Monday, April 27, 2009

Genetic perceptions and(/or) illusions?

If as we tend to think, genes are not as strongly deterministic as seems often to be argued, then why do humans always give birth to humans, and voles to voles? Isn't the genome all-important, and therefore doesn't it have to be a blueprint (or, in more current terms, a computer program) for the organism? Even the other ingredients in the fertilized egg are often to a great extent dismissed in importance, because at some stage they are determined by genes (for example, in the mother when she produces the egg cell).

It's true that most people enter this world with the same basic set of parts, even if each part varies among people. But, in many senses the person is not predictable by his or her genome. You get this disease, other people get that one. You are athletic, others are not. You can do math, others can do metalwork. And so on, but most often a genetic predisposition for these traits can't be found.

We have made a number of recent posts about that genotype to phenotype connection problem. You generally resemble your relatives, which must be at least partly for genetic reasons, but the idea of predicting much more than that about your specific life from your specific genome is not working out very well, except vaguely or, it is certainly true, for a number of genetic variants with strong effects, that are often rare or pathological. In the latter cases are the genetic causes of diseases like muscular dystrophy or cystic fibrosis.

But these facts seem inconsistent! How can your genome determine whether you are an oak, a rabbit, or a person....and yet not determine whether you'll be a musician, get cancer, or win a Nobel prize? Is the idea of genetic control an illusion?

Yes and no. Part of the explanation has to do with what we refer to in our book as 'inheritance with memory'. Genomes acquire mutations and, if they aren't lethal, they are faithfully transmitted from parent to offspring so that some people have blue eyes and some have brown, some have freckles and others don't.

Over time, differences accumulate in genomes, and with isolation of one kind or another, lineages diverge into different species, and then differences continue to build up over evolutionary time. When millions of years later you compare a rabbit to a person (much less to a maple!), so much genetic variation has accumulated that there is clearly no confusing these different organisms. The genes, and the result of their action are unmistakable.

What happens at each step during development of an embryo reflects the cumulative effect of many genes, and is contingent on what has already happened. Thus, step by step by step in a developing rabbit embryo the rabbit foundation gets laid and everything that happens next basically depends on getting more rabbit instructions, because, except for genes that are very similar (conserved) across species, they are the only instructions that the cells in the embryo can receive.

Variation is tolerated but within limits--some rabbits have very floppy ears and some don't, but none have elephant ears or a rack of antlers (as in the jackalope pictured here) or can survive a mutation that gives it, say, a malfunctioning heart. So, you get a continuum of rabbit types, but because of how development works, a newborn rabbit can't veer very far from 'rabbitness'.

The same is true for humans--among those whose ancestors were separated on different continents, during thousands of generations, genetic variation arose and accumulated in these populations, and it's often possible to tell from a genome where a person's ancestors lived. That is, that person's genome's geographic ancestry.

When the subject turns to variation within a species or a population, however, the scale of variation that we are studying greatly changes. Now we are trying to identify genetic variation that, while still compatible with its species and population, and with successful embryological development, contributes to trait variation. Relative to species differences, such variation is usually very slight. But, it happens because, within limits, biology is imprecise and a certain amount of sloppiness (mutation, in this case) is compatible with life. In fact, it drives evolution.

What does this mean about the genetics of disease, or other traits that are often of interest to researchers, like behavior, artistic ability, intelligence, etc.? Culturally, we may make much of these differences, such as who can play shortstop and who can't. But often, these are traits that are not all that far from average or only are manifest after decades of life (e.g., even 'early onset' heart or Alzheimer disease means onset in one's 50s). Without even considering the effects of the environment, it is no surprise, and no illusion, that it is difficult to identify the generally small differences in gene performance that are responsible.

Of course, like a machine, there are many ways in which a broken part can break the whole machine, so that within any species there are many ways in which mutations that have major effects on some gene can have major effects on the organism. Mostly, those are lethal or present early in life. And, many mutations probably happen in the developing egg or sperm rendering that cell unable to survive--that's prezygotic selection. But even serious diseases are small relative to the fact that a person with huntington's disease or cancer is, first and foremost, a person.

Weiss and Buchanan, 2009
So, there may seem to be an inconsistency between the difficulty of finding genetic causes of variation among humans, and the obvious fact that genetic variance is responsible for our development and differences from other species. A major explanation is the scale of difference one is thinking about. Just because genes clearly and definitively determine the difference between you and a maple tree--and it's easy to identify the genes that contribute to that difference--that does not mean that the genetic basis of the trait differences between you and anyone else is going to be easy to identify. Or between a red maple and a sugar maple. Or that genetic variation alone is going to explain your disease risk or particular skills.

And there's another point. In biomedical genetics we are drawing samples from billions of people, whose diseases come to the attention of specialty clinics around the developed world, and hence are reported, included in data bases, and are put under the genetic microscope for examination. This means that we systematically identify the very rarest, most aberrant genotypes in our species. This can greatly exaggerate the amount of genetically driven variation in humans compared to most, if not all, other species.

However, it must be said that even the instances of other species (such as a hundred or so standard lines of inbred mice, or perhaps a few thousand lines of fruit flies or Arabidopsis plants (as in the drawing), from which much of our knowledge of genetic variation is derived), finds similar mixes of genetic simplicity and complexity. In other words, one does not need a sample space of hundreds of millions to encounter the difficulty of trying to predict phenotypes from genotypes.

So while this is true, it's also true that some variation in human traits is controlled by single genes, and those behave at least to some extent like the classical genetic traits that Mendel studied in peas. This variation arises in exactly the same way variant genes with small effects that contribute to polygenic traits arises--by mutation. But, the effects of mutation follow a distribution from very small to very large, and the genetic variants affecting the extreme are easier to identify. Some common variants of this kind do exist, because life is a mix of whatever evolution happened to produce. But complex traits remain, for understandable evolutionary reasons, complex.

No comments: