Large Hardon Collider

[Shiroikage]

Well, structural biology has about three distinct parts and each of these parts have a ton of different components. I'll just say outright that when I say biomolecules, I'm more or less talking about proteins (even though nucleic acids structures can be studied too) because that's where my interests lie.

One is the simple elucidation of structures. You can do this by any number of methods, but only some actually yield an actual physical representation of structure. For example, X-ray crystallography is probably the most common method for direct elucidation of structure. In it, you crystallize a biomolecule, then bombard the crystals with X-rays. Then you can determine the structure by studying the diffraction pattern. But without going into detail , not all biomolecules can be crystallized and since biomolecules are often dynamic, you can't get all the information you need from crystallography. That's why there's a whole host of other techniques such as protein NMR, FRET, nanopore analysis, and circular dichroism.

The second aspect is discovering what determines the structure of biomolecules. A big part of this is that if you know this, you can predict the structure of unknown proteins based on the their sequence. For example, sequence homology can be indicative of a common domain between the two proteins. Truth be told, this is probably the aspect that I'm least interested in. It's largely a bioinformatics problem and I'm a biochemist (with a dash of biophysics tossed in).

The last aspect is the relationship between the structure and the function of biomolecules. A protein's physical structure and sequence directly dictate its function. For example (and I'm not talking about any protein in general), a protein has a hydrophobic interior (or core) and polar and charged amino acids are usually located on the exterior of the protein. However, an enzyme might have a hydrophobic pocket located on some part of its exterior. There might also be a mobile loop on some other part of the enzyme. The substrates of the enzyme might enter the hydrophobic pocket, the loop might come over top of the binding pocket, the reaction occurs, and then presto-chango we have a product. By mutating amino acids in either the pocket or the loop and examining how these change affect the structure and how the changes affect the activity, you can determine the how structure affects function in this enzyme.

[Chun Liroy]

So my teacher is either going to hate my paper or love it. 6 pages in this format is like the threshold where I feel like I'm not expressing my understanding of the subject, but in order to adequately do that it would take like 10 pages ~.~ Whatever, I'm liking the research that Shiro helped (read: did for) me with.

[Shiroikage]

Awesome.
Did you do it on a single type of vector or did you do an overview of all of them?

[Chun Liroy]

I did all three, 2 lengthy paragraphs on each only :/ I would have done it on only 1 after reading the instructions but I had already finished one and a half of the vector overviews and honestly didn't feel like having to start over already after finish 3 pages. I'm technically not done, finishing up AAV then have to do conclusion and abstract.

[Woozie]

The research I'm currently doing in the physics department is related to protein structures.

I hate it.

I want to go back to carbon nanotubes.

Edit: Well, I didn't really like the carbon nanotubes that much either tbh. They're not going to find any research I like unless we either start doing theoretical physics or they put me on a project where I can solve things analytically instead of writing programs.

Edit 2: And besides, computational physics and protein structures don't really mix. We would need a ridiculously powerful computer in order accurately model two real proteins interacting and folding for more than just a fraction of a second. I don't know if a very accurate model over a decent time period has ever been done, even with supercomputers.

Edit 3: Actually, you would know more about this than I. How useful is computational methods for protein modeling? Do actual biologists find use in these? Has this ever been used to make predictions or anything like that? Is this useful at all for modeling the actual protein interactions within an actual living lifeform? Or is it really as useless as I think it is?

[Shiroikage]

I took a class on protein structure and engineering this term. I'm inclined to agree with you.

Docking of small molecules with proteins certainly is effective at this time, but we don't have a complete enough understanding of protein-protein interactions at the moment or the kind of computational superpower necessary to produce a very accurate model of protein-protein interactions. Hell, the energy terms in docking are simplified (from what I recall) to speed up simulation for drug design.

[Tristam]

The research I'm currently doing in the physics department is related to protein structures.

I hate it.

I want to go back to carbon nanotubes.

Edit: Well, I didn't really like the carbon nanotubes that much either tbh. They're not going to find any research I like unless we either start doing theoretical physics or they put me on a project where I can solve things analytically instead of writing programs.

Edit 2: And besides, computational physics and protein structures don't really mix. We would need a ridiculously powerful computer in order accurately model two real proteins interacting and folding for more than just a fraction of a second. I don't know if a very accurate model over a decent time period has ever been done, even with supercomputers.

Edit 3: Actually, you would know more about this than I. How useful is computational methods for protein modeling? Do actual biologists find use in these? Has this ever been used to make predictions or anything like that? Is this useful at all for modeling the actual protein interactions within an actual living lifeform? Or is it really as useless as I think it is?

Late to the last couple pages, but computational methods for predicting protein structure are getting better. A fellow grad student used structure prediction to identify a putative domain in a protein he works on, and validated the prediction using site directed mutagenesis, and did great biochemistry on the membrane-binding properties of the protein.

He used the PHYRE protein fold recognition server, but I don't know how the "threading" algorithm or whatever works. But I'm not a structure guy, even though I'm on one of his papers lol.

edit: I should have mentioned, the membrane binding properties of the protein were entirely dependent on the domain identified by the structure prediction. Sorry it's friday and my brain is all over the place.

[Kuya]

Why are protein-protein interactions so complicated? With that in mind, what is the folding of protein? I don't even have the faintest idea about those things even though i've heard the term tossed around so often.

[Chun Liroy]

I'll let Shiro answer that Kuya since I'd probably give a highly simplified response whereas he can get very in depth for you

On a different note, would modeling proteins (What Woozie's talking about) fall under bioinformatics jurisdiction? I plan on minoring in that, but I really don't know too much about it other than that it uses computer models to predict biological processes, or something.

[Tristam]

And I guess, since it was asked, I should be a bit more specific about what I do.

I do single molecule studies of motor protein biochemistry. I fucking love it. But I'm starting a postdoc where I'm learning about cell biological studies of the protein superfamily that I study. I'm still doing some single molecule stuff, but my main focus is motor proteins in the context of the brain and how they might be involved in the formation of memories (specifically, motor neuron spine formation). I fucking love my postdoc too, even though biology is damn complicated and for every experiment I think of I can think of two reasons why it won't work or prove inconclusive.

For the gene patenting thing, I'm pretty much against it. Patenting tests or detection methods is all well and good, but since most basic gene information is derived at the level of basic (mostly academic) research, if everyone were to patent every potentially interesting gene they discovered, medical science would go nowhere.

Also, to shiro, I'm a big fan of structural biology, picking apart how a protein works is so much easier and makes so much more sense when you have a structure to work on. Some of my favorite journal clubs are when someone presents a structure paper and we all dig into it.

[Kuya]

And I guess, since it was asked, I should be a bit more specific about what I do.

I do single molecule studies of motor protein biochemistry. I fucking love it. But I'm starting a postdoc where I'm learning about cell biological studies of the protein superfamily that I study. I'm still doing some single molecule stuff, but my main focus is motor proteins in the context of the brain and how they might be involved in the formation of memories (specifically, motor neuron spine formation). I fucking love my postdoc too, even though biology is damn complicated and for every experiment I think of I can think of two reasons why it won't work or prove inconclusive.

For the gene patenting thing, I'm pretty much against it. Patenting tests or detection methods is all well and good, but since most basic gene information is derived at the level of basic (mostly academic) research, if everyone were to patent every potentially interesting gene they discovered, medical science would go nowhere.

Also, to shiro, I'm a big fan of structural biology, picking apart how a protein works is so much easier and makes so much more sense when you have a structure to work on. Some of my favorite journal clubs are when someone presents a structure paper and we all dig into it.

Would you mind going into that? Part of my studies is psychology, so this might be interesting.

[Woozie]

I'll let Shiro answer that Kuya since I'd probably give a highly simplified response whereas he can get very in depth for you

On a different note, would modeling proteins (What Woozie's talking about) fall under bioinformatics jurisdiction? I plan on minoring in that, but I really don't know too much about it other than that it uses computer models to predict biological processes, or something.

That's the word my computer science professor calls it. She works closely with my boss, so my boss had me to go her to get some help on it. Turns out she has her PhD in math but does all her research in bioinformatics. She was able to recommend me some books and programs and stuff to help with my research, and even tried to get my a bioinformatics scholarship. But since I don't plan to do this as a career, I didn't want to join that scholarship program.

She also tried to get me into an REU for bioinformatics. But I choose to go to a different one, where I'll be working with superconductors and josephson junctions (still computational. I'll never get into experimental physics, ever).

As for Kuya's question, I'll let someone else explain it, because I'm sure I'm not the best one to explain protein folding. Based on my research, protein folding is the process of spitting out infinities, collapsing into a singularity, and crashing my computer after wasting 5~15 hours of my life. After the protein folds, it then wastes another 10 hours of my life through debugging.

If you want more detail on protein folding, I can post my error messages. Or you can crash your own computer to get a feel for how proteins apparently work.

See, this is why I hate computational physics.

[Tristam]

Why are protein-protein interactions so complicated? With that in mind, what is the folding of protein? I don't even have the faintest idea about those things even though i've heard the term tossed around so often.

Basically, when a protein is made by cellular machinery (trying not to get technical), the amino acids come out one-by-one, in a chain, but the structure of the protein determines how it will work, and what parts of the protein are exposed to interact with other things. The way a protein folds is governed almost entirely by interactions between the amino acids of the protein itself, mainly hydrophobics, hydrogen bonding, and electrostatics. Hydrophobic amino acids don't like to be exposed to water, so they tend to group together in certain places, usually towards the center away from the solvent which is mostly water. Some amino acids have a net + or - charge, so they will attract each other. As these interactions begin to form, the structure of the protein begins to form through intermediates, until the entire final structure resolves. Proteins fold in a surprisingly reliable manner, and a huge field is dedicated to it, so any explanation I'm giving is horribly oversimplified and relies on my macromolecular biophysics course from 7 years ago lol. For example, there are chaperone proteins that assist other proteins to make the proper fold, and most misfolded proteins are targeted for degradation by protein eating machinery.

The final structure is usually a beautiful piece of work, and you can make predictions from it, for instance, if there is a "basic patch" of many amino acids close together that have + charges close to the surface of a protein, you might suspect that is important for some sort of protein-protein interaction.

Drug design, like Shiro mentioned, is becoming more rational because most companies have devoted great amounts of money developing small molecule inhibitors against proteins. When you know the structure of a protein, and where the most likely sites for interactions or other "biological business" are, then you can begin to rationally design molecules that "target" your protein target of interest more specifically. Shiro probably knows more about Structure-activity-relationships (I think that's what SAR is right?) than I, but I loosely remember some talks about it, and it made alot of sense to me.

[Woozie]

What's funny is how little I actually know about proteins despite modeling them for a living right now (for almost a year, I think. I can't remember exactly when I started because for a few months I was doing that and carbon nanotubes at the same time). It's the same thing with the carbon nanotubes. I spent two and a half years on those and I probably don't know anything more about them than anyone else here. I don't really even know more about them now than I did before I started the project.

[Tristam]

Would you mind going into that? Part of my studies is psychology, so this might be interesting.

Sadly, I wish my research was more related to the formation of episodic or other types of memories normally associated with the hippocampus. Every time I read a paper about hippocampal neurons and wonder how I can apply that knowledge, a more senior member of the lab basically says "Purkinje neurons don't work that way" (I study Purkinje neurons).

So for most neurons, changes in the strength of a signal between two neurons is associated with a change in a thing called a dendritic spine, which is the part of a synapse that recieves the signal. The spine looks like a little button, and it is close to a projection of another cell that looks like a catchers mit. The communication between these two things is the fundamental method of signal transport between neurons. For hippocampal neurons, if a signal is repeated (like say you're actively recalling a memory over and over) that spine gets loads of signals and gets bigger by recruiting a number of different proteins. So the repetition leads to a change in spine volume, which increases the strength and efficiency of the signal, which is part of what's called long term potentiation (LTP), which is proposed to strengthen the memory. Alternatively, if the signal stops coming, the opposite process occurs (called long term depression, LTD).

Now, for purkinje neurons, it seems like LTD is the most important aspect of memory formation. Why you ask? Because a purkinje neuron develops and sprouts thousands, if not tens of thousands of dendritic spines in the complete absence of ANY signal whatsoever! So that's a bit weird, and we're trying to figure out why that happens and how. Purkinje neurons are in the cerebellum, which is a part of the brain important for motor memory (like, how you remember how to walk). But the cerebellum, iirc, is also important for some aspects of emotion. So maybe there's a link between what I study and what a psycologist might study, but it's a great big black box at the moment and I don't think very much is known about the link between emotion and basic molecular architecture.

To get a visual idea of the type of cell I'm working on, check out these images from the cell-centered database, these things are huuuuge.

http://ccdb.ucsd.edu/sand/main?stype...mit=Go&start=1

edit: for what it's worth, as we learn more about the brain and what parts are important for what functions, I do think we will begin to see cross-talk between psychologists and molecular biologists. But it's going to require some new and interesting sorts of thinking to really make progress in that regard.

[Kuya]

Sadly, I wish my research was more related to the formation of episodic or other types of memories normally associated with the hippocampus. Every time I read a paper about hippocampal neurons and wonder how I can apply that knowledge, a more senior member of the lab basically says "Purkinje neurons don't work that way" (I study Purkinje neurons).

So for most neurons, changes in the strength of a signal between two neurons is associated with a change in a thing called a dendritic spine, which is the part of a synapse that recieves the signal. The spine looks like a little button, and it is close to a projection of another cell that looks like a catchers mit. The communication between these two things is the fundamental method of signal transport between neurons. For hippocampal neurons, if a signal is repeated (like say you're actively recalling a memory over and over) that spine gets loads of signals and gets bigger by recruiting a number of different proteins. So the repetition leads to a change in spine volume, which increases the strength and efficiency of the signal, which is part of what's called long term potentiation (LTP), which is proposed to strengthen the memory. Alternatively, if the signal stops coming, the opposite process occurs (called long term depression, LTD).

Now, for purkinje neurons, it seems like LTD is the most important aspect of memory formation. Why you ask? Because a purkinje neuron develops and sprouts thousands, if not tens of thousands of dendritic spines in the complete absence of ANY signal whatsoever! So that's a bit weird, and we're trying to figure out why that happens and how. Purkinje neurons are in the cerebellum, which is a part of the brain important for motor memory (like, how you remember how to walk). But the cerebellum, iirc, is also important for some aspects of emotion. So maybe there's a link between what I study and what a psycologist might study, but it's a great big black box at the moment and I don't think very much is known about the link between emotion and basic molecular architecture.

To get a visual idea of the type of cell I'm working on, check out these images from the cell-centered database, these things are huuuuge.

http://ccdb.ucsd.edu/sand/main?stype...mit=Go&start=1

edit: for what it's worth, as we learn more about the brain and what parts are important for what functions, I do think we will begin to see cross-talk between psychologists and molecular biologists. But it's going to require some new and interesting sorts of thinking to really make progress in that regard.

Sounds counterintuitive, kind of like how visual sensory neurons do the exact opposite of what every other neuron does. Instead of reaching an action potential when stimulated, they actually hyperpolarize, but yes i agree it is a very complex thing. But yes, what you study seems to be important for understanding plasticity, and now that i remember you once mentioned plasticity to me so i got the impression that you were well acquainted with the concept.

At the moment i'm only taking an introductory course(on the physiologicl aspect of psychology), but i am a bit dissatisfied since it's only one semester for so much information, making it so you have to study for the tests rather than actually learn the stuff.

[Tristam]

At the moment i'm only taking an introductory course(on the physiologicl aspect of psychology), but i am a bit dissatisfied since it's only one semester for so much information, making it so you have to study for the tests rather than actually learn the stuff.

Being interested in the brain and such now, I wish I had taken this sort of course. Will the course look at diseases or mutations associated with the brain in terms of changes in emotion or other whole-body aspects of mental function?

I'm completely baffled by this for instance:
http://scienceblogs.com/seed/2010/04...d_racial_b.php

I mean, what is the fundamental basis for forming stereotypes. I had always assumed stereotypes were learned things, but there may be a more fundamental basis for such a thing?

[Kuya]

Yes, the course in my university covers diseases and mutations and the associated problems in behaviour and cognition. For example, recently we studied agnosia and its relation to brain lesions. Studying the brain in relation to cognition and behaviour is a very fascinating field and i kind of wish i could go and study that instead after i finish my BA.

As for stereotypes, i learned in social psychology that it's a relatively "automatic" phenomenon, and that it seems to happen naturally in and out of groups, without of course, ignoring the political aspects of stereotypes.

Speaking of biological psychology, how do you guys think thoughts are formed exactly? Or how the brain produces what we call ourselves and our experiences? I find this question the most fascinating. For example, for many aspects our sensotions are hypothesized to manifest via volley theory and place coding. In other words, how the frequency and the groups of neurons that fire equal a specific sensory experience. In terms of thought it seems much more complicated, and it seems to be the product of various groups of neurons handling various aspects of cognition firing in specific patterns, and this somehow equals our thoughts.

Also fascinating to me is how our brain may determine our behaviour and thought, but our thought and behaviour also seems to change the brain.

[Kuya]

Actually, i have an essay on neuroplasticity to write which i'm not writing because.... because.

[Chun Liroy]

Well, I'm done came to exactly 6 pages in length. I cannot thank you enough for the help Shiro, and I shall definitely let you know if I get a good grade 3 minutes to spare too, true Quincy Fashion. Anyways, time to fold clothes and play some CoD.