It was mentioned by professor in the lecture that the shape of proteins and various chemicals are very crucial since its the shape that is responsible for all reactions. But my question is how exactly is the shape produced on the screen verified with what actually is the shape of the chemical?
@Asterix In a context like molecular dynamics, biochemical simulation, etc., what appears on the screen is not important. In other words, rendering or rasterization does not come into play. Instead, the focus is on things like the shape and forces in the environment (in graphics what we might call "geometric modeling" and "animation"). The important thing, of course, is that the models being used to capture the physics and chemistry accurately reflect reality. Building such models is not easy: you're taking measurements at very small spatial scales, of phenomena that often take place on extremely short time scales. Correctly taking measurements, acquiring data, building models, etc., is largely the work of folks in biology, chemistry, etc. (who of course have to worry about many of the same issues we've discussed in class, such as sampling and aliasing!). The job of computer scientists is then to ensure that the computation we do accurately reflects these models. In the particular context of geometry, this means picking representations that capture the functionally important features of shape, and algorithms that allow accurate predictions to be made in a reasonable amount of time. It also means working together with physical scientists to validate numerical models—for instance, running "sanity checks" of numerical simulations against known solutions.
Importantly, the models we use in graphics are often not appropriate for doing science, because the approximations we make may be guided by things like efficient computation and human perception, rather than quantitative accuracy. There is a gradual push in graphics toward more accurate physical models, more rigorous treatment of geometry, etc., but one does need to be careful not to conflate beautiful pictures with accurate predictions. On the flip side, this additional flexibility often allows researchers in graphics to take bigger risks, i.e., try weird and wild ideas (which might make perfectly good pictures) that someone in traditional scientific computing might not immediately be willing to consider (because they can't be confident it will give a correct result). In the longer term, these "weird and wild" ideas can feed back into broader scientific computing. For instance, many techniques from rendering, animation, geometry, etc., do nonetheless feed back into the broader scientific ecosystem. Even if these methods are not used directly (and sometimes they are), they can inspire very different ways of thinking beyond the status quo. This kind of interaction is part of what makes graphics so exciting: it is very much a 'nexus' at the center of many different things (science, art, mathematics, computation, theory, applications, ...).