‘Automating data analysis is core part of science’

On functionalized surfaces, influenza viruses show rich and dynamic behaviour. ‘They bind via multiple, individually weak interactions,' says Erik Hamming. ‘They show a kind of walking behaviour that would be impossible when attached via one connection only.' 

These specific interaction characteristics are known in science quite recently, from 2011, and are understood better and better ever since. Now, with advanced automated data analysis, theoretical findings are: discovered, better understood and refined. Also, a novel biosensor, based on this multivalent “emergent” behaviour, comes into view.

While talking to Erik Hamming about his PhD work, he stresses the teamwork performance of scientific work ever again. ‘With my group colleague Nico Overeem, for example, I had daily consultations for over a period of four years,’ he says.

Erik: ‘While working on a novel biosensor I saw Nico struggling with complex image data, taking him days to analyze them properly. My efforts on automating the data accelerated his research and led to shared publications, both in our field of research and in theoretical interpretations of virus behavior. So, one could say that advanced data analysis is a core part of scientific progress. I am fully happy to contribute to that, using my specific expertise: complex data analysis using image data.’


Further key collaborations were with Stephan Block from Freie Univerität Berlin, willing to share his expertise on single virus spectroscopy. Nicolas Tito, Eindhoven University (and later: Electric Ant Lab), was of great help in developing the analytical virus-surface models. ‘The list is much bigger, but I would like to add Geert-Jan Boons of Utrecht University,’ Erik says. ‘His help was indispensable in functionalizing our surfaces with biomolecular bounding agents.’

Erik: ‘And, of course, the cleanroom experts of MESA+. They make the most of the functionalities present in the cleanroom. For example by exactly knowing beforehand what chip designs will eventually work, and which won’t. Working with them saves huge amounts of time. One month of work for me, learning everything along the way, is settled in one afternoon. That’s no exaggeration.’


Erik characterizes his PhD work as fairly fundamental in nature. ‘The data analyses part I perceived as an overall bird’s view,’ he says. ‘For example, by collaborating with virologist Erhard de Vries, Erhard mentioned many times that he learned a great deal. The data analyses helped him to look more deeply into underlying and more detailed molecular processes. His more application-oriented disciplinary focus was usually on the virus level as a whole.’

The virus characterization specialist helped Erik and his colleague Nico Overeem to make strategic research choices, as well as practical design chip choices. ‘In this way the gradient receptor sensor that Nico is working on will stay in view, to be used in the mid-term and long-term future, making our research really aimed towards future application.’

Future job

For some time during his PhD project, Erik considered an academic career. It is a pity that data analysis work is somewhat underestimated, he notes. ‘But I am able to prove it’s added value. And, I am convinced that this approach is to become more and more important in future science. Not choosing a career in academics was more of a personal consideration. Finding, creating and finding subsidies for my own research topics doesn’t suit me very well, I believe.’

Now, in his daily work, Erik uses his expertise in a more societal context. For example, complex data and traffic tracking sets are analyzed systematically in order to predict future bottlenecks and safety issues.

Erik: ‘By tracking and mapping brake use manoeuvres, unsafe crossings and other dangerous traffic situations can be predicted. My scientific approach enables me to reason from the complex data into underlying explanation structures. Within the team, I am keen on interpreting the data correctly. Only from clear and clean data, one can jump to correct conclusions. My colleagues value this input, I am glad to say.’