The Futile Cycle

Perhaps many of you have seen the xkcd comic for today, called “The Difference.” Now see PZ Meyers fill in the missing bit on the comic.

I’m slowly making my way through my journal Table of Content RSS feeds. I suppose I should make this weekly habit, so that I can keep up without being swamped. I’m all caught up in PLoS Computational Biology, PLoS Biology, Nature, Cell, and Science, for example, but I still have PNAS and Molecular Systems Biology to get through, not to mention the quantitative biology section of arXiv. Mostly, I’m just reading abstracts. Every once in a while, I find a paper I find interesting enough to skim, and sometimes they’re so interesting (and short enough!) that I read them. But I haven’t read a full paper in a long, long time.

I’ve been trying out different workflows to see what sticks for reading the literature, and actually I’ve found a great program, Papers, for reading and keeping all my PDFs organized. I used to use a workflow that centered around BibDesk, downloading PDFs, importing the citation into BibDesk and linking it to the PDF, and so on. It wasn’t the worst thing in the world; actually, the best part was being able to tag papers and group them into various subjects. The metadata handling with BibDesk is pretty good.

Now, though, I think I’m slowly converting to Papers (still in Beta), which is integrated with Pubmed to allow one to download citations and organize PDFs. Papers resembles iTunes and NetNewsWire, in a way, which I guess is a complement. It’s still quite buggy and very beta, though; it’s sometimes lost a few of my PDFs, crashed a couple of times, froze a few times, and so on. Overall, the workflow is much simpler here than it was with BibDesk. There’s less “stuff” to deal with, subjectively, in terms of the user interface. It’s definitely a lot less painful to organize my PDFs now. If the next few releases increase the stability significantly and iron out some of the weirdness with losing PDFs (and maybe some user interface cleanup), I’d be quite willing to shell out money for it.

A wonderful pair of articles on archenemies and archnemeses: a blog post from Cosmic Variance on the value of having an academic archnemesis, the rules thereof and an article on how to tell the difference between an archenemy and an archnemesis. Wonderful stuff!

As always in a society largely governed by utilitarian concerns, triaging is a huge aspect of policy and the distribution of goods by the government. We have limited abilities to provide, and mostly, whenever a government program provides, the problems it addresses grow to meet the budget of that program.

Thus, Peter Schuck and Richard Zeckhauser (via Greg Mankiw) write about triaging and getting rid of “bad apples” that tend to ruin good programs for others, whether it’s the constant disruptor in housing and schools, the chronic freeloader in welfare, or even those negligent patients who don’t take care of themselves properly. I definitely think that triaging health care is something that needs to happen. I don’t really think that universal health care will be that feasible, and even if some minimal system of that sort does occur, it won’t cover all goods and services. There will still be triaging, regardless of how much the government spends on medicine or other welfare programs, because costs always increase to meet and exceed the budget of that program (and as the budget grows bigger, a larger percentage goes to government waste, I think).

Triaging sounds distasteful, but I really do think that it will help enormously, especially in health care. The health care spending distribution, for example, is extremely skewed and concentrated: the top 5% of patients (in terms of the cost of their most health care expenses) spend almost 50% of the national health care money. The bottom 50% of patients spend 3% of the health care. A more detailed table and some interesting statistics appears here. I’m not saying that the distribution in and of itself is bad—it’s sort of inevitable, I think, particularly with the massive voting power of the elderly and just the way health care is structured—but the presence of such a skew would mean that if we focus on weeding out specifically the bad apples that appear at the very top of the distribution, that we could free up a substantial amount of money to help a larger number of people.

Two interesting papers this week, one on the cyanobacterial circadian clock in PLoS Biology and the other on a global map of histones in yeast (subscription to Nature required).

The circadian clock paper is pretty neat, in that they figure out how the cyanobacterial circadian clock could be so simple, a mixture of three different proteins with ATP. It even works and oscillates with a 24 hour cycle in a test tube. They really get into the meat of the clock, and it gets pretty complex, even for “just” three proteins, to the point that they bring in a mathematician to model their ideas and verify what they suspect. The math is nothing that remarkable, but I find circadian oscillator mechanism that they finally work out to be pretty neat.

The global map of histones is quite awesome. The authors take the yeast nucleus, ChIP out the H2A.Z nucleosome-wrapped DNA (H2A.Z is traditionally a signal of the promoter and/or transcribed region), and then sequence every fragment of DNA in parallel, getting a map of H2A.Z nucleotide positioning in the yeast genome at a much higher resolution than any ChIP-chip data would give. Just a couple years ago, even to think of this sort of project would have been preposterous and probably denied funding, but with new advances in computer technology, robotics, and techniques like pyrosequencing, we get technologies like this massively parallel DNA fragment pyrosequencing. Just breath-taking. And take a look at the beautiful picture of DNA dinucleotide sequence correlation to position along the nucleosome:

This kind of huge scale experiment is a bit unwieldy, I guess, but it’s also incredibly cool when people pull off this sort of project.

Papers Referenced:
Mori et al., (2007) PLoS Biol., 5, e93.
Albert et al., (2007) Nature, 446, 572-576.

Voltage-change weak
Potassiums leak
Sodium flow
Very slow
Voltage jumps
Ion flow bumps
Opens the gates
Propagates

The bloggers at Effect Measure are doing a magnum opus blog miniseries on introducing math models in medicine and public health (or rather, one particular model) to the general public. It’s a very admirable and difficult thing that they’re tackling. Explaining mathematics to the general public is hard, because most people stop listening as soon as they hear the word “mathematics,” Just at the door, they stop without entering, not knowing how much they’re missing out.

It’s kind of sad, actually. We lament over the fact that some people can’t read, but we don’t always lament over people not being able to do math. In fact, for some circles, it’s a point of perverse pride. “Oh, I can’t do math” is an acceptable phrase in, say, literary circles, when in reality we should be as shocked about so-called “educated people” saying that as when any “educated” person would say “Oh, I can’t read.” To grasp the essential, math isn’t any harder than reading. At the highest level, math is quite difficult, but so is, say, reading James Joyce’s Finnegans Wake. Most of us don’t need to get anywhere near that level to do useful things. But people right now need a push to get to the mathematical ability level that one would consider appropriate for the “educated.”

Thus, Effect Measure’s magnificent foray into educating the public. They’re very good writers and teachers, and I highly suggest it to anyone who’s afraid of the math in science. So far, they have five articles in the “modeling antiviral resistance” series discussing one paper on pandemic influenza and antiviral resistance a little bit at a time (through 16 projected blog posts!), aimed to the general audience with no prior experience necessary:
I. What is a model?
II. A modeling paper
III. Introduction. What’s the paper about?
IV. The essential assumption
Sidebar: Thinking mathematically
Read the rest of this entry »

I recently read an interesting PLoS Biology article that covers the issues underlying intellectual property law as it would apply to synthetic biology, or the creation of new organisms and biological building blocks. It tries to address the balance between incentivizing products (such as drugs) by allowing for profit-making and some privatization and the need for openness and freedom to allow research and new innovation to happen.

Over the years, there have been lots of problems with the U.S. Patent system as it tries to keep up with the changing world. In this day and age, a lot of inventions are about ideas, information, and ways to organize such information. Software, algorithms, processes. The patent system wasn’t really designed to handle that kind of innovation, but rather the tangible kind, with physical products that would result.

Now, we have kind of a mess on our hands, and that’s starting to creep into biology. Can we patent the products of human genes? Is it legitimate to patent all diagnostics related to a particular gene? Derek Lowe often talks about patent issues with regards to the biotech and pharmaceutical industries, including the legitimacy of patenting recombinant genes as new chemical entities, and the finer distinctions in filing patents on genes and uses of genes.

For more on patenting genes and uses of genes, see this article from the Council for Responsible Genetics. One of the arguments against being able to patent the uses of genes is that it halts medicine via overly broad patents. Sure, using certain gene sequences to predict disease in a strict algorithm is patentable (and should be encouraged, to induce people to look for new, innovative ways of linking basic biology to medicine), but is using any natural gene for any diagnostic or research whatsoever patentable? That’s what Ariad did with NFkB, a gene that’s implicated in all sorts of biology. If the patent were actually enforced, research and drug development in immunology, inflammation, pain treatment, and so many other things would probably be crippled for the next 10 years. There needs to be some sort of balance, but where that should be struck is a hard question.

I’m always a big fan of intriguing ways to hire people, from the classic Microsoft puzzle interview to Google’s 10-digit prime ad to this ad for typography (h/t to Daring Fireball). I also love that company’s advertising work.

I’m reading an interesting paper from Alan Perelson, which he wrote in 1975 on network thermodynamics, particularly as they apply to biology. It’s interesting to see that it took about 25 years for the rest of the world to catch up to his ideas to create the fields of systems and synthetic biology:

“Let me end by mentioning an area in which I foresee future growth of network thermodynamics. The goal of biophysics is to understand how biological systems work. Traditionally we have approached this problem through reductionist analyses. However, when we have isolated every enzyme and catalogued every reaction that occurs in a cell will we understand how the system works? I think not, for there are complex dynamic interactions that impart to matter the property that we call life. However, if we can design and synthesize systems which have these dynamic characteristics we will have made significant progress towards understanding them. Engineers have enormous experience in synthesis and design, and it is my hope that through network thermodynamics, these techniques can be applied to synthesize chemical networks with prescribed behaviors.”

These days, these ideas are being embraced by systems biologists, who study the interactions of groups of molecules in the formation and behavior of biological systems, and with synthetic biologists, who synthesize new biological organisms with new behaviors. The insight from the past is quite amazing. It was only in the year 2000 that Michael Elowitz and Stanislas Leibler managed to synthesize their repressilator in bacteria, far after Alan Perelson described such goals in his paper.

Papers referenced:
Alan Perelson (1975). Network Thermodynamics: An Overview. Biophys J 15, 667-85