Related Experiment Video
Updated: Jul 15, 2025
09:09
Synthesis of Aptamer-PEI-g-PEG Modified Gold Nanoparticles Loaded with Doxorubicin for Targeted Drug Delivery
Published on: June 23, 2020
6.0K
Machine Learning Reinforced Genetic Algorithm for Massive Targeted Discovery of Selectively Cytotoxic Inorganic
Susan Jyakhwo1, Nikita Serov1, Andrei Dmitrenko1
1International Institute "Solution Chemistry of Advanced Materials and Technologies", ITMO University, Saint-Petersburg, 191002, Russian Federation.
Small (Weinheim an Der Bergstrasse, Germany)
|September 29, 2023
Summary
This study introduces a novel computational method for discovering nanoparticles (NPs) with selective toxicity against cancer cells. This approach accelerates the identification of effective and safe nanoparticle drug delivery systems (DDSs).
Area of Science:
- Computational chemistry and materials science
- Nanotechnology and drug delivery
- Bioinformatics and machine learning
Background:
- Nanoparticles (NPs) have been utilized as drug delivery systems (DDSs), but often lack selectivity.
- Emerging evidence shows NPs can exhibit selective cytotoxicity against cancer cells due to metabolic differences.
- Existing methods for identifying selectively cytotoxic NPs are limited in scope and throughput.
Purpose of the Study:
- To develop and validate a high-throughput in silico screening approach for discovering selectively cytotoxic inorganic NPs.
- To identify novel NP-based DDS candidates for targeted cancer therapy, specifically liver cancer.
- To explore the potential of NPs for broader therapeutic applications beyond cancer.
Main Methods:
- Trained a gradient boosting regression model to predict cell viability in NP-treated cell lines (Q2 = 0.80, RMSE = 13.6).
- Developed a machine learning-reinforced genetic algorithm for rapid screening of >14,900 NP candidates per minute.
- Validated the approach by screening DDS candidates for liver cancer treatment using HepG2 and hepatocyte cell lines.
Main Results:
- The predictive model demonstrated good accuracy in assessing NP cytotoxicity.
- The ML-reinforced GA efficiently screened a large chemical space of NPs.
- Identified silver nanoparticles (Ag NPs) with a selective toxicity score of 42% for liver cancer treatment.
Conclusions:
- The proposed in silico screening approach enables massive, targeted discovery of selectively cytotoxic inorganic NPs.
- This method significantly accelerates the identification of promising NP-based DDS candidates.
- The approach holds potential for clinical translation and expanding NP applications to other organisms like bacteria and fungi.
