EPCO-08. CONVERGENCE OF NON-CODING SOMATIC MUTATIONS ON CANCER DRIVER PLEXUSES IN PEDIATRIC HIGH-GRADE GLIOMA

Abstract

Abstract Central nervous system tumors are the most common pediatric malignancies after leukemia, and the most common cause of cancer death among persons ages 0 – 14 [1,2]. Despite its low incidence rate, pediatric high-grade gliomas (pHGGs) account for over 40% of all childhood brain tumor deaths [3, 4]. As such, it remains the focus of many research teams across the globe. The discovery of recurrent mutations in genes encoding histone H3.3 [5] was a breakthrough in the genetic basis of pHGGs and highlighted the relevance of epigenetic reconfiguration for oncogene activation. However, much work is still needed to uncover driver alterations responsible for tumor initiation or progression. pHGGs have frequent epigenetic dysregulation [6,7] and abundant somatic mutations in the noncoding regions of the genome [8], but the functional relevance of the noncoding somatic mutations has not been evaluated systematically. Here, we assess correlations between noncoding somatic mutations in pediatric brain-specific cis-regulatory elements (CREs) and changes in gene expression. We demonstrate the sparsity of recurrent mutations in CREs. Then illustrate the power of graph database in consolidating multi-omics datasets for efficient queries. Our network approach for charting the genomic, epigenetic, and transcriptomic landscape of pHGG allow us to 1) identify clusters of CREs associated with dysregulated gene expression, 2) nominate transcription factors whose binding affinities are recurrently impacted by somatic mutations found in pHGGs, and 3) establish regulatory networks potentially guiding pHGG tumor growth and progression.