Related Experiment Video
Updated: Aug 10, 2026
09:41
Formulating and Characterizing Lipid Nanoparticles for Gene Delivery using a Microfluidic Mixing Platform
Published on: February 25, 2021
High-throughput microfluidics and machine learning-assisted screening of lipid nanoparticle formulations for siRNA
Mingzhi Yu1, Luhan Wang1, Dongsheng Liu2,3
1Centre of Micro/Nano Manufacturing Technology (MNMT-Dublin), School of Mechanical & Materials Engineering, University College Dublin, Dublin 4, D04 V1W8, Ireland.
Materials Today. Bio
|August 8, 2026
Summary
Optimizing lipid nanoparticle (LNP) formulations for siRNA delivery is complex. This study developed a machine learning workflow to accelerate LNP development, identifying promising candidates for enhanced gene silencing.
Area of Science:
- Biotechnology
- Nanotechnology
- Molecular Biology
Background:
- Lipid nanoparticles (LNPs) are advanced non-viral vectors for RNA delivery.
- Optimizing LNP formulations is challenging due to extensive formulation possibilities.
Purpose of the Study:
- To establish a high-throughput microfluidic and machine learning-assisted workflow for prioritizing LNP formulations for siRNA delivery.
- To identify optimal LNP formulations for effective gene silencing.
Main Methods:
- Generated a library of 864 LNP formulations.
- Utilized microfluidics for high-throughput preparation and characterization of 203 empty LNPs.
- Employed machine learning models to predict particle size and PDI for the entire library.
- Selected 25 candidate formulations for siRNA loading and biological validation.
Main Results:
- Machine learning models accurately predicted LNP physicochemical properties.
- Lipid concentration positively correlated with size; PEG ratio negatively correlated with size and positively with PDI.
- All selected LNPs maintained particle sizes below 150 nm after siRNA loading.
- SM-102 and CKK-E12 based LNPs demonstrated potent S100P gene silencing in vitro.
- CKK-E12 LNPs showed higher lung accumulation in vivo.
Conclusions:
- Favorable physicochemical properties are essential but not sufficient for optimal biological performance.
- Biological validation is crucial after model-assisted screening for siRNA-LNP development.
- The integrated workflow facilitates efficient siRNA-LNP development.
