Data from Heterogeneous Driving Effects Guide Personalized Tumor Treatments Targeting N<sup>6</sup>-Methyladenosine

<div>Abstract<p>Alterations to N<sup>6</sup>-methyladenosine (m<sup>6</sup>A) modifications can promote malignant progression by modulating gene expression through regulation of transcript metabolism. Quantifying the causal impact of m<sup>6</sup>A dysregulation at the population level could help guide personalized therapeutic interventions. In this study, we developed a causal framework that enables precise estimation of the driving effects (DE) of m<sup>6</sup>A dysregulation on tumor survival, establishing DE-based enhanced prioritization rules for anti-m<sup>6</sup>A therapies. The global average DE of m<sup>6</sup>A dysregulation across 9,647 tumors resulted in a decrease in overall survival of 180.1 days, considering a 5-year survival endpoint, suggesting potential overall benefits of treatment. Profiling of tumors most susceptible to m<sup>6</sup>A dysregulation and application of a modifier-mining tool revealed an effect modification in the PI3K/AKT/mTOR pathway. Specifically, the benefits of m<sup>6</sup>A-targeted treatment varied depending on baseline mTOR levels, which were validated <i>in vitro</i> and <i>in vivo</i>. Overall, focusing on m<sup>6</sup>A, this study established a paradigm leveraging DE over conventional risk to optimize personalized therapy.</p>Significance:<p>A causal inference framework based on m<sup>6</sup>A modifications that quantifies driving effects and characterizes high-benefit tumor profiles offers an improved approach for guiding personalized tumor treatment decisions.</p><p><a href="https://aacrjournals.org/cancerres/pages/data-science-special-series" target="_blank"><i>This article is part of a special series: Driving Cancer Discoveries with Computational Research, Data Science, and Machine Learning/AI</i></a>.</p></div>