Development of machine learning solutions for prediction of functional and clinical significance of cancer driver genes and mutations are paramount in modern biomedical research and have gained a significant momentum in a recent decade. Given rapid proliferation of machine learning tools to address biological problems, there are several fundamental questions arising in the context of classification of cancer driver mutations. Can deep learning models achieve robust recognition of cancer driver mutations based solely on nucleotide information? What is the role of functional and structural predictors derived from biophysical analysis ? In this work, we have explored and integrated different machine learning approaches, including deep learning and convolutional neural networks for prediction and classification of cancer driver mutations by using and consolidating information across large datasets of functional and clinical mutations. The classifiers were benchmarked against their tree-based alternatives in order to evaluate the performance on a relative scale. We then integrated DNA-based scores generated by convolutional neural networks with various categories of conservational, evolutionary and functional features into a generalized random forest classifier. The results of this study have demonstrated that convolutional neural networks can learn high level features from genomic information that are complementary to the ensemble-based predictors often employed for classification of cancer mutations. By combining deep learning-generated scores with only two main ensemble-based functional features, we can achieve a superior performance of various machine learning classifiers. Our findings have also suggested that synergy of nucleotide-based deep learning scores and integrated metrics derived from protein sequence conservation scores can allow for robust classification of cancer driver mutations with a limited number of highly informative features.<br/>Through extensive comparative analysis with structure-functional experiments and multi-center mutational calling data from Pan Cancer Atlas studies, we have demonstrated the robustness of our models and addressed the validity of computational predictions. To address the interpretability of cancer-specific classification models, we have combined machine learning predictions of cancer driver mutations with autonomous molecular design of targeted mutation-specific inhibitors in protein kinase genes.

Deep convolutional neural networks for prediction of cancer driver mutations and autonomous design of targeted kinase inhibitors

Fall 2022 abstracts, posters, and presentations will launch in mid-September.

ACS Fall 2022

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.