Dead run

I am doing really badly in my QM2 class. I think that he has dropped the first exam, which I failed terribly, since everyone else also did terribly. He was a bit vague, but I'm pretty sure he dropped it.

My next exam is April 12th, and then all that's left is the final. I can't fail this one. I really need to knock this one out of the park, just in case he decides to count that first exam. (Rumor has it that my professor will let you retake the final if you need this class to graduate and you're failing, but there's no way I'm going to let it come to that.) I figure if I do nothing but study from now until the 12th, I can at least console myself that, if I do fail, I did everything I could have possibly done to pass. I need to know everything completely cold. I'm going to work and rework homework problems until I can do this stuff in my sleep.

Right. So, degenerate perturbation theory, here I come...

(...and on a happier note, my knee is feeling better, and I've finished my past three runs without incident!)

Frat boys and matrix elements

First of all, let me just say that I hate frat boys. I really do. Screw frats and the people in them. It's 11 AM and I'm at work at the SLC and there's about 20-30 frat boys here (and they ARE frat boys, you can tell just by looking at them) screaming and running around and cheering. AT THE SLC. For about 2 hours solid. It's intensely obnoxious. Anyone who would be in this area studying has to go to a different area of the building because of these imbeciles. I can't even tell them to shut up because this isn't technically a quiet area.

Anyway...

One thing that turns up over and over in QM2 is 'matrix elements'; this is because you represent operators with linear transformations, and linear transformations are written as matrices. I realized that I don't understand that well why this is, so I'm reviewing that, and using this as my notepad...

For any vector with components a1, a2,..., an where you know what a linear transformation T does to its set of basis vectors, T|a> returns a new vector |b> with components:

b1 = T11*a1 + T12*a2 + T13*a3 + ... + T1n*an
b2 = T21*a1 + T22*a2 + T23*a3 + ... + T2n*an
...
b3 = T31*a1 + T32*a2 + T33*a3 + ... + Tnn*an

And T acting on the basis vectors gives you:

T|e1> = T11|e1> + T21|e2> + ... + Tn1|en>
T|e2> = T12|e1> + T22|e2> + ... + Tn2|en>
...
T|en> = T1n|e1> + T2n|e2> + ... + Tnn|en>

These equations give you the matrix elements for T. Take the inner product of each equation with one of the basis vectors to find these; for example, to find the element T21:

<e2|T|e1> = <e2|T11|e1> + <e2|T21|e2> + ... + <e2|Tn1|en>

But the Tij's are scalars, so you can take them out of the inner products. Assuming this is an orthonormal basis, you are left with only one term on the RHS that is not zero:

<e2|T|e1> = T21<e2|e2> = T21

So this is the formula for finding the matrix elements.

C++ ftw

I've been thinking that I would like to get involved in computational biophysics research when I get to UCSF, even though my background is mostly in experimental work. The work I've been doing with Jake here at UGA is mostly computational/analytical stuff, and while the specific stuff we are doing (analysis of threshold traits in multiple environments) is not particularly interesting to me, I do like the sort of work that I've been doing. I guess I've always enjoyed working with computers, so it's not any great surprise that I like this kind of stuff.

One thing that I am planning on doing this summer is really throwing a lot of energy into re-learning C++. I knew C, and C++ to a lesser extent, fairly competently when I was in high school, so re-learning it should not be so difficult, but I think I will probably have much more success getting involved in computational research at UCSF if I show up and can demonstrate that I am a good coder. It is something that I do have a relatively strong aptitude for, but a lot of the more complex ideas and syntax fades from memory when you haven't done it for such a long time.

I actually sat down at Hot Corner tonight and spent some time reviewing simple C++ stuff, and it's coming pretty naturally to me, so that's a promising sign! I guess I shouldn't spend too much time doing this during the semester, though. I've got a little over a month left until I graduate...ALL I need to do is pass this crazy quantum mechanics 2 class and I will be done. Easier said than done, though...

Not even four-letter words

Words really can't describe how much I hate having to be at work at 8 AM.

End do

So, I ended up getting rejected from Harvard. I expected this, since their application was due back at the beginning of December, and you normally don't wait this long for an interview offer! It's a moot point anyway, though, since I've already accepted UCSF's bid.

Would have been a nice ego trip, though...heh.

So, I applied to 8 schools, one of which (the University of California, San Diego) actually returned my application fee, which I suppose means there was some undetermined problem with my application and it wasn't processed or something. Of the 7 I was considered for, I was rejected from 4 (MIT, Stanford, UW, and Harvard), and received 3 formal admission offers (from ASU, UCSB, and UCSF). And I'm going to UCSF.

Perks!

I just found out that UCSF's graduate program includes not only payment of all fees and tuition and health insurance and a $26,500 annual stipend for living expenses, but also includes a fitness center membership and they give you a free brand new Mac laptop when you begin. I mean, I'm sure this is a for-business-use-only sort of deal, and you probably have to give it back when you graduate, but...wow! How cool is that?

Hee. Excited!

I am teh uber nerd

So I'm sitting at a coffee shop downtown, reviewing some mathematics (calculus, actually), and it's spring break. Everyone else is out partying, and I'm sitting here, studying theorems about infinite series and convergence. I guess this makes me some kind of uber-nerd or something. What's more, although I am reviewing this because I need to for one my classes, I actually enjoy this stuff a lot. I like learning about mathematics. I find it satisfying, and interesting, and even fun. I think I like it more than I like, for example, drinking myself blind, which is, I suppose, sort of the normal thing to do over spring break.

It's strange that a person can go out and drink until they puke every night of spring break, and their peers will think this perfectly normal, even praiseworthy, but if a guy would prefer to sit down and study something, that's weird, and mockable.

Ah well...that's why I'm such a nerd I guess...

Degeneration...

So, it turns out that degenerate perturbation theory is, like...hard. And tedious. And my homework assignment is due tomorrow morning at 9:30, DAMMIT!!

Decision

So, I've accepted UCSF's bid. (And canceled my trip to Phoenix, which seemed like the more polite thing to do at this point...)

I'm really excited about this. Moving to San Francisco, baby! Woo!

Crawling towards a decision...

Well, the UCSB interview is over, and I'm flying back to Atlanta at noon. The interviews themselves (or 'meetings,' as they call them, for us pre-accepted visitors) went alright, although I have to say, I guess everyone here was more interesting on paper than in person, because I'm not sure what I was thinking when I decided that UCSB had cool research and I wanted to come here! This makes me doubly glad that I got into UCSF! There is some interesting stuff going on here, but nothing that really got me fired up like, well...just about everyone that I talked to at UCSF.

Also, some of the graduate students I met here seemed pretty weird. The ones vetted by the school to show us around were unfailingly polite, cheerful, and had their heads up the schools ass, of course, but the other grad students I talked to that were not part of the recruitment session were pretty sketchy. They took us out drinking last night, which was fun, and a lot of the people there were cool, but some of them were just... I mean, damn. Nails in the coffin, guys, I definitely don't want to come here if the other grad students are douchebags. Overall, while the school and the town were beautiful, I got kind of an off vibe from the place. I don't think I will come here.

I have considered simply canceling my trip to Phoenix and accepting UCSF's offer, but it's only a week from now, and I am genuinely curious to meet Gro Amdam, who is so interested in working with me that she requested that I come early and spend almost an entire day meeting with her and her lab. Her research sounds pretty interesting, but this could also potentially be the most boring thing I've done to date... But I've told the UCSF admissions people that I will let them know by March 12th (when I get back from Phoenix), so even if UCSD or Harvard make me some kind of offer at this point, it's hard to see what they could offer than UCSF isn't doing better.

Santa Barbara is incredibly beautiful, though, I have to say. The pictures of this place, nice as they are, do not do this place justice. That is the main reason why I would want to come here, to be honest, after these interviews...and it would be so nice!! But, of course, I can't make a graduate school decision that way.

So, it looks like I'm going to be throwing all my stuff into the Civic this summer and driving to San Francisco. Sounds like fun to me!

En route

So, I'm on the plane to LA, headed to the UC Santa Barbara interview. This is interesting, because since I've already been accepted to both this school and to UCSF (which I am almost certain I prefer), there is absolutely no stress associated with this, so I can just breeze through, and enjoy Santa Barbara...which is an insanely beautiful area!

I am keeping an open mind, though. People have been telling me all along that I shouldn't take the US News rankings too seriously (and if you look at their ranking methodology, it's easy to see why!), so even though UCSF is much better ranked than UCSB (or ASU), I'm still going to their interviews, just in case they absolutely blow me away and I decide that I actually want to go there, instead! After all, as far as location goes, I'd be perfectly happy with any of my three options - San Francisco, Santa Barbara, and Phoenix are all fine by me.

So, I still haven't heard back from Harvard or UC San Diego yet, and I'm not exactly holding my breath. To be honest, I'm fairly certain I'd prefer UCSF over either one of these places, and god knows I sure don't want to move to Massachusetts! San Diego's neat, because it's right on the beach, but San Francisco is better ranked, and it seems like there's more cool research going on at UCSF, anyway. Actually, that's kind of how I feel about these other schools: they're alright, and there is some cool stuff going on at each place, but UCSF has just so much cool research going on there. At this point, I am about 95% certain that I will end up accepting UCSF's offer.

Also, using this extraordinarily large laptop on an airplane really is no good. Wonder if Colin's managed to sell his tiny Vaio yet...

Yes indeed

I GOT ACCEPTED INTO UCSF!

OH SNAPPPPPPP!!!

More extreme happiness

Well, I got my rejection letter from MIT yesterday.

Also, I failed the CRAP out of my quantum mechanics test two days ago. And when I say 'failed,' I don't mean that I might have gotten below a 60% or some shit, I mean I would be super-duper stoked to even have gotten a 20%. I couldn't figure out ANYTHING on that test. I don't know what the hell I'm going to do about this class (obviously I should have taken Emag instead, but too late for that...). I NEED to pass this class, or I won't graduate.

Still crossing my fingers waiting to hear back from UCSF; they said I ought to hear back by the end of the week. Getting an acceptance there would make this whole rotten week alright again.

I hate days like today

I got my rejection letter from Stanford today.

I got my rejection letter from the University of Washington today.

I got my rejection letter for my story, Rally, from Asimov's today.

I'm working a split shift today, even though it's President's Day.

I went snowboarding yesterday, so today I hurt all over.

Also, I have a quantum mechanics test tomorrow that I am going to cram intensely for tonight, even though the material is this close to being impossible to comprehend, and hope that somehow I manage not to fail.

In summary, I would like to extend a hearty 'fuck you' to days like today!

The evolving grad school situation...

So, the UCSF interview last weekend went well. At least, I feel like I did well. I was very impressed by the school. The entire Mission Bay campus at UCSF is brand new; nothing there is over 3 years old, and the school has so much money it's just sick. UCSF is unusual in that there are actually no undergraduates there -- it's purely a graduate and medical school, and the med school is not part of the new campus, so basically it's an entire campus devoted entirely to graduate education in the biomedical sciences. That kind of focus is really cool, and I think that's a big part of why the faculty there are as energetic and passionate about their work as they are (or at least, that's how everyone came off while I was there). UCSF is ranked 5th in the country for the discipline that I am interested in (biophysics), but honestly, after seeing the place, I think this may be my first choice (even if I am accepted into Stanford or MIT, both of which are higher ranked). There's also some cool aging-related research going on there, which of course is neat. Plus the stipend is $26,500/year...not bad, while you're in school!

So, it was satisfying in the sense that I really felt like I did as well as I could have done. If I don't get in, there won't be any if-only-I-hadn't-fucked-up moments, because I interviewed as well as I could have, and prepared as much as I was planning on preparing. So I won't stress about it, whatever the answer is. And if I don't get in, then well...I have to go live on a beautiful beach in southern California for the next 4 to 6 years. There are worse fates, I think!

Also, San Francisco seemed really cool. I do think the whole 'SF-is-gay' thing is really overblown, too...there was a gay part of town, which Amanda actually drove me through just for the hell of it, and indeed...that part of town was really, really gay. But aside from that, I didn't get the impression that the city had more gay people there than any other big city. I did think most people there seemed unusually attractive -- weirdly so, in fact. I came to the conclusion that it's because everyone there is in remarkably good shape -- there are almost no overweight people anywhere in the city. I think this is because the weather there is so mild all the time, that outdoor exercise is really easy and appealing. (Not like here, where my 30 degrees Fahrenheit 5-mile runs this time of year start to get old pretty fast...) A few of us went hiking along the coastal trail on the northwest end of the peninsula, and damn, there were some cool views there! I could definitely see wanting to do walks like that regularly. And let's face it, northern California has some of the best hiking you can find. When Isabel and I visited the area back in '04, I absolutely fell in love with the outdoors around there. I'd love to visit Sequoia and Yosemite again.

I've got the UCSB interview next weekend. This is kind of an unusual situation, because I've already been accepted into the school, so I guess it's more of a recruitment than an interview! I am looking forward to visiting, though; the area is insanely beautiful, and if I don't get into UCSF this is probably going to be my fall-back school. Although who knows, maybe I'll get to UCSB and be so impressed there that I decide to choose it over UCSF...

Compatibility

I'm trying to wrap my head around the idea of compatible versus incompatible observables (in quantum mechanics, natch). If two observables are compatible, that means they can share a common set of eigenfunctions. (Does this mean they CAN share a common set, or can one be a subset, or do they have to be the same set? Hm.) If they have a common set of eigenfunctions, that means their stationary states are described by the same wavefunctions.

So one example of this is the commutation/noncommutation of the various operators related to angular momentum. For example, the square of the angular momentum commutes with all of its components: [L^2, L_x or L_y or L_z] = 0. This means that the square of the angular momentum and any of its components have stationary states described by the same wavefunctions. However, the components do not commute with each other: if i != j != k then [L_i, L_j] = i*hbar*L_k (for even permutations). So the components of the angular momentum do not have stationary states described by the same wavefunctions as the other two components.

This is a very strange idea. If you know the eigenvalue returned by the eigenequation for one component, you can't know either of the other two. Okay, mathematically, I see why that is, in terms of commutation relations...but intuitively, I find that kind of baffling.

More lack of insight later.

What is beauty?

According to the Speculist, beauty is a program:

The program is deliberately designed to be perceived as the will of the host organism (the guy who's looking) playing out - rather than a preprogrammed directive. This deliberately stealthy program takes over most of the identity and action of the host during puberty and - at least until reproduction succeeds - and in all truth, continues without any letup until the hormones and other agents driving the program are reduced in effect by the ravages of aging. By way of example, a dog probably perceives that when it chases a stick it is "choosing to chase the stick" - yet, a human would say "well, dog's are IMPELLED to chase the stick by influence of instinct"...in just such a fashion, humans believe they CHOOSE to follow the siren song of beauty (a babelicious one) - yet, the very belief of such choice prevents them from ever knowing that they are NOT in control of their actions (and thus - lives).

Fascinating stuff. I think he's right, and I find that to be oddly frightening.

Note to George Bush

"Allow the President to invade a neighboring nation whenever he shall deem it necessary to repel an invasion, and you allow him to do so whenever he may choose to say he deems it necessary for such purpose, and you allow him to make war at pleasure. Study to see if you can fix any limit to his power in this respect, after having given him so much as you propose. If to-day he should choose to say he thinks it necessary to invade Canada to prevent the British from invading us, how could you stop him? You may say to him, — 'I see no probability of the British invading us;' but he will say to you, 'Be silent: I see it, if you don't.'

"The provision of the Constitution giving the war making power to Congress was dictated, as I understand it, by the following reasons: Kings had always been involving and impoverishing their people in wars, pretending generally, if not always, that the good of the people was the object. This our convention understood to be the most oppressive of all kingly oppressions, and they resolved to so frame the Constitution that no one man should hold the power of bringing this oppression upon us. But your view destroys the whole matter, and places our President where kings have always stood," - Abraham Lincoln, in a letter to William H. Herndon, Feb. 15, 1848.

Tang paper

The point of this paper was to characterize a model for structural and dynamic properties of networks based on functional constraints. Evidently most previous research has looked at structural constraints on models instead, so this sort of approached the problem of signaling networks from a different perspective. So, the first thing they did was develop an algorithm for sampling from a maximum entropy probability distribution of functionally constrained networks (a 'functional ensemble,' for short). The point of using a maximum entropy distribution, as I understand it, is just to minimize the amount of information built into the system: i.e., not making any unnecessary assumptions about the nature of the network. So the idea is that they want to use these functional ensembles to test whether or not some dynamic/structural property of a network can be adequately explained by its function, or whether there is some additional selection at work.

Since everybody loves linear algebra so much, the authors decided that the most efficient way to store information about the state of the network is to encode each node as a boolean-value coefficient for a state vector. They're defining these as column vectors, so this looks like:

S(t) = [s1(t) s2(t) ... sN(t)]^T

Where T just means transpose (not an exponent) and the sj(t)'s are elements of {0, 1}. Then they define a connectivity matrix that describes how the nodes interact with one another, so its components are elements of the set {-1, 0} along the diagonal and {-1, 0, 1} elsewhere. So for a given initial state vector, you have a progression through a trajectory of state vectors, each transformed into the next by the same connectivity matrix. That is:

C*S(0) = S(1),
C*S(1) = S(2),

etc. So the general recursion formula would be C*S(t) = S(t+1). Also, since each row only affects its corresponding component of the new state vector for each transition (row 1 only affect component 1, for example), each row of C can be checked to satisfy the constraints imposed by the state vector trajectory independently of the other rows. This is important, because it makes iterating through all of the incoming arrow combinations for each node computationally feasible. Their sampling procedure for this model is simply: let alpha be an index for the rows of C between 1 and Z (where Z is the number of allowed incoming arrow combinations that satisfy the trajectory constraints). Pick alpha randomly, but use the same alpha for every node: this gives you a C that satisfies all required trajectory constraints.

Voigt paper

So the basic idea here was to show that they could set up a genetic circuit using the invasin gene as an 'output module' -- that is, the piece of the circuit that actually affects the environment in some way. Invasin is an E. coli protein that binds to proteins called beta-1-integrins on the surface of a number of mammalian cell types and can induce bacterial uptake by the mammalian cells. So the authors thought, well, wouldn't it be nice if you could, for example, set up a genetic circuit that caused invasin to bind selectively to tumor cells?

This was pretty painstaking, and boring to read about, but the end result was interesting. So they're obviously careful about genetic characterization of invasin, first of all. They checked whether invasin required other E. coli adhesion systems in order to bind to cell surfaces (it didn't), then showed that inv+ E. coli were able to successfully invade a few different human cancer cell lines (they examined osteosarcoma, hepatocarcinoma, and HeLa cells). The bacterial recovery rates were different for the different cancer types, possibly because of varying expression of beta-1-integrins on the different cell types.

Next they constructed a arabinose-inducible circuit for invasin. So first what they discovered is that background transcription of this circuit's promoter was actually sufficient for invasin synthesis, and to reduce this background expression, they altered the ribosome binding site of the circuit: they replaced the 5' untranslated region with randomized sequence and also randomized the first base of the first and second codons. So to figure out what they got from the resulting library, they did a positive selection for the inducible phenotype, then a negative screen to check which of this subset kept on invading without induction.

So next they linked the invasive phenotype with an anaerobic environment. Their strategy was the same as before: alter the ribosome binding site, generate a library, then do a positive selection and negative screen. Next they note that, since a cancerous environment tends to attract a whole bunch of different bacterial species to it, it would be nice if they could link the invasive phenotype with bacterial cell density. Specifically, quorum sensing, a method by which bacterial cells can communicate with each other. What they did was find an existing genetic switch (called lux), and created a plasmid where the inv gene was placed under the control of the lux promoter. They found pretty much what you'd expect -- when inv was expressed via a constitutive promoter, invasion was independent of cell density, where under the lux-promoter's control, it was detectable only at high bacterial cell densities.

So, this is actually a pretty cool idea. Follow-up work, and I think a good line for comments, would be along the lines of, well ok, so invasin gets you into tumor cells, but now what do you do? Also you could use this sort of genetic circuit to as a vector for a number of different applications aside from targeting cancer cells.