Prediction of Anti-Glioblastoma Drug-Decorated Nanoparticle Delivery Systems Using Molecular Descriptors and Machine Learning

Abstract

The theoretical prediction of drug-decorated nanoparticles (DDNPs) has become a very important task in medical applications. For the current paper, Perturbation Theory Machine Learning (PTML) models were built to predict the probability of different pairs of drugs and nanoparticles creating DDNP complexes with anti-glioblastoma activity. PTML models use the perturbations of molecular descriptors of drugs and nanoparticles as inputs in experimental conditions. The raw dataset was obtained by mixing the nanoparticle experimental data with drug assays from the ChEMBL database. Ten types of machine learning methods have been tested. Only 41 features have been selected for 855,129 drug-nanoparticle complexes. The best model was obtained with the Bagging classifier, an ensemble meta-estimator based on 20 decision trees, with an area under the receiver operating characteristic curve (AUROC) of 0.96, and an accuracy of 87% (test subset). This model could be useful for the virtual screening of nanoparticle-drug complexes in glioblastoma. All the calculations can be reproduced with the datasets and python scripts, which are freely available as a GitHub repository from authors.

Prediction of Serum Adsorption onto Polymer Brush Films by Machine Learning

“…Machine learning algorithms evaluating quantitative structure–property relationship (QSPR) emerged as a game-changer in this situation. The algorithms have been employed to design drug carriers, antibacterial nanoparticles, and other biomaterials. − The algorithms can also unveil the complex relationship between the chemical structures of molecules constituting SAMs and their affinity to proteins. However, very few machine learning-based studies on various antibiofouling films have been reported.…”

Systematic Review of Supervised Machine Learning Models in Prediction of Medical Conditions

“…This model could be applied for the screening of nanoparticle-drug complexes in glioblastoma. [31] The bagging ensemble method improves reproducibility of both cortical and subcortical functional parcellation of the human brain neuroimaging (more than 300 samples). [32] AdaBoost was used to differentiate colorectal neoplasia from normal tissue (AUROC up to 0.95 on 64 samples from 16 patients).…”

“…Neurological conditions are among the most common medical system subject to ML model applications. [29,31,32,38,46,53,61-63,66,69,70] The most frequent type of data used in these applications were imaging data. Images consist of spatially coherent pixels in a local region, meaning that pixels close to each other share similar information.…”

“…Accurate predicting the development of cancers or complications of cancers could indicate earlier diagnosis and therapeutic approaches that would improve outcomes. [26,31,33,35,37,39,41,53,55,58-60,73] The majority of these ML applications use imaging data (most often histologic type) for classification of malignant versus benign tumors. Cardiovascular conditions and DM are among the most common medical conditions used in predictive analysis.…”