My partner has adored chess ever since he was a child.While I’m no aficiando like he is, there’s something easy to romanticize about the game. The sheer artistry of the pieces themselves, the beautiful organization of the alternating black and white squares, and the intellectual skill a game such as this demands. It’s easy to see how its popularity has stood the test of time and maintains an ardent group of enthusiastic devotees.
In order to be a successful chess player, one must be capable of thinking numerous moves ahead and entertaining a variety of possible alternatives. This is the primary reason the majority of people are not good at chess because, though humans are arguably the smartest animals on earth, our predictive and multitasking capacities are still, well– shall we say “less than desirable”? And I certainly don’t exclude myself from this group when it comes to chess. I’m afraid I’m more a collector than a player.
Biology, on the other hand, is like a chess game, only whose complexity is multiplied by an almost-infinitesimally-large number. How do we ever hope to predict it?
We humans aren’t too bad at holding two objects in our head at one time. Ball A rolls down a hill and hits Ball B which causes Ball B to move. Okay, no biggy. But once we add in a Ball C, then we may start having problems. And heaven forbid we throw into the mix Balls D, E, F, and G!
Very rarely can we identify an issue in biology that’s as simple as Molecule A interacts with Molecule B which causes Molecule B to change its chemical behavior. That’s not to say that adding a given catalyst to a reaction can’t cause changes in chemical behavior. But more often it’s this scenario: Molecule A interacts with Complex B (made up of Molecules B, C, D, E, F, and G) which causes Complex B to change its chemical behavior. But in order to predict how Molecule A would interact with Complex B you already have to know how Molecules B, C, D, E, F, and G interact with each other especially if these subunits vary. Strictly from a probalistic calculation, the potential complexity of Complex B could be represented by:
B x C, B x D, B x E, B x F, B x G, B x C x D, B x C x E, B x C x F, B x C x G, B x D x E, B x D x F, B x D x G, B x E x F, B x E x G, B x C x D x E, B x C x D x F, B x C x D x G, B x D x E x F, B x D x E x G, B x E x F x G, B x C x D x E x F, etc. etc. etc.
In other words, 720 (6 x 5 x 4 x 3 x 2) possible combinations of interactions which we could feasibly study and which, chemically, could each have divergent outcomes. Now add Molecule A and I think you might grasp how even the simplest of biochemical interactions can become overwhelming. You can undoubtedly generalize Complex B as a single unit, but in cases in which the constituents vary you need to know at that given moment in time what subunits comprise the larger unit.
Have your eyes gone crossed yet? 😀
I like to remind people, laymen and scientists alike, how complex biology is. The purpose of this isn’t to overwhelm and stalemate scientific progress but hopefully to remind people to be cautious with their hypotheses and avoid broad generalizations that rest upon a Ball-A-hits-Ball-B foundation. Let me give an example:
Genetics. We love the idea that a single gene codes for a single gene product (usually a protein) and that if a mutation occurs within that gene, that will translate into a change in the gene product. We then also like to take that construct and then stretch it even further to say things like, “Because Single Nucleotide Polymorphism (SNP) B occurs in Gene A more frequently in people who do really well on intelligence tests, that means that this mutation in Gene A is causally linked with intelligence.” This kind of data interpretation happens all the time, and this in fact is taken from a recent example of research performed on the HMGA2 gene and intelligence. Here’s a sample report from BioNews:
“[HMGA2] affects the overall size of the brain and links to intelligence. People with a small change in this gene had larger brains and performed slightly better on IQ (intelligence quotient) tests in studies.”
And here is a quote from a blogger who’s taken it even a step further:
“A mutation in a single gene, [HMGA2], has been identified as being responsible for intelligence. [HMGA2] has been found in people with larger brains and has a positive correlation with the level of IQ. But if a large brain is associated with high IQ and the likes of Einstein apparently had a smaller than average brain surely his ingenuity must be down to the nurtured brain as supposed to this genetics, bearing in mind that the mutation in [HMGA2] only increases IQ by an average of 1.3 points. The high IQ in Einstein’s brain must therefore be due to other factors including the nurturing of this non-mutated [HMGA2] gene to behave as if it were in a larger brain to encourage the shift above the average IQ.”
Obviously if there’s any kind of relationship between a gene mutation and a gross phenotype there’s a causal link right? In the world of Ball-A-hits-Ball-B that would be true, and if only it were. But let me give you a single example of how a given gene mutation can have a relationship with a larger phenotype without being causally associated with it. In fact we’ll use a hypothetical (though feasible) example involving big brains:
The molecule, PTEN, sits in an interesting place within the cell. Not only does it help suppress cell proliferation and growth, it also plays additional roles in DNA stability and repair processes. Now let’s just say that the HMGA2 gene for whatever reasons is inherently less stable than your average gene. And let’s say that the activity of PTEN is mildly suppressed during prenatal brain development (maybe a genetic proclivity or an environmental suppressor, doesn’t matter). This subsequently leads to increased proliferation in your neural progenitor population, leading to more neurons and bigger brains. Because you’ve suppressed PTEN this may also lead to increased occurrences of mutations due either to instability or poor repair. Because HGMA2 is a more likely target for mutation than your average gene, on average it will be mutated more frequently in this scenario. But does that mean that the HGMA2 mutation led to the increase in brain size? No, the primary cause in this instance was the reduced PTEN activity. What this scenario actually resembles is this:
In the above scenario, Ball A is PTEN while Balls B and C are brain growth and the HMGA2 gene. In the case of HMGA2, brain growth, and intelligence the only thing we can be certain of is a relationship, not causation. And in fact, what is currently known about the functions of HMGA2 don’t immediately reveal any obvious relationship to neural proliferation. Here is what GeneCards has to say about it:
“This gene encodes a protein that belongs to the non-histone chromosomal high mobility group (HMG) protein family. HMG proteins function as architectural factors and are essential components of the enhancesome. This protein contains structural DNA-binding domains and may act as a transcriptional regulating factor. Identification of the deletion, amplification, and rearrangement of this gene that are associated with myxoid liposarcoma suggests a role in adipogenesis and mesenchymal differentiation. A gene knock out study of the mouse counterpart demonstrated that this gene is involved in diet-induced obesity. Alternate transcriptional splice variants, encoding different isoforms, have been characterized.”
As it says, HMG proteins are essential components of the enhancesome, which is a complex that binds to enhancer regions of DNA and upregulates or “enhances” the rate of transcription. Obviously these will be involved in brain development, but they’ll also be involved in every other kind of tissue. At first glance, there is nothing overtly obvious linking HMGA2 to neural proliferation.
Now, I want to state categorically that I’m not saying certain types of HMGA2 mutation don’t affect brain size and subsequent changes in intelligence; it could well happen. What I am saying is that we’re getting far too ahead of ourselves drawing causal inferences without even knowing what this gene product does in such a scenario. And in the case of known SNPs in this gene, we don’t know how those mutations ultimately effect– or don’t effect– its gene product(s). We need to answer these questions first before we begin to address whether it leads to changes in brain tissue formation and ultimately behavior.
It’s all too easy to get sucked into Ball-A-hits-Ball-B thinking. One thing which I find helps me from jumping ahead of the data too frequently is to remind myself of specific scenarios like the PTEN one above. It keeps me skeptical, it keeps me thinking and retesting my assumptions. And, in short, it helps keep me honest.
I’d like to invite any readers to proffer their examples of anti-Ball-A-hits-Ball-B thinking within the comments. The more the better!