Mining cancer genomes for copy number alterations identifies glycosylation enzymes as oncogenic drivers
- Pranoy Sahu
- Francesco Russo
- Domenico Russo
- Ilenia Agliarulo
- Pasqualina Ambrosio
- Riccardo Rizzo
- Matteo Lo Monte
- Nicola Normanno
- Silvia Soddu
- Francesca Carlomagno
- Alberto Luini
- Seetharaman Parashuraman
2026-04-01
Altered cell-surface glycans are established cancer biomarkers, yet no oncogenes have been identified within glycan biosynthesis machinery. This represents a critical gap, as defining a gene as a true oncogene, rather than merely a component of an oncogenic pathway, reveals targetable dependencies that can improve clinical decisions. To date, no gain-of-function mutations have been detected in glycogenes, and the search for such mutations is largely saturated. To address this gap, we developed a bioinformatic–experimental pipeline to identify copy number alteration (CNA)-based driver genes, overcoming noise from passenger genes. The approach recovered known oncogenes and tumor suppressors, while revealing novel candidates, including glyco-oncogenes. Focusing on the glycosphingolipid (GSL) biosynthetic pathway, we validated B4GALT5 as a bona fide glyco-oncogene whose genomic amplification drives proliferation, oncogene addiction, and poor prognosis, effects that can be reversed by targeted pathway inhibition. Mechanistic studies show that B4GALT5 promotes cancer cell survival via integrin-Src signaling under anchorage-independent conditions. Collectively, these findings establish glycosylation enzymes as a druggable oncogene class and provide a resource of high-confidence CNA-based cancer regulatory genes.