Polymer design via SHAP and Bayesian machine learning optimizes pDNA and CRISPR ribonucleoprotein delivery
Problem
Designing polymer carriers that efficiently deliver pDNA and CRISPR-Cas9 ribonucleoprotein into cells is hard because the structure-property relationships in combinatorial polymer libraries are complex and poorly understood.
Why it matters
Effective delivery carriers are a bottleneck for gene editing and gene therapy reaching patients.
What the AI did
Explainable machine learning (SHAP) analyzed 432 formulations to reveal design parameters for pDNA vs RNP delivery, and closed-loop Bayesian optimization of 552 formulation ratios further enhanced in vitro performance.
Results
Top three polymers gave higher signal and stable transgene expression over 20 days in vivo, with 1.7-fold enhancement over controls.
Impact
ML-guided design demonstrably improved polymer carrier performance in vitro and in vivo.
Record
| Venue | Chemical Science |
|---|---|
| Published | 2024 |
| Work type | article |
| Authors | Rishad J. Dalal, Felipe Oviedo, Michael C. Leyden, Theresa M. Reineke |
| Citations | 47 (OpenAlex, at import time) |
| References | 54 |
| Model or method | SHAP explainable ML; Bayesian optimization |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1039/d3sc06920f |
| Open access | diamond |
| Keywords | polymer design, Bayesian optimization, SHAP, CRISPR delivery, gene therapy, machine learning |
Importance rating
significant — thousands to millions affected
Delivery carriers are a key enabler for nucleic acid medicines and gene editing; abstract shows ML-guided design with in vivo validation and 1.7-fold enhancement, a genuinely new design capability, but no patient-scale evidence.