Tuesday, April 25, 2006

A guide to covert communications

Advice for paranoid reporters was written (somewhat sardonically) to help reporters deal with the age of Bush, but it's a handy catalog of techniques to use when one wishes to retain a measure of confidentiality and privacy.

Frothing mad: Colonel Wilkerson

One of the way Rove, Cheney and Bush destroy their enemies is to "feminize them". They does this to women and men alike, but of course it works better on males. Rove uses words historically associated with feminine ineffectiveness, words like "shrill" or "hysterical" or "histrionic". He and his minions ascribed these traits to many democrats in the last election, including Paul Krugman of the New York Times.

When someone uses this kind of effective grade school tactic, one response is to invert the pejorative. "Shrill? Hell, yes I'm shrill. You should be too." Shrillblog headlines those of us who are justly shrill.

Even with its new respectability, however, the word "shrill" seems a bit understated for Colonel Wilkerson's incendiary rage:
Colonel Wilkerson Is Really Shrill!

From the Kyoto accords to the International Criminal Court, from torture and cruel and unusual treatment of prisoners to rendition of innocent civilians, from illegal domestic surveillance to lies about leaking, from energy ineptitude to denial of global warming, from cherry-picking intelligence to appointing a martinet and a tyrant to run the Defense Department, the Bush administration, in the name of fighting terrorism, has put America on the radical path to ruin.

Big science: Regulating the master regulator genes

When I was a child of 19 or so, I helped write Fortran code to simulate the braking system of an ancient freight train. These pneumatic systems were not entirely designed, they evolved over time. A large amount of lost expertise was used to create a kludged system that would reliably activate at roughly the same time over a long train despite large signal latencies. Analysis showed that some of the pneumatic subsystems merely counteracted others, but the emergent behavior was reliable.

I don't recall, at the time, realizing that I was working on a metaphor that was widely applicable to biology, economics, software, and all other complex evolved systems. I think of those lessons again as I read about the very reliable and fantastically arcane bio-nano-machinery that controls the master regulator genes:
Studies Find Elusive Key to Cell Fate in Embryo - New York Times

A question of interest for biologists studying cell identity is what regulates the master regulator genes. The answer has long been assumed to lie in the chromatin, which determines which genes are accessible to the cell and which are excluded. The chromatin consists essentially of millions of miniature protein spools around each of which the DNA strand is looped some one and half times.

The spools, however, are not mere packaging. They can lock up the DNA they are carrying so that it is inaccessible.

Or they can unwind a little, so that the strand becomes accessible to the transcription factors seeking to copy a gene on the DNA and generate the protein it specifies.

... there are protein complexes — essentially sophisticated cellular machines — that travel along the chromosome and mark the spools with chemical tags placed at various sites on the spool.

A complex known as polycomb ... tags spools at a site called K27.

This is a signal for another set of proteins to make the spools wrap DNA tight and keep it inaccessible.

Another complex tags spools at their K4 site, which has the opposite effect of making them loosen their hold on the DNA.

The chromosomes of the body's mature cells are known to have long stretches of K27-tagged spools, where genes are off limits, and other regions where the spools are tagged on K4, allowing the cell to activate the local genes.

... In the current issue of Cell, a team led by Bradley E. Bernstein and Eric S. Lander reports that they looked at the chromatin covering the regions where the master regulator genes are sited.

They found to their surprise that these stretches of chromatin carried both kinds of tags, as if the underlying genes were being simultaneously silenced and readied for action.

... Each cell must avoid being committed to any particular fate for the time being, so all its master regulator genes must be repressed by tight winding of the spools that hold their DNA. But the cell must be ready at any moment to activate one specific master regulator as soon as its fate is determined.

The Broad team then looked at the chromatin state of the master regulator genes in several kinds of mature cell.

... they found that the bivalent domains had resolved into carrying just one type of mark, mostly the K27 tag, indicating the master genes there were permanently repressed.

But in each kind of mature cell one or more of the domains had switched over to carrying just the K4 tags, within which genes would be active.

... Dr. Bernstein's team worked with mouse cells, but its findings have been confirmed in human embryonic stem cells by Tong Ihn Lee and Richard A. Young of the Whitehead Institute...

The new findings raise the question of how the embryonic cell knows where on its chromosomes the bivalent domains should be established. Dr. Bernstein and Dr. Lander believe that the answer lies in the structure of the DNA itself.

The bivalent domains occur at regions on the chromosome where some of the DNA sequence is highly conserved ...

... These particular sequences, however, do not contain genes, so must be conserved for some other reason.

The highly conserved non-gene sequences were first detected in the dog genome, which was decoded last year. It was in trying to figure out what these regions did that the Broad team stumbled across the bivalent domains.

Although only half of the highly conserved regions contain master regulator genes, something in their DNA structure may be the signal that tells the cell where to create the bivalent domains.

... Dr. Young's team has studied another aspect of embryonic stem cells which ties into the new finding about bivalent domains. Three genes, known as oct4, sox2 and nanog, are known to be particularly active in the cells and are regarded as a hallmark of the embryonic state.

Dr. Young showed last year that the genes make transcription factors that act on each other's control sites in ways that in effect form a circuitry for controlling the master regulator genes.

He has now found that these transcription factors bind at many of the bivalent domains created by the polycomb complex.

... a working definition of cell status may be almost at hand, in Dr. Lander's view, in terms of a cell's chromatin state and the transcription factors that can bind to its available genes.

So, how many of these researchers will end up making the trip to Stockholm? There are a lot of moving parts that are coming together here. The Nobel committee will have its hands full sorting things out. I do think it's cool that one of the fundamental breakthroughs appears to have been an unanticipated side-effect of decoding the dog genome.

Fame aside, note the shape of things. The "master signal" that says "control me here" is the shape of the DNA. In other words, this control signal is topological. Signaling by shape is fundamental to DNA and protein alike, in the world of the cell topology is about creating a distinctive signature of electromagnetic fields (and perhaps quantum signals too). On the other hand, there is a binary system of "wrap" and "expose" that controls what DNA is read. On the third hand there's set of 3 genes for transcription factors that regulate each other's activity -- a configuration that will be familiar to anyone who's studied simple transistors.

If one abstracts the control systems as shape and charge, simple circuits and binary actions, it becomes possible to see how such an emergently complex nano-world could evolve a little bit of a time. I am still waiting, however, for the announcement that biologists have uncovered the DNA equivalent of the LZW-compression algorithm. [Note to software designers -- if you're looking for new compression algorithms, look how DNA solves this puzzle.]

This is huge news. The biggest thing I've read for a while.

Monday, April 24, 2006

Missed comments: Gmail eating them!

GMail has been eating (filtering as spam) Blogger's comment notifications. So if you're commented and the comments didn't get approved for posting, the blame goes to GMail's atrocious spam filtering. It appears to be untrainable -- I've several times marked these as 'non-spam'.

I think I've found all the comments and approved them.

Revolt against Rumsfeld: It's the torture.

This week's Economist features a Lexington column baying for Rumsfeld's head on a pike. I think this is the "new" Lexington, since the prior Lexington was a GOP poodle this is a bit of a change. There's no denying attacks on Rumsfeld are also attacks on Cheney and Bush.

Salon has more to say. They claim the root of the general's revolt is not so much in Rumsfeld's strategic failures and tactical blunders, but rather in his execution of Cheney's torture program:
More top brass blast Rumsfeld | Salon.com News

..."I sense a great deal of distress among senior military officers over what's happened with prisoner treatment," Irvine said. "I believe the abuse is playing a significant part in how these generals are feeling and why they're speaking out. There's an understanding that whatever we're doing at Guantánamo and elsewhere constitutes license for others to do to us when our soldiers are taken prisoner in the future. There's the realization that we've pretty much trashed the high ground along with the Geneva Conventions."

On April 14, Salon revealed that Rumsfeld was personally involved in directing the harsh interrogation of a prisoner at Guantánamo Bay, according to a sworn statement by an Army lieutenant general who investigated prisoner abuse at the U.S. base in Cuba...
The thesis is that emotional fire beneath the rebellion is a hatred for the dishonor Rumsfeld has brought upon America and the American military.

Honor. I figured Americans had forgotten what it meant.

Sunday, April 23, 2006

Space Race. Brain Race.

I have to give FuturePundit credit here. I've long assumed that economic competition within the US will overwhelm any resistance to messing with IQ genes, but FP points out that competition will likely be nationalistic as well.

Obvious in retrospect; I'm sure it's shown up in a science fiction story somewhere. China is the obvious suspect to start things off, but the US is about as good a candidate. One or the other will make altering IQ genes a national agenda, and the race will take off.

Practically speaking nationalist competition only speeds things up a bit compared to internal economic competition. Either way the outcome will be a lot of people with very high IQ, and a lot of people suffering from the expected and unexpected consequences of this kind of gene manipulation.

It will be an interesting problem, but it's manageable. The much bigger problem will be artificial minds, I suspect the merely linear enhancements of gene-engineering will be a historical footnote; assuming there's a history.

Saturday, April 22, 2006

If it's not mini-B, it's no good

My Canon Digital Rebel XT and my Canon s410 both use mini-B USB connectors. The Motorola RAZR and the Blackberry use mini-B connectors. On the other hand, the Palm Treos use yet another of Palm's proprietary cables.

As a general rule (the iPod being the noteable exception), one could do worse than base all buying decisions on the use of the mini-B USB connector for data transactions.

The evolutionary biology of reading - and of sign language

Forget all this creationist blather about how hard it is to make a retina. The much more interesting challenge to evolution is that humans can read. Of course, as we'll see below, it's probably not going to be that much of a challenge ...

The evolutionary history of reading has interested since my undergraduate neuropsychobiology class in 1981. Translating visual input into perceived sound, or even directly into the mysterious connections that are the basis of thought and what-we-call-consciousness -- it's tough to imagine where that came from. It seems like a kludge of the first order, something that ties together very disaparate systems. Hardly surprising then that a significant portion of humanity are unable to read well, even with above average IQs.

My best guess was that it had something to do with sign language, and that maybe we "signed" before we could speak well, so signing language and spoken language evolved together. It was natural then to "read", reusing visual parsing and mapping subsystems that evolved contemporaneously with spoken language. [There's nothing original here btw. All those people who trained chimps to do sign language in the 1970s must have been thinking along the same lines.]

So given my interest, it's neat to read that new progress has been made in sorting out how we read (at least how we read phonetic languages!). Emphases mine:
Science & Technology at Scientific American.com: Controversial Theory Linking Reading Ability to Specific Brain Region Gets a Boost
April 20, 2006
Controversial Theory Linking Reading Ability to Specific Brain Region Gets a Boost

More than a century ago, a French neurologist suggested that a specific region of the brain processes the visual images of words. Without it, he postulated, people cannot read except by laboriously recognizing letter after letter, rather than whole words. Yet humans have only been able to read for several thousand years--perhaps not enough time for such a trait to evolve, some scientists have argued. New research, however, supports the idea that reading does rely on a localized set of neurons.

Previous imaging studies using functional magnetic resonance imaging (fMRI) or positron emission tomography (PET) showed that a small region buried deep in the left rear of the brain lit up with activity when subjects read, or recognized words, as opposed to perceiving other objects, such as faces or tools. Victims of stroke with damage in this region often reported reading difficulty. But because stroke damage in these patients was never confined to this region alone and imaging studies can only demonstrate correlation, not causation, controversy persisted.

Neurologist Laurent Cohen of the Hopital de la Salpetriere and his colleagues received a rare opportunity to explore this hole in scientific understanding when a 46-year-old epileptic came to them for treatment. His chronic seizures indicated that a small portion of his brain--roughly contiguous with the so-called visual word-form area--should be removed.

Prior to removing the damaged section, the scientists performed a series of tests on the man, including a wide array of reading challenges and the temporary placement of electrodes in his brain. He proved normal in all regards, including his ability to quickly recognize words no matter how many letters they contained.

But two weeks after the operation, though cured of his epilepsy, the patient complained of difficulty reading and tests showed that his ability to comprehend longer words had slowed by half. Even six months later, he needed roughly an additional 100 milliseconds for each additional letter to recognize a word.

The finding seems to support the contention that this region of the brain is critical to reading, but it does not answer questions as to how it developed. "One possibility is that the [visual word-form area] performs a visual processing function that predisposed it to being co-opted for reading," Alex Martin of the National Institute of Mental Health writes in a commentary accompanying the paper in today's Neuron. Nevertheless, the French team has provided more evidence that this region is critical to your ability to read this article.
This sure sounds like it supports the sign language theory. There's no way this evolved in the past 1000 years; Native Americans can learn to read and they only started about two hundred years ago. It had to be a subsystem that evolved long ago for a different purpose.

The theory that sign language and spoken language co-evolved is stronger than ever ...
































More than a century ago, a French neurologist suggested that a specific region of the brain processes the visual images of words. Without it, he postulated, people cannot read except by laboriously recognizing letter after letter, rather than whole words. Yet humans have only been able to read for several thousand years--perhaps not enough time for such a trait to evolve, some scientists have argued. New research, however, supports the idea that reading doe"

Friday, April 21, 2006

Iranian bomb: Krauthamer and the cost of havoc problem.

The March 31st issue of Time Magazine included an essay by Charles Krauthamer. The essay is now behind a paywall, but a blog search found this excerpt (from a neoconservative-type web site). Emphases mine.
Clear and Present:

Today Tehran, Tomorrow the World

... We're now at the dawn of an era in which an extreme and fanatical religious ideology, undeterred by the usual calculations of prudence and self-preservation, is wielding state power and will soon be wielding nuclear power.

We have difficulty understanding the mentality of Iran's newest rulers. Then again, we don't understand the mentality of the men who flew into the World Trade Center or the mobs in Damascus and Tehran who chant 'Death to America'—and Denmark(!)—and embrace the glory and romance of martyrdom.

This atavistic love of blood and death and, indeed, self-immolation in the name of God may not be new—medieval Europe had an abundance of millennial Christian sects—but until now it has never had the means to carry out its apocalyptic ends.

That is why Iran's arriving at the threshold of nuclear weaponry is such a signal historical moment. It is not just that its President says crazy things about the Holocaust. It is that he is a fervent believer in the imminent reappearance of the 12th Imam, Shi'ism's version of the Messiah. President Mahmoud Ahmadinejad has been reported as saying in official meetings that the end of history is only two or three years away. He reportedly told an associate that on the podium of the General Assembly last September, he felt a halo around him and for 'those 27 or 28 minutes, the leaders of the world did not blink ... as if a hand was holding them there and it opened their eyes to receive' his message. He believes that the Islamic revolution's raison d'etre is to prepare the way for the messianic redemption, which in his eschatology is preceded by worldwide upheaval and chaos. How better to light the fuse for eternal bliss than with a nuclear flame?
Krauthamer starts the article by quoting Richard Feynman. The rumbling you hear is my old professor (ok, so I went to two of his Physics-X lectures ...) approaching a relativistic spin rate. Insults to the honored dead aside, let me deconstruct the essay as such:
  • Fanaticism is not new, but the falling cost of havoc means it has a new significance for the survival of humanity.
  • Iran's leader is a religious zealot who wants to bring on the end-time (elsewhere in the same issue it's noted that he now has strong support among the young for a nuclear program, albeit perhaps not for the end-time).
  • Iran must be stopped, with a strong implication that military action will be required.
For the sake of discussion let's assume that President Ahmadinejad is indeed an apocalyptic madman. True, Saddam was accused of this as well and was found to actually quite calculating and not inclined to suicide. (Alas for us all, Saddam was no better at such calculations than were Rumsfeld/Bush/Cheney). True, Bush has been accused of this and he's probably not suicidal. No matter, let's assume it's really true of Ahmadinejad.

What can and should we do in this case? I'd say, not damned much. Bush has dug a deep hole for us. Thanks to Bush Iran is stronger than ever, and US actions in Iraq have enormously strengthened Ahmadinejad's political base. Thanks to Bush the US has no credibility to push for sanctions, and no international support for any serious action. Probably the best we can do is horse trade to get China to take the lead on this, doing whatever they can do to slow things down. (Putin seems as blind and incompetent as Bush.)

Could we nuke Iran and solve the problem? Maybe their bomb would be delayed a few years, but probably not much more than that. In the meantime there'd be enormous sympathy for Iran, and in many circles there'd be support for an anonymous counter-strike against the US. Incidentally, this idea of nuking Iran is morally repugnant.

So, what can we do? We delay, retreat, appease and hope for a miracle. Voting the GOP out of the house this year, and the presidency in 2008, might help a bit.

Anyone who voted to reelect Bush is at least partly responsible for this mess. Competence matters, rationalism matters, thinking matters.

Thursday, April 20, 2006

Minnesota GOP Platform: Protect creationism

Gotta love the MN GOP. PZ Myers picked this out of their platform:
Pharyngula: Good ol' MnGOP

Protecting educators from disciplinary action for including discussion of creation science, adopting science standards that acknowledge the scientific controversies pertaining to the theory of evolution.
Discussing birth control too? I didn't think so. Only creationism gets special protection. What a bunch.

Civil liberties protection officer?

Normally it's obvious that anything the Bush administration does is bad. This is unusual in that it's not obviously in which way it's bad:
WSJ.com - New U.S. Post Aims to Guard Public's Privacy:

Alex Joel ... named as the first civil-liberties protection officer for the U.S. Office of the Director of National Intelligence...

... 'There is no silver-bullet answer,' he says of balancing privacy and national security. 'There are actually a lot of silver BBs and if you put enough of those together in a coherent way, wrap it with good policy, procedures and training, then you can have the same impact as a silver bullet.'
Experience mandates we assume this is somehow covertly bad, even if we can't figure out how it's bad. It does remind me of the several good cybersecurity appointees -- all of whom quite within months of taking office.

Still, by the standards of our government, this is good in a not-so-obviously-bad way.

Wednesday, April 19, 2006

Outsourcing the grunt work of a web article

I wonder how many workers today privately outsource tedious tasks like this:
Philip Greenspun's Weblog

... I estimated that it would take me 10 hours to assemble these data by clicking around at Amazon. It is a bit more involved than you'd think because for many of these items, Amazon requires you to 'add item to cart to see price'. Anyway, I put the project up on www.rentacoder.com and a guy from Pakistan did the job in two days for $10. He made only a couple of mistakes."
This is where huge productivity boosts can arise. Good to see a real-world example of what's long been theoretical.

A corporation I know of recently implemented a Dilbertian "educational" program on corporate security. Important topic, horrendous delivery and evaluation mechanism. Next time they do something like that employees may be tempted to outsource the process of doing the training and taking the exam ...

Neuregulin: Monster gene of madness

Neuregulin is a massive and complex gene that produces six proteins in the brain. Protein four is found to be over-expressed in some schizophrenic patients, possibly due to a point mutation in a regulatory gene.
Schizophrenia as Misstep by Giant Gene - New York Times

... Neuregulin is one of about 10 genes so far linked to schizophrenia ... It is one of the largest genes in the human genome, sprawling over some 1,125,000 units of DNA, and it generates at least six types of protein ...

... the variant stretch of DNA ... linked to schizophrenia does not lie in the neuregulin gene itself but just upstream of it, meaning that it presumably affects not the actual proteins produced by the gene but the way the gene is controlled.

... the first component of the transcript that makes the Type 4 proteins lies at the very beginning of the neuregulin gene and closest to the upstream genetic segment that is statistically linked to schizophrenia.

The researchers found that people who inherited two versions of the variant segment, one from each parent, were producing 50 percent more of neuregulin's Type 4 protein than those who inherited one or no copies.

... The role of neuregulin's Type 4 proteins is unknown, but they may be involved in making neurons migrate, a property of great importance when the brain is being constructed.

Dr. Law said that the variant segment linked to schizophrenia had a single DNA unit change at the center of one of the binding sites recognized by the transcription factors that control the gene. Loss of the binding site presumably upsets the regulation of the gene, causing too much of the Type 4 protein to be generated, she said.

Having slightly more than usual of a single protein may seem a very subtle derangement for so devastating a disease, but subtlety is to be expected, Dr. Weinberger said. "We know that all mental illnesses are about very subtle aspects of the wiring biases. They are about how you process complicated environmental stimuli, not about how you walk down the street.

It's extremely difficult to make a functioning brain. Autism, schizophrenia, mental retardation, all are felt to be related to possibly minor defects in brain development. Given how hard it is to make the brain work "out of the box", I'd bet there are a vast array of corrective processes that take place until the brain matures -- at age 25 or so in men, earlier in women. Some of the disorders we see may represent partial failures, and partial successes, of repair processes. (ie. sacrificing the mirror neuron network to support the primary neuron network ...)

Windows Vista: a disaster

Microsoft in the past few years has become the gang that can't shoot straight. For a time I thought they were faking it. Surely this ruthless bunch of cut-throat morals-be-damned competitors couldn't have become ineffective. I remember back when they ran the rags, playing PC Magazine and its ilk like violins. They were pirates then, and unfortunately successful.

Lately it's seemed that they were moving away from being pirates, and also from being effective. (Thank heaven, given the fruits of their success.) Thurott, a former Microsoft-groupie who's become a real journalist over the past few years believes things are very bad indeed.
Paul Thurrott's SuperSite for Windows: Windows Vista February 2006 CTP (Build 5308/5342) Review, Part 5: Where Vista Fails

... For Windows, specifically, the situation is dire. As I've noted in the past, the Windows Division retains, as employees of the software giant have told me, the last vestiges of the bad, old Microsoft. This is the Microsoft that ran roughshod over competitors in order to gain market share at any cost. The Microsoft that forgot about customers in its blind zeal to harm competitors....

... So what went wrong? What didn't go wrong? When Bill Gates revealed in mid-2003 that he was returning to his roots, so to speak, and spending half of his time on what was then still called Longhorn, we should have seen the warning signs. Sadly, Gates, too, is part of the Bad Microsoft, a vestige of the past who should have had the class to either formally step down from the company or at least play just an honorary role, not step up his involvement and get his hands dirty with the next Windows version. If blame is to be assessed, we must start with Gates. He has guided--or, through lack of leadership--failed to guide the development of Microsoft's most prized asset. He has driven it into the ground.

Promises were made. Excitement was generated. None of it, as it turns out, was worth a damn. From a technical standpoint, the version of Windows Vista we will receive is a sad shell of its former self, a shadow. One might still call it a major Windows release. I will, for various reasons. The kernel was rewritten. The graphics subsystem is substantially improved, if a little obviously modeled after that in Mac OS X. Heck, half of the features of Windows Vista seem to have been lifted from Apple's marketing materials.

Shame on you, Microsoft. Shame on you, but not just for not doing better. We expect you to copy Apple, just as Apple (and Linux) in its turn copies you. But we do not and should not expect to be promised the world, only to be given a warmed over copy of Mac OS X Tiger in return. Windows Vista is a disappointment. There is no way to sugarcoat that very real truth.
No great loss. When I replace my last PC with a Mac, I'll run either XP or Win2K in the virtualizer for the few apps I need, including Microsoft Access, Visio and my old version of Quicken. Every week or so I'll boot Windows, otherwise I'll do very well without. Vista is irrelevant and unwanted.

Fresh organs from China: kill to order

We've known for years that China harvested organs from condemned prisoners. I, like many others, expected this practice to be monetized. I didn't anticipate this practice, though in retrospect it's obvious:
Brad DeLong's Semi-Daily Journal: Involuntary Organ Donors

Professor Stephen Wigmore, who chairs the society's ethics committee, told the BBC that the speed of matching donors and patients, sometimes as little as a week, implied prisoners were being selected before execution.
In other words the prisoners are kept alive until their organs are needed, then they are killed. Sort of like cooking lobster.

Maybe there's still time to reconsider those Olympics? Not that the US is in any position to point fingers, Bush has reduced our reputation to the national equivalent of "psychotic pedophile".