Cancer Research in the Age of Big Data

This article is part of a special series: Driving Cancer Discoveries with Computational Research, Data Science, and Machine Learning/AI.Over the last two decades, cancer researchers have elucidated the clinical correlates and signatures of a wide range of molecular features in cancer. We now know that there are more than 400 different types of cancer, each with distinct genomic, epigenomic, transcriptomic, proteomic, metabolomic, and other signatures. Heterogeneity of these signatures can be profound due to the intrinsic features of cancer cells, as well as the surrounding tumor microenvironment and host macroenvironment, and this can be further exacerbated by cancer treatments. Elucidation of these features of cancer is essential for understanding the fundamental biology up- and downstream of key molecular, microenvironmental, epidemiologic, and clinical manifestations of cancer, which requires collection and analysis of large quantities of data.The massive volume of cancer data generated and the pace at which they are being generated are intertwined with the rapid increase in data science, encompassing computational biology, mathematical modeling, machine learning (ML), and artificial intelligence (AI). The next decade of data science research will increasingly exploit large sample sizes to better map out the complex relationship of drivers, processes, and features that shape the biology of cancer over time and space and in the context of clinical trials. Yet, the rapid advancement of data science in cancer research has outpaced the traditional structures for publishing and sharing this work. Cancer Research aims to be a central platform that acknowledges the unique contributions of the data sciences to cancer research and fosters in-depth interdisciplinary dialog.Currently, articles at the intersection of data science and cancer research are published throughout the literature, and authors may find it difficult to identify the right home for publishing their studies. Some journals may not include data scientists on their editorial boards, whereas others may underappreciate the clinical and translational implications of a computationally focused study. Many excellent articles might be too technical for general medical journals and not technical enough for data science audiences. The lack of suitable datasets for validation of computational and ML models can also impede publication of otherwise high-quality data science research. Together, these issues culminate to delay the dissemination of knowledge, which slows progress toward the goal of improving cancer care. Clinical and translational researchers may not learn of cutting-edge discoveries that could inform diagnosis and treatment of their patients. The lack of a centralized venue also stalls cross-disciplinary collaboration. When cancer data scientists and cancer biologists are siloed, opportunities for synergy and translation are missed.At Cancer Research, we appreciate that data science is an interdisciplinary form of study, and this is particularly important for tackling the complex problem of cancer. The convergence of data science, foundational cancer biology studies, and translational oncology published within Cancer Research will stimulate ideas and create new opportunities for cooperation to work on innovative solutions. Given the broad audience and scope of Cancer Research, it is ideally situated to foster integrative efforts by publishing cancer data science alongside all other aspects of cancer biology within the journal.As the foundational journal of the American Association for Cancer Research, Cancer Research intends to fill the need for a publishing home for data science articles that is both rigorous in its technical expectations and deeply embedded in the oncology research community. We have developed a two-stage strategic plan to achieve this goal. First, when I became Editor-in-Chief of Cancer Research, I revised the journal sections to include ones related to Computational Cancer Biology and Technology and Convergence Science. At the same time, new senior editors were recruited to our editorial team to complement the already talented group of editors overseeing the review process of manuscripts related to data science.The second stage that we are now implementing is to launch a special series of articles within Cancer Research devoted to all aspects of data science, including but not limited to computational research, mathematical modeling, systems biology, and ML/AI. The goal of this series is to publish highly impactful research that implements cutting-edge computational and data-driven approaches to accelerate biological discoveries and facilitate therapeutic and clinical innovations. To assist me in oversight of this special series, I recruited three advisory editors whose work spans the spectrum of the cancer data science we hope to attract for this series and for Cancer Research over the longer term. These advisory editors are Dana Pe’er, a leading computational biologist combining single-cell and spatial profiling technologies with ML approaches to study cancer, immunity, and development; Regina Barzilay, an expert in computer science who develops ML methods for drug discovery and clinical AI; and Peter Van Loo, a renowned researcher in cancer omics working to elucidate the evolutionary history and subclonal architecture of tumors.Together with the advisory editors and our dedicated senior editors, we are devoted to publishing high-impact articles focused on the intersection of data science and cancer. In doing so, it is our hope to stimulate discussion to help spur the field forward and drive scientific discoveries to bring us closer to understanding and curing cancer.C.A. Iacobuzio-Donahue reports other support from Bristol Myers Squibb outside the submitted work.