Philosophy of Genetics

Personalized medicine

According to a recent article in Time, the number of people who must take a drug to avoid a single adverse event (e.g. a heart attack) is called the "number needed to treat" (NNT).  For statins, the NNT is around 50 people.  In other words, for every heart attack avoided, 49 people were taking statins to no benefit, with all the potential side-effects (and cost) that entails.

Pharmacogenetics and personalized medicine are trying to change that by rationalizing who is treated.  One way to do that is by creating genetic tests, to see who responds better to certain drugs.  If the NNT were reduced to 1, that would mean everyone would receive only the drugs they needed. 

But it would also mean a 50-fold decrease in drug sales, which is why pharmaceutical companies are trying to downplay personalized medicine (or, rather, to use it to predict drug safety, but not efficacy).  Saving healthcare costs is probably what Senator Barack Obama had in mind last year, when he introduced the "Genomics and Personalized Medicine Act of 2006" that would "improve access to ... genetic tests by all populations".  That bill hasn't passed, yet, but it could revolutionize the way healthcare is administered, especially if Obama is elected president in 2008.

What is the meaning of life (at a molecular level)?

What is the meaning of life at the level of individual molecules?  At what point does an individual molecule attain a purpose that allows it to contribute to living beings?

My favorite example of a “purposeful molecule” is a catalyst.  A catalyst simply helps two reactant chemicals bond more quickly, by reducing the barriers to their pairing, like a matchmaker who gets a couple married more quickly than if they were left alone.  After all is done, the catalyst is left unchanged by the reaction.

Because there are around 1,670,000,000,000,000,000,000 molecules in every drop of water, you can see that there is plenty of opportunity for reactant chemicals to bond on their own (just by accidentally bumping into each other), if they are already predisposed to do so.  The predisposition exists when the product of a reaction is more stable (having less energy) than the individual reactants themselves (i.e. if the married couple is happier together than as separate individuals). 

However, adding a pinch of the right catalyst greatly increases the rate of a chemical reaction, by reducing the barriers for the formation of a bond between the reactants.  Life relies heavily on such catalysts in the body, called “enzymes”, which are made from proteins as specified by our genes.

Still, a catalyst molecule by itself has no meaning, especially when we enclose it in a small box without potential reactants.  Or, more precisely, the catalyst has no meaning in that place and time.  The catalyst is simply a three-dimensional form, without a purpose.

Only when the catalyst exists in the same box as a sufficient number of reactant chemicals does its purpose become clear to an observer.  Its purpose is then defined by the increased rate of chemical reaction, a purpose which is necessary (but not sufficient) for life.  You might say the purpose of the catalyst "emerges" in that context -- or “is discovered” -- where previously there was no purpose.  But unless an external observer makes a record of the event, the catalyst's purpose is lost once the context -- the spacetime box -- goes away.

Brain genes

The brain is a three-dimensional form, weighing about 3 pounds in adults (less than a pound in newborns).   Each of the 100 billion cells (called "neurons") in our brain connects with thousands of other neurons.  [There are also another 900 billion or so "supporting cells" in the brain, but we'll ignore those for now].  All told, there are around 1 quadrillion (10¹⁵) connections between neurons (called "synapses") in the entire brain, where activity happens to create the mind.

So the development of the brain is largely a problem of how to connect 100 billion neurons using 1 quadrillion synapses, so that the brain can operate, consciousness can emerge, signals can be processed, memories can form, and responses can be effected (free choices can be made?)

About 6,000 of our genes seem to be active only in the brain; genes (or gene-produced proteins) like Robo which induce neurons to grow in specific directions inside the skull, and others that allow them to recognize friendly neurons and cling to them (making a synapse), and then allow signals to be transmitted across those synapses.  Many specialized proteins (such as Reelin) help in the formation of synapses once two neurons find each other and “dock” together.  Reelin also helps the brain develop its characteristic six-layer structure.

Cadherins are sticky molecules that guide neurons as they migrate inside the skull, to find their permanent position.  Think of them like Spiderman climbing a building, using a sticky substance to cling and move against gravity and friction, propelling against other neurons until the right one is found with which to form a more permanent synaptic connection.

The Emx family of genes is involved in establishing the identity of certain regions in the brain.  The brain is full of specialized areas (vision, speech, planning, etc) which are set up in the course of development.  The Eph family of genes help lay out the basic topography map of the brain, by setting up a chemical gradient (like the latitude and longitude on a GPS device) which allows migrating neurons to find their home.  The Hox genes also help to establish basic layouts of the brain and body.

Since genes largely function to create proteins, I’ll use genes and proteins fairly interchangeably.  However, some genes can code for multiple proteins depending on the context, so it’s not as simple as “one gene = one protein”.  The FGF8 gene (fibroblast growth factor 8), for example, can be sliced and diced in different ways, leading to the production of different proteins (depending on context).  Those proteins are also responsible for laying out some of the gross anatomy of the brain.

Once neurons have made their specific connections together in the course of their development, those same synapses can be used not only to send signals from one neuron to other.  Synapses are not truly connections - but gaps - between neurons into which signalling chemicals are injected.  Usually, those chemicals are neurotransmitters (like serotonin) which are used to send signals from one neuron to the next across the synaptic gap.  But hormones and other compounds (like anti-depressants) in the bloodstream are also able to influence the signal of many synapses (and other receptors), at a global level.

For example, vasopressin is a hormone released by the pituitary, that can affect social behavior.  The most promiscuous male prairie voles, for example, have fewer vasopressin (V1a) receptors in a specific region of their forebrain (ventral pallidum region).  Artificially introducing more vasopressin receptors into their brain immediately makes them seek monogamous relationships.

I've often thought that the strategic placement of hormone receptors throughout the brain and body during development (like the placement of troops on a battlefield) is much more interesting than the hormone itself (or "command from the general"), which is really just a molecule (or, to follow the analogy, a single word -- "Charge!").  The signal simply triggers a response that was already planned, practiced and ready.

Other examples of "brain genes" include Pax6, important for the formation of the eye, and NMDA receptors which seem to play an important role in establishing memories when the activity of two neurons coincide closely in time.  Also, in simple organisms like the “sea slug”, entire complex behaviors can be triggered by a single hormone, such as the ELH (egg-laying hormone), due to the way receptors have been set up in advance in the synapses of its neurons.

I've described some of the genes that are active in the development of the brain, but I haven't begun to describe the variants of some of those genes.  Obviously, if two people don't have exactly the same version of the NMDA receptor, perhaps one of them would have a better ability to recall facts.  This is obviously an interesting area for future exploration under the subject of "genetic variation and social fairness".

Diabetes, obesity and genetics

An article in Nature finds that variations in 4 specific genes (TCF7L2, SLC30A8, IDE–KIF11–HHEX and EXT2–ALX4) can explain 70% of who gets diabetes.  It makes you wonder why people have those gene variants to begin with.  What purpose do they serve?

Also, newly published research in Science shows that having a variant (or two copies) of the so-called FATSO gene can often lead to obesity.

I'm wondering if the susceptibility to diabetes is similar to that of hypertension (due to a salt imbalance).  An article a couple of years ago described how variants in the CYP3A gene are linked to salt retention in the body.  Africans who live near the equator have one form of the gene, and others (living farther from the equator) have another form of the gene!  According to a press release at that time:

In the sub-Saharan African regions where humans first appeared, available salt must have been limited and quickly lost through sweat. People who were better at retaining salt may have had a significant survival advantage.

The problem is (and anyone who's stopped by McDonalds for super-size fries well knows), salt is no longer scarce in the modern world.  So people with the stronger "salt retention" version of the gene are at greater risk for hypertension these days.

Evolution is about trade-offs.  Having the genes for better salt retention in warm climates can give you hypertension in the era of fast food.  Does diabetes work the same way?

“This could change the way we look for disease genes,” [said study author Anna Di Rienzo, Associate Professor in Human Genetics]. “Historically, we have searched for mutations, altered or damaged versions of genes that cause rare disorders, like cystic fibrosis or phenylketonuria. Now, we are starting to look for common genes that may have been beneficial in an environment of scarcity, but have become harmful in a world of plenty. In the modern setting, it may often be the genes that aren’t damaged that predispose to disease, such as the ‘thrifty genes’ associated with type 2 diabetes.”

The Education of Bill Gates

Bill Gates recently told Time Magazine that "Learning is mostly about creating a context for motivation".  The Gates Foundation often provides funding for educational programs, and it's interesting to see his driving principles.

I agree that motivation is the key, since kids spend more time practicing things they're interested in ("time on task" leads to skill).  But by reading between the lines, I can see that Gates is ignoring the politically sensitive issue that some kids have a harder time learning than others, because their motivations themselves are innately different.  In other words, you can't teach the ability to be motivated, you can only exploit motivations that already exist in people, which have to come from within.  Some people have the "right ones" for a successful schooling experience, and others don't.

I think Bill Gates is proposing the creation of temporary, motivational frameworks using technology and other means.  In other words, teachers should find something (anything!) that motivates a child (and it may have to be customized for each child), and then use that motivation to keep them interested in the actual subject you want them to learn.

Anyway, this is what I imagine Bill Gates is talking about.  If a child is motivated only by sports (but hates geography), the teacher would create a sports games that teaches him geography as a biproduct.  Sounds like a nice approach at first glance.

But I believe the personalization and customization for each child would be prohibitively expensive, even for the Bill and Melinda Gates Foundation, due to differences in human nature among students.  Optimistically, Gates says in a recent Business Week article that "the setbacks don't mean they have squandered the $1 billion the foundation has spent so far."  Still, we may need to wait until version 3.0 of the program to get it right.