Friday, February 25, 2005

Specialization sometimes considered harmful

Bill Tozier has an interesting post about, well, a lot of things, but at least in part about the importance of sometimes trading off depth for breadth. At least, that's one of the things I got out of it. Read it for yourself and judge.

It made me think of a couple of related things: first, the sections about "problem solvers vrs problem creators" and the skills of problem creators in Michael Nielsen's post about the principles of effective research.

Second, the fact that the people I learned the most from during my time at Microsoft really were the "problem creators" and generalists -- they were the people who didn't necessarily know all the ins and outs of a specific technology, but knew something much more important: they knew enough about a particular topic, and other topics, to ask the right questions. These were the questions that exposed an angle that I'd never thought about before or a hole in my logic that would never have occurred to me [or only much later, at the most inopportune time].

Here's the catch, though: I think you need to acquire deep expertise in one [or more] areas in order to really be an effective generalist who can ask the insightful questions and sniff out the anomalies that will prove to be fruitful areas to probe into further. You need the skills, scars and skepticism you acquire in the course of learning all the ins and outs of something, whether that's becoming an expert in process X in species Y, or whether it's building a real-world software system. You need some specialization.

Personally, I've always been torn between being a specialist and a generalist. On the one hand, I enjoyed being The Man when it came to my part of the feature set of Microsoft Application Center 2000, knowing all the ins-and-outs of the code. On the other hand, I also liked getting to influence some of the higher-level decisions involved in building Microsoft's P2P SDK in my dev manager role.

Overall, though, I'm more of a generalist than a specialist. In keeping with that, I think I'd like to end up being a biotech VC, but I won't feel competent to do that until I've spent some time in industry developing expertise in one or more areas [
and nobody would hire me as a biotech VC until then anyway =)].

iPod iSolation

Andrew Sullivan has an opinion piece in the Times Online about the large number of people that are walking around enveloped in the soothing musical swaddling provided by an iPod. He links this to the increasing personalization of content: only reading a set of websites that agrees with your viewpoints, listening to satellite radio stations that play exactly the kind of music you want etc, and makes the argument that this is leading to people becoming more and more isolated. His main point with respect to the iPod is that by cocooning yourself in your own soundtrack, you miss out on things that you don't explicitly seek out, like overhearing funny conversations, kids laughing and miscellaneous other Hallmark card-worthy images, and that that's bad.

I agree with him that you miss out on some good stuff. However, on balance, I think that the good stuff is so infrequent that the bad stuff you're subjected to [if you're not listening to music] easily negates it. Based on the last few months, the chances of me having to listen to a bunch of teenagers who seem to be on a pure sugar-and-fat diet, loudly chanting "F!ck you, b!tch" [or something of that sort] while listening to their latest gangsta rap CD and doing pullups on the metal handholds on the train are much, much higher than hearing a funny/interesting conversation. [And, for the record, I'm actually not making up that last scenario -- it's what caused me to write my post about stereotypes].

All that said, I don't actually listen to an iPod other than while working out. The hassle of dealing with the headphone wires, finding space for the iPod in my pocket with all the other stuff that's already in it etc, while dressed in full winter gear and lugging a heavy bag, just isn't quite worth it to me. So it's not like I'm trying to justify why I listen to an iPod, it's more that I think Sullivan is living in la-la land a bit.

Thursday, February 17, 2005

Jumping on The Gates Bandwagon

No, not the Bill Gates bandwagon. I got off that a few months ago [and now realize what a cushy ride it was ;-)].

Rather, I refer to the art installation in New York that's the latest talking point for ... well, for people who talk about stuff like that. Christina and I watched a bit of a documentary on the couple that created it and were blown away by the cost: $20 million, of their own money. We both thought that was a bit excessive, a sentiment that is reinforced by the [as usual, pretty damn funny] transcript from The Daily Show that's available at Quantum of Wantum.

State of the academic union

3 weeks into the semester and the overall situation is ... manageable. I haven't been totally overwhelmed with work due to my 4 classes yet, but it's still early days. One class in particular may yet prove to be my undoing, more on that below.

Individual situation reports:

My cell biology class is great, primarily because the professor, Frank Solomon, is a really funny man. Among my favorite quotes are "[This molecule] is like happiness -- fleeting", "Here's a little shout-out to all the neurobiologists in the room" and "I haven't won a Nobel prize yet, but that's only because I'm tied for second in the department".

I like my molecular biology class -- it's neat to learn about all the "mechanistic" details of how DNA gets replicated, repaired etc. It's really reinforcing what an amazing machinery it is that keeps us all alive. My favorite analogy so far is this: DNA polymerase is the molecule that synthesizes [ie makes] DNA, by basically matching 4 different types of pegs with 4 different holes -- each peg only fits properly into one hole, and if you fit a peg into the wrong hole, you've messed up the DNA, a bad thing. Now, imagine that your DNA is a pipe that's 1 meter thick. Then, the DNA polymerase molecule in the bacterium E.coli is
- about the size of a FedEx truck
- moves at about 375 miles an hour
- puts around 1000 pegs into a hole each second
only makes a mistake [ie puts the peg into the wrong hole] once every 40 minutes

Now that's a mighty machine !

There was also a funny bit today when we were talking about a protein called "Dam methylase"; the professor kept saying "the Dam protein" and I, being juvenile, couldn't help snickering each time she said it. Didn't see anybody else laughing, though. Nothing funny about molecular biology, jawohl !

The "Fundamentals of Computational and Systems Biology" class is OK so far, not great and not bad. My favorite thing about it is that the assignments involve doing a bit of programming, in Python, so I get to do some programming and learn a new language, which is fun. The algorithms we're covering in the class are sort of interesting and good to know as a basis, but not something I can really get excited about.

"Functional Genomics" is the dragon that may yet rise up and slay me, because it relies heavily on statistics and probability, two things at which I've always sucked. Maybe that's because I never took a college-level class in it, but even in high school questions like "If you have 5 red balls, 4 black balls and 5 green balls in a bag and you reach in and take out 3 balls, what are the chances of getting 2 red balls and 1 green ball ?" always gave me an instant headache and the desire to go do some straightforward integration or differentiation.

In this case, the situation is made worse by the fact that one of the lecturers is, ahem, confusing, shall we say. He has a tendency to write formulas containing a bunch of symbols [each with at least one subscript or superscript] on the board, not really define what the symbols are, and then start messing around with the formula, adding a subscript here, putting a summation sign there, taking the log of all of it, while he's giving an explanation that I can't follow. The end result is that, as the German phrase goes, "Ich verstehe nur Bahnhof". The English translation of this is "The only thing I understand is 'train station'", which doesn't make any sense in English and, in an indirect, recursive kind of way, is exactly my point.

The saving grace is that his lecture notes are actually very good and the TA has been very responsive to my expressions of bewilderment, so I may make it through the class yet.

Ok, time to go flesh out an implementation of an Expectation Maximization algorithm which finds promoter motifs in DNA based on initial motif matrices and a given background distribution. [No, I'm actually not kidding. That's a question on my latest Functional Genomics problem set. Can you hear that train whistle ? I can ...]

Sunday, February 13, 2005

Wonder Twin Powers, Activate !

Last night, while Christina and I were on our way to have dinner with some friends, we saw a truly scary thing: identical twins, dressed exactly alike [purple hat with dangly bits, fluffy gold-brown coats, black leggings, reddish sneakers on disproportionately large feet], carrying the same shopping bag ... and they were women in their fifties.

Here is a sketch of them, courtesy of the court artist [aka Christina]:

[see, Christina is so good she doesn't even need a camera to take pictures ;-) Her blog shows some of the interesting pictures she's taken, together with humorous and/or educational commentary]

While dressing twin babies in the same clothes is reasonably cute, it's rather disturbing to see adults that are still dressing exactly the same. You wonder why they didn't develop separate personalities and why they're still content to be mirror images of each other. The only theory I could come up with was that they were a little "special" and lived with each other so they could take care of each other. However, that theory wasn't really supported by a bit of their conversation which I overheard: they were discussing the merits of various makers of LCD screens, with one of them making the bold statement that Sony is overrated. Somehow, not exactly a topic of conversation you expect between two women in their fifties, let alone between, well, mirror images. The speech patterns of their conversation were interesting too: they talked with the same cadence, responding at the same intervals, a bit like a metronome.

Things like this make me wish I had the social graces that would have allowed me to start talking to them and figuring out what their story was.

Friday, February 11, 2005

It must be "biology month" at the NY Times

I guess "Black History Month" wasn't keeping them busy enough ;-)

1. Last week they had an op-ed piece by a supporter of the "Intelligent Design" school of thought [I use the term "thought" advisedly -- "propaganda" might be a more accurate term]. Supporters of this theory hold that living things were "designed" [by who/what, they cannily refuse to say] and couldn't have evolved, and that hence evolution shouldn't be taught in schools, or, at most presented as "just a theory", as if there were no actual facts and experiments backing it up. There are a couple of rebuttals to this particular piece of nonsense, here and here, in addition to a website dedicated solely to critiquing ID.

*Update: Bruce Alberts, one of the scientists who Behe misquotes in his article, fires back.

2. They report on an attempt to do "open source" Biotech by a couple of Australian researchers who have come up with a system for genetically modifying plants without infringing on any of the patents held by companies like Monsanto.

3. Today, there's an op-ed piece
talking about the perils and promises of synthetic biology in this op-ed piece.

At least the last two pieces feature actual science, not religion dressed up as pseudo-science, like the first one. How and why the NY Times actually lowered itself to printing that kind of babble is beyond me. I guess occasionally you have to let the voice of the wingnuts be heard, just to keep the conversation lively ...

Thursday, February 10, 2005

Some more on the [ethical] problems of biotech

First, a couple of pointers to Slate articles about two of the most divisive issues in biotech: stem cell research and mucking with embryonic DNA: "The creepy solution to the stem-cell debate" and "The creature genetic engineers fear most". The first article especially gave me pause -- the notion of starting the growth of a human embryo but then mucking with it by turning off some genes essential for proper development so you basically end up with a totally twisted, well, "thing" of incompletely-formed organs [depending on how far you let it get] etc just seems wrong. On the flip side, arguably abortion is worse than this because you're terminating something that did have the potential for becoming a fully-formed embryo, and I support allowing abortions, so I should be OK with this too. I guess it's a question of how far you let the "broken" embryo develop ...

Second, I went to a seminar on Tuesday where Jef Boeke talked about the work his lab is doing at Johns Hopkins. They're working on something called something called "retrotransposons", which, for the purposes of this discussion, you can think of as segments of DNA that are present in large numbers in your genome and sometimes, just for the hell of it, "move" ie they cut themselves out of one position and insert themselves into another position in your genome. Now, if they happen to insert themselves into a functioning gene, that gene may get disrupted; if that disrupted gene happens to be in one of your egg/sperm cells, it could get passed on to your children who might then suffer the ill effects of having that gene disrupted. So, clearly, having these retrotransposons moving around a lot is a Bad Thing.

What Dr.Boeke's lab has done is take the most abundant retrotransposon in human beings and re-engineered it to be 200 times better at moving around. This was all in the context of figuring out how the thing worked, a fine goal in and of itself, but the end result definitely makes me go "Now, at what point did you start thinking this was a good idea ?". Boeke of course presented it as a positive, in that it would allow easier genetic modification of test animals like mice etc, which is true, but my initial reaction [and a few other people's] was to question the wisdom of doing something like this.

We're heading full steam for the Brave New World.

Tuesday, February 08, 2005

Awww, look at that cute stereotype ! Quick, let's feed it !

Riding the subway over the past few months has led me inexorably to two conclusions:

1. If somebody is acting stupid on the subway, the probability that that person is black is higher than the probability that the person is any other color. In probabilistic terms,
p(black | acting stupid) > p(any other color| acting stupid)
2. If a person is black, the probability that they will act stupid is higher than if they're any other color. In other words, p(acting stupid | black ) > p(acting stupid | any other color)

- I'm not saying anything about intelligence, innate abilities etc. My statements are purely reflective of exhibited behavior. There are lots of talented individuals who are @$$holes.
- I'm not saying this out of some misguided self-hatred. I'm happy I'm black. The observations above are just facts, however inconvenient they may be.

I guess the reason this bothers me so much is that I'm concerned that that sort of behavior is too easily taken as representative of all members of the race, including me. Stereotypes are convenient shortcuts that we all use [to varying extents, but I assert that everybody uses stereotypes in one way or another], so when a stereotype that concerns me is being fed, it bugs me [other stereotypes being fed bugs me too, but they don't have the same sort of personal impact]. On the flip side, though, it's not like I ever really feel like I'm "representing my race" or I have to "act properly" to avoid bringing shame to a set of people, so I suppose I'm applying a bit of a double standard: I want other black people to not act ignorant because I think they're feeding a stereotype but I don't have the same concern about my actions.

Ok, I've got that out of my system, now it's time to go to class.

Friday, February 04, 2005

And this ... this is the best that you can do ? You're NASA ...

[Title taken from the funniest bit of dialogue in "Armageddon"]

One area that I've been meaning to read a bit about so I can form my own opinion is bioethics, which deals with questions like "Should we be allowed to create embryonic stem cells for research ?", "Is it ever acceptable to clone a human being ?" etc. The problem with all the books I've found so far is that they seem to fall into two camps: one camp's attitude is basically "If somebody wants to graft a man's head onto a lion's body and give him the wings of a condor, why not ? Biotechnology will free us from all our limitations !" whereas the other side seems to think medical knowledge should have stopped at penicillin. There's very little middle ground, which is where I think I fall.

In any case, this article by the director of Penn's Center for Bioethics, no less, seems like a good showcase for the weakness of some of the arguments used in these debates. It centers on the recent case of the 66-year old Romanian woman who had a child and argues that that's too old to have a child. His argument, stripped down to its essentials, is: her baby was born prematurely via a C-section, it had a stillborn twin, she lost one fetus early on in her pregnancy, and [the capstone of the argument] when her child is in high school, she'll be 80. Hence, she should not have had a baby. That's a pretty weak chain of reasoning to base his conclusion on, given that all the issues he listed can and do happen regardless of the mother's age. And saying "She'll be 80 when her child is in high school and hence this should end the argument" isn't even an argument, but just falls into the "some random guy's opinion" category.

So while I agree with him that I feel like 66 is too old to have a child, I can't really subscribe to his chain of reasoning for why that's the case. And while I don't necessarily have a much better argument, I would think the director of a well-known bioethics center, who has presumably thought about this sort of stuff a lot, should be able to do better, especially in a public forum. If this level of debate persists, we're going to have a hard time making any progress on the knotty issues that biotechnology brings up.

Tuesday, February 01, 2005

First day of spring semester

Had the first lectures of 3 of the 4 classes I'm signed up for, and, at a first glance, they all look pretty good. If my fourth class ends up also looking good, I'll have a bit of a dilemma, namely whether to absorb the pain of taking 4 classes while also trying to get something done in the lab or having to pick a class to drop.

The demographics in my molecular biology class were different from anything I've seen so far: women outnumbered men 3:1 or 4:1, and 50-60% of the women were Asian. Even taking into account something I seem to remember reading about the current trend being that more women than men are going into biology [and cheerfully ignoring such statistical issues as whether this time around the class just happens to have a demographic significantly skewed from the usual distribution], that's quite a departure from the common norm of male-dominated science/engineering classes. Weird.

And another entry from the "Only at MIT" category: the subway car I was in had a poster advertising the magazine "Science", and made reference to Thomas Edison. Somebody had scrawled "Thief of Tesla's work !" beside the bit about Edison, a reference to Nikolai Tesla, a physics genius who invented the alternating current (AC) system of electricity distribution [and lots of other stuff]. Apparently, Tesla and Edison had a bit of a falling-out over Tesla's redesign of Edison's direct current (DC) generator, as well as the fact that Tesla's AC equipment was technically better than the DC stuff Edison was pushing. Now, I ask you, where in the world would somebody actually know this somewhat obscure fact and be moved to call attention to it by scrawling it on a poster ? Only at a place deeply committed to geekery [like MIT]; while a student from Hah-vahd [which lies on the same subway line] could technically also have done it, that possibility seems much less probable than it being an MIT student. [I'm disregarding all the other people who ride the subway because, by a super-scientific process of elimination based on no evidence whatsoever, I just don't think they're serious candidates =)].

More ruminations on biology

After a six-week hiatus, classes start again today. I'm registered for 4 classes but I expect that I'll have to drop one, or change to just being a listener, at least if I want to sleep more than 4 hours a night. I'm actually pretty happy that classes are starting again; I've been kind of in limbo the past few weeks since I couldn't really get started on doing research because I knew I'd have to stop spending any real time on it once classes start. So I've been reading a lot, but there's only so much reading you can do.

One interesting book that I'm just about to finish is "The Eighth Day of Creation: Makers of the Revolution in Biology", by H Judson. It's a very-well written account of the key experiments and people involved in the major discoveries in biology starting from the early 1900's till mid-1975. One of the things I like most about it is that it gives a lot of insight into all the blind alleys that people went down before hitting on the correct interpretation that unified a bunch of seemingly contradictory data, showing once again that science doesn't proceed in a straight line but rather via a series of zigs and zags. It also gives a sense of the personalities of the people involved, from Francis Crick's bold theorizing [he actually hated experimental work and did very little of it, yet he was one of the most influential biologists of the century] to Max Perutz' singlemindedness which led him to work on haemoglobin for over 30 years.

One thing that has struck me in all the "history of biology" books that I've looked at is the fact that just about all of them stop their accounts in the mid-70's. What seems to have happened is that the period from 1930-1980 produced the fundamental insights into what happens at the molecular level, whereas
the "biggest" thing to have happened since 1980 is that we sequenced the genomes of a bunch of organisms. [Take a look at this timeline to see what I'm talking about] While sequencing these genomes is undoubtedly very important, what biologists have been doing for the last 20 years is working out all the implications and fine-tuning our understanding of the mechanisms that were laid out between 1930 and 1980. Basically, my impression is that they've been fact-gathering, with very few "major" new mechanisms being discovered [the only thing I can think of are the recent discoveries that RNA is a lot more "active" in various ways than previously thought, but I'm of course speaking from a position of extreme ignorance].

While fact-gathering is all very well, part of the problem with biology right now is that biologists are drowning in lots of very detailed data, and don't really have any way to cut through the thicket and get to the next level of understanding, which is how to go from all the detailed data back "up" to understanding how life functions at a higher level. I suppose that's where "systems biology", with its focus on understanding biological systems as a whole, comes in. While there are already some examples of this sort of system-level understanding, I wonder how long it'll take before it all comes together in a coherent whole, and whether we're also going to end up with a small number of [relatively] simple, yet extremely powerful mechanisms that have the same sort of global explanatory power as the ones discovered in the beginning of molecular biology.