In a study that has the potential to add significant value to the landscape of early cancer-detection methods, DPhil candidate Hira Javaid and colleagues have used machine learning to predict SETD2 mutation status with remarkable precision using DNA methylation. The study, published in BMC Cancer, sheds light on the critical role of SETD2-dependent H3 Lysine-36 trimethylation (H3K36me3) in DNA methylation dysregulation across multiple cancer types and opens the door to more effective diagnosis and prognosis for patients.
SETD2, a gene often found to be mutated in various cancer types, has been associated with the deposition of de-novo DNA methylation. Until now, the functional consequences of SETD2 loss and depletion on DNA methylation and tumorigenesis remained elusive. However, this new study has unveiled a ground-breaking connection.
In a pan-cancer analysis spanning 24 different cancer types, Hira and her colleagues observed that both mutations and reduced SETD2 expression were consistently linked to DNA methylation dysregulation in 21 of these cancer types. This finding suggests a broader role for SETD2 loss in not only tumorigenesis but also cancer aggressiveness through DNA methylation disturbances.
(C) Disease ontology of expression-methylation correlated genes shows that many of the differentially methylated genes are associated with kidney neoplasm and neoplasm invasiveness
The implications of these discoveries are far-reaching, particularly in renal cancer. In this context, DNA methylation alterations were correlated with changes in the expression of vital genes such as TP53, FOXO1, and CDK4 – genes pivotal to oncogenesis, tumour suppression, and neoplasm invasiveness. Such findings have the potential to reshape our understanding of the mechanisms driving cancer development and progression.
However, the most novel aspect of this research lies in the integration of the unique machine learning approach for biomarker selection. Through a rigorous and robust machine learning methodology developed by Alessandro Barberis, the team developed and validated a 3-CpG methylation signature that accurately predicts SETD2 mutation status. This approach allows a better estimate of biomarker performance in the real world, which will help scientists develop biomarkers that are more readily translatable into the clinics. This signature not only holds promise for earlier and more accurate cancer diagnosis but also demonstrates a strong correlation with patient prognosis. This could enable oncologists to tailor treatments based on the individual molecular profile of each patient, leading to more personalised and effective therapies.
Hira Javaid, DPhil candidate, expressed excitement about the potential impact of this research and the impact of support from the Clarendon Fund and the Ioan & Rosemary Scholarship:
"I’m incredibly grateful to the Clarendon fund and Ioan & Rosemary scholarship without which I would not be able to pursue this DPhil at Oxford. It’s been a fulfilling and rewarding time and has paved way for me to continue to work in early cancer detection in the future”.
Read the full research article on the BMC Cancer website.
The Ioan & Rosemary James Undergraduate Scholarship for Overseas Students
The Ioan and Rosemary James Scholarship supports undergraduates from outside the UK to cover University and College fees, and to provide a living cost grant and travel support.
The Clarendon Fund
The Clarendon Fund first welcomed scholars to Oxford in 2001. The scheme was expanded in 2012 to include students from the UK and EU, therefore providing funding for all fee statuses. Throughout this period, the Fund’s aim has remained unchanged; to assist academically outstanding graduate students through their studies at the University of Oxford.