threats to global health.
Madam Therapeutics developed synthetic antimicrobial and anti-biofilm peptides (SAAPs) to treat resistant infections in hospitals. But designing new peptides that are both effective and non-toxic is complex. Experimental testing alone could not keep pace with discovery needs.

They needed a partner who could combine biological understanding with data science to predict antimicrobial activity and toxicity before compounds reached the lab bench.

Our approach

BioLizard’s team of Lizards applied a combination of natural language processing (NLP) and machine learning (ML) to transform unstructured biological data into predictive insights.

1. Data foundation

Using NLP and text mining, we built the largest annotated AMP database to date, cataloguing sequence, structure, activity, and toxicity data across thousands of peptides.

2. Model development

Created and validated ML models capable of sequence-based prediction of antimicrobial activity and toxicity.

3. Validation

Tested the models on new peptide sequences, showing high precision and recall and confirming their robustness across bacterial species.

4. Knowledge transfer

Delivered a reproducible workflow that can be extended to future peptide classes.

The Outcome

By combining biology-informed modeling with machine learning, the Lizards helped Madam Therapeutics:

  1. Accelerate discovery by reducing experimental screening time
  2. Gain new biological insight into sequence and toxicity relationships
  3. Validate lead compounds with strong predicted activity and low toxicity
  4. Advance promising peptides that are now in clinical development

The result was not just a successful project but a shift toward data-driven antimicrobial design.

Why it matters?

Traditional antimicrobial discovery depends on trial and error. By integrating explainable AI and curated biological context, BioLizard enables faster and more confident candidate selection. This project shows how data-driven design can directly feed clinical pipelines and how working with the right Lizards turns complex data into a competitive advantage.

Talk to a Lizard

Curious how AI and machine learning could speed up your discovery process? Let’s discuss your next therapeutic challenge over a coffee.

We are very happy with the time and dedication put into the project by an experienced and knowledgeable bioinformatician.

Leonie de Best
CBO, Madam Therapeutics