RNA Sequencing-Based Antisense Oligonucleotide Drug Modeling for Personalized Lung Cancer Therapy and Tumorigenic Inhibition

Lung cancer accounts for 18% of all cancer-related deaths in the United States, leading the list. Surgical removal of small tumors remains challenging and current treatments are not completely cost effective. Oligonucleotides, particularly antisense oligonucleotides (ASOs), represent a promising noninvasive therapeutic approach to inhibit the production of cancercausing proteins by binding to their target mRNAs. This study aims to create a computational framework that designs oligonucleotide drugs targeting mRNAs associated with lung cancer and the expansive tumor microenvironment. Using RNA-sequence data from the GDC data portal, a multi-step pipeline including differential gene expression analysis and feature selection was developed to pinpoint key biomarkers affecting patient survival. A stacked ensemble combining gradient boosting and a multilayer perceptron (MLP) classifier was trained on both clinical and genetic data to enable personalized biomarker targeting. A custom ASO design tool generated proposed sequences and binding affinities were confirmed using the IntaRNA and RNAFold software. The deep learning model identified the genes IGKC, FTL, and MT-CO1 as prime ASO targets, achieving a model accuracy of 89% with supporting metrics (precision, recall, and F1-score) all above 85%. Additionally, RNA-RNA binding predictions showed high affinity with minimal off-target effects. This study proposes a computational framework for ASO-based therapies in lung cancers as a potentially scalable and noninvasive strategy. To evaluate the variability in ASO effectiveness, an ANOVA and Eta Squared analysis indicated a statistically insignificant link between sequence choices for genes and binding affinity, displaying uniform hybridization ability. However, the intricate biophysical RNA structure interactions remain unresolved. Future research will concentrate on enlarging datasets, refining ASO sequences, and using experimental assays to improve clinical applications for various other diseases.