Tuesday, June 3, 2014

Folding@home Arrives at Wilkes University!

In the Lucent Lab, we are proud to announce the arrival of our own Folding@home server! This means that as part of the Folding@home consortium, we will be able to use this amazing resource in collaboration with The Pande Group at Stanford University to explore in vivo protein folding, protein - drug docking, and enzyme design. But what is Folding@home and why should you care that our lab at Wilkes is involved? What do we plan on doing and how will our plans benefit Wilkes University as well as science as a whole?  Well, read on to learn more!

What is Protein Folding?

Proteins are the main functional components in all living organisms. They are the things that digest your food, make your muscles move, tell your body when it is sick, and perform nearly every other task your body needs to survive. In order for proteins to do their job however, they need to fold up into the correct shape. Here's an example: have you ever tried to make an origami swan that flaps its wings? If you make a mistake during the folding process, you may end up with a swan that doesn't flap, or something that isn't even much like a swan at all. Your proteins work in a similar fashion. Right after they are assembled inside your cells, they fold up into their functional shape.  Sometimes proteins mess up when they fold (we call this misfolding). When this happens, a number of terrible diseases can result including Alzheimer's disease, Parkinson's disease, cystic fibrosis, ALS, and even some forms of cancer.
 Your DNA is simply a set of instructions for how to make a protein.  The process of going from DNA to protein is called the "Central Dogma of Molecular Biology"
So what makes a protein "know" how to fold up correctly? Although we know the gross features about protein folding (i.e the hydrophobic effect, formation of secondary structure, etc.) we still don't understand the process well enough to be able to predict exactly how it will happen for an arbitrary protein or when it will mess up. What steps are involved and in what order do these steps have to occur for a protein to fold? How long do these steps take? Which ones are more important? These are hard questions to answer in the laboratory, mostly because it is very difficult to see objects as tiny as proteins while they are actually folding. We can see what they look like when they are folded, and we can figure out some features of the folding process (like how long it takes) but we can't actually see most of the details. One way to solve the problem is to use computers. If we combine what we know about protein structure with our knowledge of physics, we should be able to simulate the protein folding process. It turns out that this works really well... if we have a very very powerful computer to let loose on the problem.

What is Folding@home?

Folding@home is a distributed computing network specifically designed to study the essential biological process of protein folding. Rather than using a centrally located super-computer, the Folding@home network uses more than 250,000 CPUs around the world, donated by individuals who want to contribute to the study of protein folding. This represents approximately 10 petaflops of computing power, making this one of the most powerful computing resources in the world.

 Map showing location of Folding@home donors world-wide.
How does this project work? It works kind of like a screen saver. Each person who wants to donate time on their computer downloads a client. When their computer is idle, the client will use this time to perform protein-folding simulations. When the donor starts using their computer again, the simulation stops. This way, individuals are able to donate spare computer resources to perform simulations.

 Folding@home graphical client
This network represents an incredibly powerful resource for studying this type of problem. Why is this method so effective?  Well, unlike the processors of a super computer, the donor computers do not need to talk to each other, they just need to talk to the servers that distribute work (in the form of inputs for protein folding simulation). We call this type of problem pleasingly parallel.

Why does this work? It turns out that certain biophysical problems are pleasingly parallel. To understand what I mean by this, consider the following simple analogy. Imagine playing the lottery with a 1 in 365 chance of winning. You can play the lottery every day for a year and by the end of the year, you will likely have won the lottery at least once. Alternatively, you can go to the store one day and buy 365 tickets. Doing it this way, you will have a high chance of winning the lottery in a single day. Protein folding is similar to this (in statistical physics we call this sort of process ergodic). By giving out little chunks of protein folding simulation to donors around the world, we increase our chances of seeing a protein fold.

Professor Vijay Pande at Stanford University developed the Folding@home network. His group, as well as the research groups of his students currently maintain this network. The Folding@home project has led to more than 100 peer reviewed publications and is internationally famous both for its colossal size and power, as well as the high impact it has made towards our understanding of protein folding.

What about the Lucent Lab and Wilkes University?

In the Lucent Lab we are interested in protein folding in vivo. This means that we are interested in simulating the things that make protein folding different inside a test tube (where we study proteins in the laboratory) compared to inside a cell (the place where proteins are actually doing their jobs).  This involves studying things such as molecular crowding, chaperones (proteins that help other proteins to fold), and the proteasome (the cellular garbage disposal that tries to grind up proteins when they misbehave).

We are also interested in other processes that are related to protein folding including enzyme design and protein-drug interactions.  Both of these processes involve studying ways to figure out if a small molecule will stick to a protein.  In the case of enzyme design, we are trying to figure out how to change or mutate the protein so that it will speed up a given chemical reaction.  In the case of protein-drug interactions, we are trying to figure out which molecules from a large set will bind to a given protein.  For enzyme design, we hope to use our findings to design enzymes that can be used to do useful things like cleaning up chemical waste or creating biofuels.  In the case of protein-drug interactions, we hope to use our techniques to predict compounds that will be useful for curing illness.

By hosting a Folding@home server at Wilkes University, we will have the ability to use the Folding@home network in an attempt to solve these pressing biological problems. Furthermore, students performing research in the Lucent Lab will have the opportunity to contribute to this amazing project and its noble goals.

Thursday, May 29, 2014

What is "spin"?

Hello Everybody! This is the first of what I hope will be a number of short little entries explaining some interesting things in the world of physics, biology and chemistry. One thing I'd like to discuss is the quantum mechanical property known as spin. Spin is an interesting little property that is not commonly understood outside the world of physics and chemistry, but is responsible for an enormous number of the things we observe in the world around us. Spin is responsible for how chemical bonds can and cannot form between atoms (the Pauli principle and Hund's rules) and is also one of the major quantities that gives rise to the phenomenon of magnetism (meaning it is responsible for why medical imaging techniques like MRI work). But what is spin really?

Spin is kind of a poor name, and the main purpose of this article is to explain why. When we think of spin, what comes to mind is an object that is rotating about some axis (like a bicycle wheel or a planet). But does this have anything to do with the physical quantity we call spin? First of all, let's start out by saying that spin is a property of subatomic particles. These are things like protons, neutrons, and electrons that make up atoms. Let's talk specifically about electrons. These are very small and very light (they are about 2000 times lighter than protons) and they have a negative charge. This charge bit is important. It turns out that this is actually one way to create a magnetic field. An electric current (i.e. a moving charge) gives rise to a magnetic field. This is known as the Biot-Savart Law. Now, let's imagine that our electron is like a little planet spinning around its axis. This little ball of spinning charge would create a magnetic field. And it is this little magnetic field that causes magnets to work (and allows a doctor to give you an MRI).
 Figure 1: if an electron is a spinning ball of charge it would create a magnetic field.
The reason this property is called spin is because it is a type of angular momentum. We know from quantum mechanics that spin (angular momentum) can only take on specific values called quanta. As a matter of fact, this is where quantum mechanics gets its name: when objects are really small their physical properties (such as energy, momentum, or position) can only have certain allowed values. In the case of the spin angular momentum of an electron this can be $\pm\hbar/2$ ($\hbar$ is a special number in quantum mechanics and its value is $1.05457\times10^{-34}m^{2}kg/s$; it has an interesting history and a profound meaning but we won't talk about that now). For arguments sake we can say that $+\hbar/2$ refers to an electron spinning counter-clockwise and $-\hbar/2$ to one spinning clockwise.

Now let me pose a question: if you were somehow sitting on the surface of the electron (which has an angular momentum of $\hbar/2$) how fast would you be moving? To figure this out we need to know what the radius of the electron is. Modern particle physics does not give an electron a radius or a shape (it is essentially a point particle), but we can use classical physics and relativity to set an upper bound on how big an electron could be.

If we assume that an electron is a little ball of uniform charge, and we know both its mass and charge, how big would it have to be? Well we can calculate the electric field associated with a spherical charged object using Gauss' Law and from that we can calculate the potential energy associated with that electric field. If we set this equal to the energy arising from the rest mass of an electron (you know Einstein's good old $E=mc^2$) we can figure out the approximate size of an electron.
$$U=\int{\mathbf{F}\cdot d\mathbf{r}}=\int{q\mathbf{E}\cdot d\mathbf{r}}=\frac{q^2}{4\pi\epsilon r}=mc^2$$

Don't worry if you don't understand the particulars, all this is saying is that we will assume that the energy associated with an electron's mass is equal to the energy associated with its charge. Solving this little equation for r gives us:
$$r_{c}=\frac{q^2}{4\pi\epsilon mc^2}=2.819\times10^{-15}m$$

This is known as the Compton radius for the electron (AKA the classical electron radius). This is unimaginably small (ten billionths of a millimeter!). As we said previously the spin of an electron is quantized in amounts of $\hbar/2$. If we assume that corresponds to the angular momentum of a spinning object, we can use classical physics to figure out the speed on the surface of that object. We would use the classical formula for the angular momentum of a sphere:
$$S=I\omega=\frac{2}{5}mr^2\omega=\frac{2}{5}mrv$$

Now if we plug in the values for S (spin angular momentum), m (mass of an electron), and r (Compton radius of an electron) we can solve for v:
$$v=\frac{5\hbar}{4m_{e}r}=5.135\times10^{10}m/s$$

How fast is this? It is 171 times faster than the speed of light. As you may already know, objects with a non-zero rest mass must move slower than the speed of light (more on this in a future blog post). So the "spin" of an electron cannot really correspond to it spinning around its axis!

So how should we think of spin? A better term than spin is "intrinsic angular momentum" or better yet, a particle's "intrinsic magnetic moment". We can convert spin into a magnetic moment using the simple formula:
$$\mu=g\mu_{b}S/\hbar$$

This way, we can think of particles like electrons as little bar magnets and the spin of these particles is really just their intrinsic magnetic moment. So rather than thinking about spin as an electron spinning clockwise or counter-clockwise, you can think of it as a little bar magnet pointed up or down (and in quantum mechanics we do in fact call $+\hbar/2$ as "spin up" and $-\hbar/2$ as "spin down").

 Figure 2: electrons are more like little bar magnets than little spinning balls!
I hope you liked my first little blog post. There are lots of other cool things to talk about related to spin: superfluidity and quantum computing to name just a few. But we will leave that for another time. If you have any comments or corrections please feel free to leave them below. Thanks and stay tuned for more fun science stuff!