Antimicrobial Peptide Engineering With Microfluidic ML Screening

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing technologies struggle to engineer antimicrobial peptides with desired activity ranges under specific culture conditions, particularly against pathogenic bacteria in environments like human gut microbiota, and there is a need for efficient methods to tune microbial populations in industrial processes.

Innovation Solution

A method involving in vitro transcription and translation of antimicrobial peptides in microfluidic systems, using translation and transcription solutions with reagents, and iterative machine learning to design variants until desired activity is achieved, allowing for the engineering of peptides with tailored activity against microbial organisms.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If traditional methods are used to engineer antimicrobial peptides, then the process is simple, but the ability to achieve desired activity ranges under specific culture conditions is poor

Engineering Contradiction:
Improveactivity range precisionVSAvoidengineering process complexity
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The patent implements an iterative engineering process where peptide sequences are dynamically optimized through multiple cycles of variation, selection, and testing. Machine learning models continuously update design parameters based on experimental results, enabling adaptive optimization of antimicrobial activity under specific culture conditions rather than relying on static traditional methods.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system incorporates feedback loops where the activity of engineered peptides under specific culture conditions is measured and fed back into machine learning models. This feedback enables the models to refine predictions and guide subsequent engineering iterations, progressively improving activity range precision while managing process complexity through intelligent automation.

Inventive Principle:
Principle #23Feedback

2Reliability

If antimicrobial peptides are engineered for high activity against pathogenic bacteria, then the therapeutic effect is improved, but the risk of resistance development increases

Engineering Contradiction:
Improvetherapeutic effectivenessVSAvoidresistance development
Core Design Contradiction:
ReliabilityVSObject-affected harmful factors

Solution Approach 1:

The patent engineers peptides with localized optimized properties by designing sequences that target specific structural features of pathogenic bacteria while preserving compatibility with beneficial microbiota. Machine learning models identify and optimize local sequence motifs that provide high antimicrobial activity against pathogens without triggering broad-spectrum resistance mechanisms.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The system optimizes multiple peptide parameters simultaneously (sequence composition, charge distribution, hydrophobicity, length) to achieve high therapeutic effectiveness. By carefully balancing these parameters, the engineered peptides maintain potent activity against pathogens while reducing selective pressure that drives resistance development.

Inventive Principle:
Principle #35Parameter changes

3Manufacturing precision

If iterative testing of peptide variants is performed to achieve desired activity, then the activity precision is improved, but the time required for engineering increases

Engineering Contradiction:
Improveactivity precisionVSAvoidengineering time
Core Design Contradiction:
Manufacturing precisionVSLoss of time

Solution Approach 1:

Machine learning models perform preliminary design and prediction of peptide variants with desired activity characteristics before experimental testing. This pre-screening of virtual peptide libraries identifies promising candidates that are then prioritized for experimental validation, reducing the number of iterative cycles needed and accelerating the overall engineering process.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent replaces traditional trial-and-error experimental iteration with machine learning-driven in silico optimization. Computational models predict peptide activity and guide sequence design, substituting physical iterative testing with virtual screening and prediction, thereby maintaining high activity precision while significantly reducing engineering time.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentEP3802789B1Engineering antimicrobial peptides
Publication Date: 2026.04.08 SYNGULON
  • EP3802789B1 patent drawingFigure 1
  • EP3802789B1 patent drawingFigure 2

AI summary

Embodiments herein relate to methods, systems and kits for engineering antimicrobial peptides such as bacteriocins, for example to have a desired range of activity in a desired range of culture conditions. The antimicrobial peptides may be engineered to have a particular activity for a particular culture, environmental conditions or a range of conditions. Some embodiments include screening an antimicrobial peptides or several candidate antimicrobial peptides for a desired activity. Some embodiments include an iterative process for engineering antimicrobial peptides such as bacteriocins. In some embodiments, the process is performed by automated machine learning.