For the visualization of proteins, interaction with the displayed model is indispensible in order to understand the three-dimensional protein structure. Accordingly, an efficient design of the user interface including display and input devices is crucial. A large number of input devices, some in 2D, some in 3D, and a large variety of display devices exist, which are commonly used for exploring large-scale 3D data structures. Unfortunately, not all combinations of input and display devices work together in a suitable and useful manner, either because they are not capable of 3D input or because, for instance, they need a hard table surface to work properly, making their utility for 3D navigation impractical. For example, in a large-screen, stereoscopic environment, where user immersion and freedom of motion in the display space are required, such a device would bind the user too much to a fixed location.