Magnifying the minute
Making all those things we normally cannot see visible is the challenge that fascinates Fons Verbeek, Leiden's new Professor of Computational Bio-Imaging. Contributions from chemistry, optics and in particular computational sciences to making and processing these images give greater insight into the world of the minute. Inaugural lecture 12 January.
Observations are an essential element of empirical science, but there are many things that escape human observation, for example because they may be too far away or too small. It is with this last category that Professor Fons Verbeek works. 'To make small things visible you have to enlarge them. That's why the microscope was invented, an instrument that has since become iconic for science.'
‘We use knowledge from a very broad range of fields for microscopic imaging: physics for knowledge about optics, biology for the objects that we are examining, and chemistry for knowledge about the constrast agents used to view the object. And electronics is important for all the sensors and electronic components in modern microscopes.' The advent of the digital camera was the starting point for the rise of large-scale imaging. 'Digital cameras allow us to make pictures relatively quickly that we can then process directly.' Processing the images is the specific field of computational biology in which Verbeek is engaged.
Making 3D images
‘If we want to study an object, we ideally want to have it in our hands so we can look at it from all different angles and get to understand its structure.' How a 3D image can be made from microscopic recordings is an important aspect of Verbeek's research. A simple technique for doing this is to cut the object into thin slices, between 2 and 10 micrometres, and make a recording of each slice. Putting these images together again neatly gives you a 3D image. 'Another way of doing it is using tomography, a technique that we know from the CT scanner in hospitals.’ With this method the object under the microscope is rotated 360 degrees so that recordings can be made from every angle. Using a computer, these recordings are then converted into a 3D image. ‘This technique lets you do such things as making the 3D structures in a cell visible.'
All these techniques need ever more powerful computers. But it also works the other way round, Verbeek comments. 'As soon as there is more calculating power available, more new applications are devised.' One example is the recent collaboration with Naturalis. ‘We try to classify wood using microscopic images of its anatomy. We let the computer do this, with the help of deep learning.' This is a technique whereby a computer is fed an enormous number of examples of different types of wood, and the computer then learns to recognise the different types. 'This should eventually lead to an automated system that can trace imports of illegally sourced wood.' This is one example of how advances in computer technology go hand in hand with developments in microscopy.
Image above: a microscopic recording of the wood of a Douglas fir tree.