Big Data to find new questions

For centuries, data visualisation has helped humanity identify patterns and gain new insights. When London suffered a cholera outbreak in 1854, the English physician John Snow used a map to trace the source of the infection to the water pump in Broad Street, Soho. Now – in the era of Big Data – visualisation is more important than ever.

EMBL alumnus Nils Gehlenborg is pushing forward the visualisation of complex genomic and clinical datasets with highly innovative contributions. “I use computer science to build tools and visual interfaces that enable researchers to efficiently interact with biomedical data,” he explains.

As a PhD student in the Brazma Group at EMBL-EBI, Gehlenborg worked on making large collections of gene expression data visually accessible so that biologists could discover patterns more easily. He is now Assistant Professor of Biomedical Informatics at Harvard Medical School, where he develops tools to visualise various types of data from large-scale cancer genomics studies such as The Cancer Genome Atlas. He also tackles visualisation problems across scales as co-investigator of the 4D Nucleome Data Coordination and Integration Center, funded by the US National Institutes of Health. With his team, he has created visualisation tools that allow scientists to see patterns at the chromosome level, and then to zoom down to find patterns at the level of DNA bases – the individual letters of the genetic sequence.

“It’s a tool that helps us formulate new questions, rather than one which gives us all the answers.”

Gehlenborg’s lab is also exploring clinical applications of data visualisation by using data from electronic health records. “My long-term vision is that all the data generated by clinical analysis, sensors, and smartphones will be integrated,” he explains. “Rightly visualised, this will help doctors diagnose and treat patients using precision medicine. The visualisation and accessibility of the data will also help individuals understand how their health might be influenced by their behaviour.”

In the era of Big Data, biology is increasingly becoming a data-driven science, and this is changing the way we think about biological problems. “Traditionally, biology has been driven by hypotheses,” explains Gehlenborg. “You observe something in nature, come up with a hypothesis, and design an experiment to test it. But today we’re often collecting large amounts of data without a clear hypothesis. We try to find patterns and then maybe propose a hypothesis that we can test. This is where data visualisation is really strong. It’s a tool that helps us formulate new questions, rather than one which gives us all the answers.”

Berta Carreño