Happy Independence Day

Wicked.

Linkdump for June 19th through June 26th

So true

distractions

Still funny after 5 years. From elephantitis of the mind

QOTD

Pujols was it. He was the Bill Brasky of baseball. He peed singles and crapped doubles. Give him a 32-ounce Louisville Slugger and a bottle of Fruit Punch Gatorade and he could have demolished old Busch Stadium by himself in under an hour. Men in St. Louis would have stepped aside proudly if Pujols wanted to have his way with their wives. But he didn’t, because he’s too classy for that.

–Joe Sports Fan

Politics vs Academic Field

From The Monkey Cage, via adaptive complexity:

Fact: Academicians tend to be politically differentiated according to discipline, with those in the social sciences and humanities on the left, those in the natural sciences in the middle, and those in engineering and business on the right.

Argument: This is no coincidence. Students’ political attitudes are being shaped by their professors.

Counterargument: Agreed. It’s not a coincidence. But the operative force is self-selection, not socialization

Having noticed this myself, my vote is that it is absolutely self-selection. It’s no coincidence that the fields with the most earning potential (business, engineering) are populated by mostly conservatives, and that those with low earning potential (art, music, history) are liberals. It’s the same reason that you don’t see many liberals on Wall Street or conservatives working in art museums. Conservatives in the US just value money and power more than liberals, who may actually shun these things as “selling out”.

The natural sciences have a mix of the two, due to the different career options. Conservatives are much more likely to want an big pharma or industry job, while liberals prefer to stay in academia and pursue basic science.

I’m not saying that this is true for all the grad students all of the time, but it certainly seems like an accurate depiction of the general trend.

Linkdump for June 11th through June 18th

New Media Fail

The funny thing is, having read all the crazed twitters today, I actually have less of an idea of what’s going on in Iran. It’s almost like you need people whose job it is to take conflicting data in a dangerous part of the world and make sense of it. We could call them reporters…

via

Again with Big Genomics

Since my last post, the question of why we need big genome centers has been bouncing around in the back of my mind. One commenter on that post said:

The reason is quality. A large center can (and in my experience does) perform much more quality analysis and produces more sequence more cheaply.

While I don’t disagree, with sequencing price dropping dramatically, couldn’t this be done just as well by one of the dozen or so commercial operations that now offer sequencing? They’re all doing business on a large scale and this have just as much ability to offer quality product. Why should the NIH continue to subsidize something when it’s being offered at reasonable rates from the corporate sector?

So no, I don’t believe that big genome centers will be any better than industry at preserving read quality. As I said before, they will still be the ones helping to get next-gen sequencing platforms off the ground, but that’s research, not providing a service. So what other advantages might they offer?

The idea that I’ve been kicking around is that they might offer a critical mass of talent in the same place. Instead of meeting twice a year at conferences, scientists work right down the hall from each other, and can thus bounce ideas off each other more quickly and easily. This probably also helps prevent duplication of effort. If you know Professor X next door is working on problem Y, you won’t be wasting your time trying to beat him to the punch. Instead, you’ll be contributing and collaborating with him.

That said, I think physical proximity and openness about research are quickly becoming less relevant, with the advent of online science. We can develop close-knit communities of researchers online and do more of our work in the open, which will actually help prevent scooping.

So how about it - what am I missing? What are your ideas for why big genomics centers can or will remain viable?

The fade of big genomics

In a recent post, David Dooling asks why genome centers are forced to release their data early, when other smaller labs with a sequencing machine aren’t. In responding to some of the comments on his post, David had this to say:

Most of what large genome sequencing centers are doing now are not the type of projects that were done five to ten years ago, i.e., not the type of projects that fall under Bermuda/Fort Lauderdale. The goals and end results of many of the projects are no different than those of single investigators.

I agree completely, and think it begs the question: What do we need the sequencing centers for, then?

For many years, the centers were essentially acting as contractors. The NIH said: We’ll give you $50M, you give us a bunch of genomes. Back then, big centralized projects were necessary, because producing reasonable amounts of sequence required a lot of infrastructure and the economies of scale. Those times are quickly coming to an end.

It’s no secret that sequencing is becoming a commodity. In just a year or two, any lab will be able to get hundreds or thousands of genomes sequenced, either in-house, or by sending DNA off to a company. The few remaining genome centers have tried to keep themselves relevant by focusing more on the science, but it’s getting to the point where most of that science could be done anywhere.

Let me be very clear that I’m not anti-genome center (in fact, I’m affiliated with one here at BCM). I think that they still have a tremendous role to play in developing new sequencing techniques and bringing them into the mainstream. They won’t fade away completely, but I think that we will see a marked reduction in the size and scope of the projects that they tackle.

Really, I think of this process as analogous to the early days of computing. Supercomputers at universities were the only game in town for a long time. They played their part and did amazing work, but the cost-reduction and decentralization of the computing infrastructure was ultimately a good thing. We’re seeing the same thing happen in biology, and we’ll all be better off because of it.

So it begins.

hot and humid

Oh yeah, and a temp of 97 equals a heat index of about 107.

Should be a fun few months.

Linkdump for May 29th through June 8th

The future is now

After working out a problem on the whiteboard, I snapped a picture of it with the tiny camera in my phone, used bluetooth to sync the picture to a laptop without any cables, then used wireless internet to send it to my always-accessible email which resides on a computing cluster in another part of the world.

I’m awfully glad I live in the future.

Linkdump for May 21st through May 27th

Train Talk

The Infrastructurist and Michael Dukakis talk trains:

It’s a chicken and egg thing. If you build first-class rail-based public transportation systems, over time you’ll get the kind of cities you want. Boston was in bad shape in the 60s when we were having this ten year debate over the so-called master highway plan. We were doing the same thing as everyone else, building highways and inviting people to leave town. Well, today Boston is arguably the most successful city in the country. Why? Because we stopped building highways. Admittedly we had a subway in place. But it was in terrible shape. This isn’t complicated: fixed rail produces dense, walkable cities. Highways produce sprawl. Take your pick.

Update: Part II of the conversation is now online

Powerpoints and Pedagogy

An instructor recently asked a question about teaching on AskMetafilter that hit a number of nerves with me.

In short, they posted that their students are dissatisfied with any course where lecture slides, done in powerpoint, don’t get posted online. Then they really touched a nerve, by saying this:

Also, we have tried not doing PPT. They were NOT happy about it”

You know what? Tough shit.

At the beginning of my undergrad, I loved easy classes like that. Who didn’t? All the info was on the slides and it was all posted on the internet. You could skip class every day, pull the powerpoints off the web, then show up and regurgitate info for the exam, all without ever learning anything. Now that I’m a few years out and have some perspective:

Screw that.

Instructors need to resist the urge to give the students what they want. Instead, give them what they need. The best lecturers I ever had all used chalk and a chalkboard.

Now, there’s nothing inherently wrong with powerpoint, but in the hands of an already poor presenter, it’s murder. For pete’s sake, don’t fill the slides with all the information you’re lecturing on. Throw up a picture and talk about it for 5 or ten minutes. Remember, your slides are a visual aid, there to assist your presentation. The slides shouldn’t stand alone, because they’re missing the key component: You.

You are the presenter, and you need to present knowledge in a way that is stimulating and accessible. If students can learn everything they need to know from your powerpoints or the textbook, you’ve marginalized yourself. Then you get surprised when your students don’t respect you or your class?

Look, if someone misses a lecture, they SHOULD be missing important information. If they’re sick, they can talk to their classmates, or come to your office hours and get a copy of the notes. Your job is not to pamper these young adults, it’s to teach them.

Air bubbles

Have you ever left a glass of water sitting out overnight, then found little air bubbles all over the sides the next morning? Yeah, so did this guy, and he wanted to know why. Since I spent ten minutes answering him, I figured I’d repost part of the answer here:

To put it simply, solubility is the amount of solute that a solvent can hold in solution (got all that?). It might be easier for you if you think about stirring sugar into water. You add a little, it stirs right in and all the sugar dissolves. If you keep adding more and more, the water eventually won’t be able to hold any more, and the sugar will start precipitating out to the bottom of the glass.

Water also dissolves other things, like gases. In addition to trace amounts of minerals, some of the nitrogen and oxygen from our air is dissolved into the water.

Now here’s the important part: The solubility of water isn’t constant, and is affected by temperature and pressure. When these things change, the amount of stuff that your glass of water can hold changes.

Hot water will be able to dissolve more sugar than cold water. This fact is often demonstrated by making rock candy in grade school science lab. Heat water up to boiling, and stir as much sugar in as you can. Then, stick a popsicle stick in the solution and let the water slowly cool. Over a day or so, crystals will start to form.

What happened? Well, a few sugar molecules stick to rough spots on the stick. These are called sites of nucleation. A few more molecules of sugar will start to stick to those, then a few more, and soon visible crystals will start to form. All this happens because the solubility of the water has decreased as it cools. So the water can’t really hold all the sugar anymore, and the sugar desperately wants to get out of solution.

The same principle is at work with your glass and air bubbles. There are dissolved gases in the water, and imperfections in the glass serve as nucleation sites. That’s why you see bubbles sticking to the edge of the glass. Over time, if the bubbles keep growing, the force compelling them to rise will overcome the forces sticking them to the side of the glass, and the bubbles rise to the top and disperse. To speed this up, you can tap the glass to help the bubbles break free from their nucleation site, and a bunch of them will rise at once.

QOTD

“The first time you do something, it’s science. The second time it’s engineering. After that, you’re just a technician.”
–Clifford Stoll

via Cass

Selfless, Cold, and Composed

One of the three standout tracks from Ben Fold’s latest release, which is a compilation of University acappella groups sining his songs. (”Magic” and “You Don’t Know Me” are the others). The rest of the album isn’t bad, it’s just not stellar either.

Linkdump for May 8th through May 12th

Linkdump for May 4th through May 8th

« Previous Entries |