State-of-the-art cancer research relies upon the latest technologies to make advances in our understanding of tumors and to develop better treatments. Powerful technologies, such as next generation sequencing or advanced digital medical imaging, are rapidly evolving and produce ever increasing amounts of data.

Today’s scientific technologies can generate an enormous volume of data from a single analysis. For example, to identify molecular changes specific for cancer cells, new technologies now routinely generate datasets as large as 250Gb. The storage, analysis and interpretation of these large-scale datasets in cancer research and management demands a whole new approach. “The analogy often made is with trying to drink from a firehose,” explains Roel Verhaak, Professor and Associate Director for Computational Biology at The Jackson Laboratory (JAX) for Genomic Medicine in Farmington, Connecticut, USA. In January 2021, he was appointed part-time Professor of Computational Biology of Brain Tumors in the Department of Neurosurgery at Amsterdam UMC. Prof. Verhaak’s research seeks to gain new insights into brain tumors using computational analysis of high throughput datasets.

Computational analysis: an essential skill

Data is only good if insights can be gained from it. Rather than trying to interpret a dataset consisting of billions of bits of information, researchers are using computers to search for consistent patterns and behaviors within and between datasets. “This process of data analysis helps reduce the complexity and transforms the firehose into a more gentle stream that can readily be consumed,” says Prof. Verhaak. “Computational analysis is a skill that a lab can hardly do without.” Accordingly, bioinformatics and computational biology skills aimed at analyzing and interpreting large biological datasets are in increasing demand to advance interdisciplinary cancer research.

Finding patterns in the noise

The opportunities provided by high-throughput technologies and computational biology are plentiful. The concept of tumor-derived biomarkers that are used to guide clinical cancer management is not new, but increasingly these biosignatures are obtained by analyses of big data sets. For example, biomarkers which predict tumor response to immunotherapy have been identified thanks to computational biology. Information mined by algorithms can also be used to build models of complex molecular interactions in tumor cells to better understand the aberrant biology of cancer. In collaboration with Amsterdam UMC researchers, Prof.. Verhaak recently published a study in Nature detailing how certain brain tumor cells evolve over time.

A balancing act

Enormous datasets not only offer vast opportunities for discovery, but come with huge challenges. Algorithm design is critical to interrogate vast amounts of data. Data storage and computing power is also a challenge. “We are constantly balancing between our ability to afford [to store] data, while continuing to do the science we think is important,” says Prof. Verhaak. To ensure continuity and progress, his group has developed knowledge management tools such as web-based tables where the location of datasets, data files and tissue samples are centrally recorded. Cloud computing has some possibilities, but is not yet financially viable, although in the future Prof. Verhaak expects an integration between local and cloud-based facilities could meet most needs.

An eye to the horizon 

To stay current, researchers and research institutions must adapt their mindsets to embrace bioinformatics, says Prof. Verhaak. “It is hard to imagine a long-term successful career in research without developing a basic understanding of common computational approaches.” Dr. Verhaak advises that every young person who is preparing for a career in life sciences should learn how to perform basic computational analyses and develop the ability to interpret high-throughput analysis-based results. “Along the same lines, research centers need to invest in a computational infrastructure to store million gigabytes of data and enable computation via high-performance computing clusters. Such investments are not a luxury but a requirement for long-term success.”

Sharing data is also key. Roel Verhaak fosters worldwide multi-center collaborations to study brain cancers. Amsterdam UMC is part of the Glioma Longitudinal Analysis (GLASS) Consortium aimed at understanding brain tumor evolution and finding therapeutic solutions. Cancer Center Amsterdam extends a warm welcome to Roel Verhaak as a part-time professor at Amsterdam UMC.