 That was quite a lot of theoretical simulation concepts before I take you through a workflow how to do this in practice. Let me just wet your appetite by showing you what both students and postdocs in my group are using this for and the type of problems we tend to go after with simulations. This is worked on by Anna Johansson, one of my first PhD students and she was using simulations in the early 2000s to understand what happens when you insert helices in membranes that contain residues that are not hydrophobic either slightly polar or in some cases even charged. Very simple simulations by today's standards but they actually enabled us to understand first how these sites in so-called snorkel they stretched out into the water they occasionally pull in water in the membrane and in some cases they end up tilting or distorting the helix and it was even so popular that this particular image ended up in a one textbook by camp. This is another project performed by Samuel Murai who was a postdoc in my group some 10 years ago. He was using simulations of ligand-gated ion channels and particular studying where ligands, alcohol, binds. You see those purple and orange regions he identified those as places where the alcohol molecules spontaneously diffused and bound. That was somehow related to one cavity found but Samuel found a second cavity that nobody had seen in experiments before and the amazing thing is that now 10 years later we have been able to confirm this in experimental methods. So it's another example where his simulations were able to predict that before NFS saw it experimentally. So it's not just a matter of understanding simulations can predict things. This is a similar system called the Gamma Aminobutyric Acid Receptor where one of my current PhD students, Johan Chuang, has been doing amazing simulations together with Rebecca Howard also in the team. So the GABA receptor is important in your nervous system and one of the molecules that we know binds us is something called diatopan. You might have heard of this more under the trade name Valium which is it's a strong sedative and the relaxing. If you overdose sorry this is the reason why Valium has a strong effect on your nervous system but occasionally people overdose on it and if you overdose on it you can almost see the properties here right that it's tightening up this channel and changing the way the channel behaves. If you overdose on it you likely have too much of this and it's bound to all your channels and bad things starts to happen. If you then come into the emergency room they're going to give you flumazenil which is a known antidote to this. Do you see how the flumazenil compound kind of creates the opposite effect? It spreads the subunits apart. Johan Reba published this last year together with the Ryan Hibbs group. It's an amazing piece where they've actually been able to explain the fundamental way anesthetics and benzodiazepines act on these channels in the nervous system both when they close the channel and when they open the channel. Again would not have been possible without a combination of simulations and amazing cryo experiments by the Ryan Hibbs group. This is a very old story love story of mine. When I was a postdoc we worked together with the panda group at Stanford to simulate and folding of small proteins. Normally we couldn't fold proteins in those days this was around 2003 or so but remember those kinetics the transition barriers that is never a matter of zero probability but it's just that the probability of crossing a barrier is very low. What if instead of running one long simulation I run many many short simulations? We did that in a project called Folding at Home that Vijay pioneered at the time and this is one out of the 10,000 trajectories of a small protein called BBA5. It's called BBA5 because it contains two beta strands beta beta and then one alpha helix BBA and it was the fifth constructed such molecule. Let's hit the button and see what happens. You get a rapid hydrophobic collapse and then you're going to fairly quickly going to see that the helix is forming that's great because it agrees with the theory from lecture five. Here you're actually starting to see the helix forming too but that helix sheet is not quite right it's kind of opposite but since it's a hydrophobic collapse it will take a lot of free energy for this protein to expand and it eventually will expand somewhere roughly now or so yes the helix sheet was unfolding and then the sheet is going to refold again and there we actually have a sheet that corresponds closely to the experimentally determined structure I think it's roughly three angstrom away. This showed many things to us first we showed that we can fold proteins it showed that the approach of using screensavers all over the world works remarkably well and simply by having statistics because we had like 10 folding trajectories and today folding at home frequently has hundreds of folding trajectories we can study kinetics and ask things with much better statistics for instance is the formation of this helix independent of the formation of the sheet or do they help each other we will come back to that in the lecture when we talk about protein folding in this particular case it turned out that they were independent so that's a handful of examples I will show you a few more this is you Juan trying in action again he's great at providing illustrations isn't he. Large ligand gaited ion channel two and we have a ligand here that's responsible for the gating it binds in the outermost part here the extracellular part that creates a protein earthquake that opens up the central pore where an ion will diffuse through and we also know that this process is somehow controlled by lipids bound on the surface these systems are too large to start in atomic detail each of these lipids is 130 atoms there are other techniques so called core screening where you might represent that lipid as just a bead of 10 or 12 atoms and then you can create almost lego like building blocks and simulate things on even larger scales in this case milliseconds and really identify what lipids are binding towards parts of this protein and correlate that with experiments and this turned out to work remarkably well which is somewhat amazing given the simplicity the final example from my group in this round is that this is Christian Wenberg who was a PhD student at the time working with colleagues determining cryo-e images of the horny layer of skin and what they were able to do is they want to study the formation process of skin as lipids are dried out with less and less and less water and during this process they move from this inverted hexagonal phases to essentially become stacks you see you you can almost see that you see that you've started to form individual bilayers of lipids here it's going to be too slow for me to show the entire process here but this is yet another process that worked remarkably well to at least qualitatively explain with molecular simulations and see what happens I'll come back to that in a second