Eduventure: The AR Edition - School Chat

Aaaaaaaugh three exams in one week (5th, 7th, and 10th), I could give all you genetics fetishists summaries of my course material as I go over it if you wanted? I'm probably gonna use what I just said as a chance to procrastinate, anyway

Here we fucking go oh and may as well start this on a decent note: Quantitative Genetics (aka Statistics in a Genetics outfit) awaaaaay:

So. As I think I concluded having hissy fit about a few weeks back, the majority of phenotypes (observable traits) are not dictated by single-gene loci. They may either be the result of a "cocktail" of genes which contribute, or the environment or "developmental noise" may play a degree in traits such as height or weight. The trouble with determining this kind of shit for humans is that eliminating either the environment or genetic variation is necessary to figure out to what degree each will contribute for any given trait.

Case in point: Language. It's erroneous to say that there's no genetic basis for language (developmental genetics contributes to correct formation of the cerebral and laryngeal structures necessary to produce any language, and by some statements the language you speak could be mistaken as having a genetic basis (because your parents speak it, so it's "passed down".)), because it's obviously dictated by environment. This whole thing humans have with looking after their offspring through the course of their development makes quantifying genotype as a contributor to phenotype really. Friggin'. Difficult. Either way, language spoken is an example of a familial trait, that isn't otherwise heritable.

This whole thing about either removing the "genetic variation" factor or the "environment as a contributor" factor (because any good scientific experiment changes only one variable at a time, then measures the other!) is why twin studies are a biggie. Sadly, there are a lot fewer twins than people who aren't, so we're still only looking at scattered data points from which we could extrapolate how different environments act upon the development of a particular phenotype.

One way to test for the "heritability" (a.k.a the genetic basis of any given phenotype) is to use (lemme go look it up) marker-gene segregation. This is where you pick a "marker gene" which usually has no obvious relation to the phenotype in question. If the inheritance patterns of the "marker gene" have any correlation with the phenotype, then that raises the possibility that there is a genetic basis for the phenotype of interest. (One example of a marker gene might be the one which crops up as the ABO blood groups in humans.) Another way (not so good for humans) is to pick a trait (let's say, seed weight) and self-cross individuals with the maximum and minimum seed weights in a population, then raise the offspring in an identical environment. If the two extremes produce offspring with different sizes in the same environment, that's good evidence for a healthy genetic contribution to the phenotype. If the two groups of offspring show a similar (often middle-of-the-road, mean/mode/average value) when raised in the same environment, the genetic basis is diminished.


The other issue with all this shit is that the same genotype may produce different, non-linear results in different environments. The degree to which this can occur is called the "norm of reaction." This means that any given genotype may well display a range of phenotypes given a range in environment. (Which is kind of a given, it's like leashing a goat with one length of rope tied to one post, then trying to predict where it'll be in an hour's time. That kind of thing gets easier to estimate if you leashed it to more bits of rope (eliminated more variables).