28 February, 2010


I found out last week that I've been chosen to give an Ignite talk about information processing in biological systems. I considered talking specifically about just 1 aspect of that processing, like receptor signaling, but found that this was far too involved for the format. Instead, I've decided to try to explain the entire Central Dogma of Molecular Biology to an audience of about 500 people in 5min with 20 slides that auto-advance every 15s. I've been working heavily on my slides, because each one has to be optimized for its content in this format.

And, because I'm ornery stubborn I decided to paint many of my slides by hand with watercolors. Here's Sonic Hedgehog:

This is going down this coming Thursday, March 4th, at 7p. It'll be live-streamed, so I'll post a link as soon as I have one.

15 February, 2010

Microcluster Condensation in Immunological Synapses 1

I remember reading one of the popular science magazines I subscribed to back in high school, probably either Popular Science or Discovery, and coming across a brief article on the discovery of the immunological synapse. It included pretty pictures. I was intrigued by the spatial and sequential alignment of disparate signaling effectors, even though I knew absolutely nothing about the context at the time. Now, I know more about the molecules and pathways involved in dendritic cell::T-cell signaling than I am able to concisely put down here in words without rolling myself up in a cloak of jargon, and even so I barely know anything (with comparison to both the experts in the field and the scale of the unresolved questions). The immunological synapse is fascinating, and to me it is beautiful in its absolute parsimony (that's a whole other post for later).

In beginning the activation of the adaptive immune system, dendritic cells process and present sampled antigen in distinct molecules (MHC) that T-cells can recognize (via the TCR and CD4/8). Due to chunk recombination of V, D, and J regions of the TCR binding motifs and subsequent pre-programmed random mutagenesis* there is extremely high heterogeneity in the recognition cognates of the TCRs. So as dendritic cells (DCs) crawl through the thymus, lymph node, spleen, or other, they have many distinct antigens loaded into their surface display molecules, and every once in a while a TCR that has some binding affinity for that antigen will bind. What follows is the immunological synapse.

The immunological synapse starts out with the binding of the TCR and CD4/8 to the MHC, which nucleate the formation of the central supramolecular activation complex (cSMAC), when all of the TCR/MHC complexes from microclusters and merge into 1 more stable site. A lot of other things happen downstream of that, most of which are very interesting**, but what I find intriguing about this is: what sort of topology do the kinetics of microcluster condensation add up to?

All optimized networks have some sort of topology. This means the hierarchy of one node over another, because to have all nodes processing the same exact bandwidth is rather energetically inefficient. As such, there can be strictly hierarchical topologies like those found inside human corporations with management, there can be scale-free topologies in which hierarchy arises due to through-put optimization and is not strict (a good example of this is the server structure of the Internet), or others I don't know anything about yet.

Is microcluster condensation hierarchical or scale-free?

Unfortunately, it is extremely difficult to answer that question empirically because it happens so fast and because the cells involved are rather camera-shy unless given very exacting and munificent conditions. Therefore, this is more or less a thought experiment.

I posit that microcluster condensation is both hierarchical and scale-free, in turn; first, hierarchical and then, later, scale-free. Cook your noodle on that for a bit, and I'll explain my reasoning for why within the next few days.

*Not a contradiction of terms. The immune system allows for random mutagenesis of a restricted set of amino acids residues on T-cell receptors and B-cell receptors to greatly increase the range of possible binding motifs without great additional informational storage costs (DNA).
**E.g., I find the activation of such factors as NFAT, mTOR, et al to be interesting, but generally find the dynamics of histone deacetlyation to be rather dull.

13 February, 2010

Smokey Honey Chili

3kg ground beef
1 medium yellow onion, coarsely chopped
2 smallish yams, cut into ~5x10x7mm chunks
2 handfuls carrots, peeled and chopped like the yams above
1 large can tomatoes
1 can kidney beans
1 can black beans
4 cloves garlic, crushed
75ml whiskey
pinch lapsang souchong
25ml honey

1) Combine onion, garlic, yams, and carrots in large, deep saucepot.
2) Add generous dash of oil, salt, and black pepper.
3) Heat on high until onions are sweaty and garlic is nutty, then add whiskey and 10ml honey.
4) Stir in, continue stirring on high heat until most of alcohol smell is gone (we're after the colloidal flavors of the whiskey to caramelize them in suspension, not the ethanol itself).
5) Stir in ground beef until it is barely done.
6) Add in undrained can of diced tomatoes and tomato paste, stir.
7) Add drained cans of beans, stir.
8) Stir in remaining honey and lapsang souchong*.
9) Add generous cayenne pepper, ancho chili powder and Tabasco, moderate amount more of salt, black pepper, and file powder, and small amount of cloves and cardamom.
10) Reduce heat and let simmer, uncovered, until the protein leaking out of the beans begins to thicken the sauce to your desired consistency.

This was absolutely delicious, and it made enough to feed me for 2-3d. It was at once spicy and savory, but not sweet as one might expect from the ingredients because the low-weight sugars from the honey, cardamom and whiskey caramelized and added depth to the smoke of the lapsang souchong, ancho chili powder and cayenne. There was a faint hint of a mellow, orange sweet echo on the top of the palette, but this was largely obscured by the very well-complexed clay and blue flavors below it.

I set out in inventing this spice combination to get the mild tones of traditional chili while using sweet potatoes as the base. I succeeded on both accounts, I believe, and I really do recommend giving this one a try. I know I will be making it again sometime soon.

*Lapsang souchong is a Chinese tea that's much better at cooking than being drunk. It's a black tea that is dry-roasted over a fire of dry pine needles, and a hot cup of it tastes like literally drinking smoke. So I just added a very small amount this time to get the smokiness without resorting to liquid smoke. The tea leaves steep in the tomato juice, and they're completely edible.

08 February, 2010


In those restless gaps in the middle of the night*, washed in the faint glow of my screen, I make music. I produce a lot of bad music that should never be heard by others' ears, but some of it turns out to be fairly decent. When all I had was an analog bass guitar and guitar, I was frequently frustrated by my inability to record and multitrack, but once I procured a virtual synthesizer suite, I was suddenly freed of this constraint. The immediate consequence was a lot of noise and I produced such tracks as "Sneeze Bucket Hat", "Los Pantalones Brillante" and "Pomegranite Ballet". But time progressed and my skills improved and I eventually found that I have 3+ hours worth of music that I think is good enough that I don't mind other people hearing it. This music tends to fall into 2 categories: crunchy, fast electronica smashing subgenres from 8-bit to breakcore jazz together and a smoother music of mostly classical instrumentation running over hip hop/industrial types of backgrounds. I could probably pull 3 full-length albums out of everything, but it is only recently that I finally decided to make an album out of the latter genre.

Toaster Sunshine - "Arc"
1. Oxidative Periphery
2. A Retrospective Abnegation of the Postmodernist Solipism
3. Swiftly Daftly Boom Boom Doom
4. Density Earache
5. Dishware Flossery
6. Additivity (Damitol)
7. Suburban Dirge
8. Erosion
9. Thermal Counterpart
10. Thunder Melted Your Words
11. Läpikuultava Syndan**
12. Aftermath
13. The Nineteen Shades of Concrete Unbeknownst to Sunlight
14. Saccharine Nocturne
15. Somewhere the Magic Happy

This is just a working title, and all the songs are in rough draft form. I think "Retrospective Abnegation of the Postmodernist" needs some more slices mixed in to make it more interesting (it was composed entirely out of about a dozen samples from Shostakovich and Bartoks' string quartets), "Suburban Dirge"'s violins need EQ tweaking, and I have noticed that there is a high pitched percussive artifact in "Thermal Counterpart" that needs to be found and dealt with. Also, something weird happened with "Saccharine Nocturne" upon conversion to an .mp3 that I'll have to track down.

Anyway, if you'd like to listen to the raw tracks in all their muddy glory, you can download a zipped version of the above album at: http://drop.io/ji4trez#. If you do so, please, by all means, criticize it as bluntly as you'd like.

*Notwithstanding the background music I tend to produce whenever there's a guitar handy, despite whatever else is going on around me.
**This translates to 'translucent heart'.

03 February, 2010

Passionate N00bity

I was somewhat disappointed when I first started attending non-class seminars for my projects in the lab and my interests to find that there wasn't much rambunctious cheering or passionate cheering. Sure, everyone is very polite and asks Very Intelligent Questions and I understand that much of this is due to an inherent respect for the speaker's effort in gathering and interpreting their data and an abstract need for civility. However, if we're doing science because we love the intellectual challenge, why don't we get more worked up about it and express that excitement? I know that science is not the same thing as sports, but why can we not get up the same passionate lather as sports fans in our arenas of data and theoretical abstraction?

Part of this is due to my dream of one day delivering a data talk in the form of a power ballad, but it's also due to confusion as to when I may break down the barrier of formality and ask a douchey question within the cloak of excited admiration? I recently saw a speaker present some really awesome new biotechnology she had developed in the course of her project and its applications. The tangibility of the data (real-time fluorescent tracking of cells in a living matrice by 2-photon microscopy; which basically meant actually seeing immunotyped immune cells crawling around and interacting with each other) was surreal and exhilarating, but when she talked about her methodologies I was left wanting. I don't want to out the speaker, so I cannot say exactly what my beef was, but it had to do with the algorithms by which she chose to quantify her data. I refrained from asking this question in the Q&A session because of the very civility I mention above, it just seemed like too rude a question to ask in such a polite intellectual subculture as an academic presentation. Of course, on the other hand, there was also the possibility that my question would have revealed my ignorant n00bity, and to be honest that also played a part in me not asking.

Nonetheless, when can we whup it up and get excited about the science all around us? Can I ever high-five or fist-bump a speaker when they've shown me something awesome or would doing so break down this fifth wall academic politeness has surrounded us with?