Computational and functional screens to identify actionable neuroblastoma drivers

The National Cancer Institute, part of the Department of Health and Human Services National Institutes of Health, awarded Duke University $685,186 on August 6, 2026, under the Cancer Biology Research program (CFDA 93.396) to conduct computational and functional screens identifying actionable neuroblastoma drivers.

The award funds a multifaceted research effort applying novel computational analysis pipelines to existing genomics datasets, integrated with functional genomics screening and high-throughput drug sensitivity studies. Deliverables include new insights into neuroblastoma biology, methods to identify targetable genes or pathways within individual tumors, identification of new drugs or drug pairs informed by computational analyses, and new prognostic molecular biomarkers to guide treatment. The project addresses the poor outcomes of neuroblastoma, a leading solid cancer in children affecting approximately 750 children and young adults annually in the United States, for which current therapies remain inadequate and precision medicine approaches targeting recurrent mutations have not fully succeeded.

Duke University's Office of Research Administration Division serves as the recipient. Work is performed in Durham, North Carolina. The period of performance runs through July 31, 2031. The assistance type is a Project Grant (R01).

Abstract

Our overall goal is to improve outcomes for children with forms of cancer that cannot be eradicated using current therapies. Here we focus on neuroblastoma, an especially challenging cancer that accounts for a large proportion of childhood cancer deaths each year. To combat neuroblastoma, we have assembled a multifaceted team to explore a new approach in which novel computational analysis pipelines are applied to existing genomics datasets and integrated with a functional genomics screen and high-throughput drug sensitivity studies. Our deliverables will include new insights into neuroblastoma biology, a new way to identify targetable genes or pathways within an individual tumor, new drugs or pairs of active drugs that will be informed by our analyses, and new prognostic molecular biomarkers to guide treatment. In the US each year, approximately 750 children and young adults develop neuroblastoma, among the most common solid cancers in children. For many, the chance of cure is small, and this bleak outlook remains despite the use of multifaceted therapies in recent years. New molecular biology and genetics tools at the close of the last century brought important insights into the underpinnings of neuroblastoma. However, that knowledge has not sufficiently transformed treatment, nor does IT fully explain the poor outcomes. Many anticipated that the genomics revolution would usher in precision medicine focused on the mutant forms of proteins anticipated to drive the disease. That promise, too, has not been fully realized in cancers like neuroblastoma that lack highly recurrent, targetable mutations. During the pandemic, our team came together to close this GAP using a new approach and funding from an NCI R21. Given the few recurrent mutations in this disease, we proposed neuroblastoma to be a cancer in which normal developmental programs are corrupted by altered gene expression and that the altered gene expression is hard-wired by copy-number gains in oncogenic drivers. In the last two years, we applied our own, novel, computational pipeline, called iexcn, to identify approximately 45 oncogenic drivers in neuroblastoma, and we have begun to confirm the importance of those genes in neuroblastoma cells and how aberrations in iexcn genes may help guide therapy. We will move the project toward the clinic by completing three new aims: 1) use a high-throughput functional genetics approach to reveal how the iexcn genes drive neuroblastoma; 2) use a chemical biology approach to reveal how iexcn gene status influences anti-cancer drug effectiveness; and 3) apply AL-based deep learning to incorporate survival data into a new algorithm, cnexsurvnet, to better understand the genes driving poor outcomes in neuroblastoma. Nearly all of the experiments are conducted in vitro or in silica using cultured neuroblastoma cells and computational approaches. Beginning in year 2, we plan to use patient-derived xenografts in laboratory mouse models to secure additional confirmation of potentially targetable iexcn genes. Historically, such pre-clinical testing has been necessary to justify testing new therapies in children with neuroblastoma and other forms of cancer, which we anticipate to shortly follow the completion of this project.