Monday, July 25, 2011

Mendelian Inheritance: Basic Genetics or Basic Mistake? Part I.

Gregor Mendel
Gregor Mendel wanted to improve horticulture in his native Moldavia.  The idea was to use hybridization to bring desired traits from different pea strains together by controlled breeding.  To do this effectively depended on an understanding of inheritance.  General selective breeding worked for some traits--and had been carried out for thousands of years.  But some traits in the available strains of plants seemed to be dichotomous (two-states, like round and wrinkled), and perhaps there would be effective ways to take one state, say, round peas, and breed that into some other strain that had attributes the round-strain didn't currently have.

Mendel carefully chose strains of basically inbred  pea plants that had simple patterns of inheritance: the dichotomous traits 'bred true' from  parent to offspring. He knew there were many traits that did not just have two states, or that did not breed true in this simple way.  But to understand the effects of hybridization--or we now wrongly say, inheritance--he picked appropriate traits. His seven carefully chosen traits all had an additional attribute that was important in this sense.  One of the states was 'dominant' to the other state, which was therefore 'recessive'.   These are essentially the terms Mendel himself used.

The 7 traits of interest to Mendel
This meant that a plant produced by crossing two parents with different variants of the trait would have only one of the traits.  If a plant received an 'A' allele (variant) from one parent, that was enough to give it the 'A' trait, even if it had received the other, the 'a' allele from the other parent.  So AA's and Aa's had the dominant trait, and only aa's the recessive.  From these, he observed the famous Mendelian ratios that are in all the standard genetics texts and that every geneticist 'knows' to be true.  So if an Aa plant were crossed with another Aa plant, since each parent has a 50% chance of transmitting each of its alleles, 1/4 of the offspring of this cross would be AA's, 1/4 aa's, and 1/2 of the offspring would be Aa's.  This and other similar conclusions depending on the particular breeding scheme, was what Mendel showed, and what revolutionized the understanding of inheritance and opened the door to powerful experiments in genetics that are our legacy and working basis to this day.

Note right off the bat that there was a very big mistaken conclusion that followed:  Mendel was showing the nature of trait inheritance, but it was interpreted to mean the laws of genetic inheritance.  It worked only because of the 100% correspondence  due strictly to the careful choice of 2-state, highly determinative traits in his experiments.  Although even that isn't strictly true (see my 2002 paper in Evolutionary Anthropology), it was close enough that even now we confuse inheritance that strictly applies only to genes, with the appearance of traits in offspring compared to their parents.

Classically, many 'Mendelian' diseases were identified, because they approximately followed Mendel's laws.  Modern biomedical genetics began, around 1900, with Archibald Garrod's studies of 'recessive' traits that arose in inbred marriages, that raised the chance that a child would inherit the recessive allele from both parents (because the genetically related parents shared the allele from their common ancestor).  Step by step, we built in the illusion that genes were just waiting their transubstantiation into traits.

Until the nature of DNA was understood and we could examine DNA directly we had to work through  traits rather than genes, even though we called the field 'genetics'.  For decades it was observable traits in experimental species, such as fly eye color, or 'Mendelian' disease, or similar traits in plants.  Genetics split into two parts, one  dealing with this kind of 'clear cut' particulate inheritance, which eventually led to understanding of how protein-coding areas were arranged along chromosomes, the nature of chromosomes, the nature of genes coding for proteins, and the transmission of DNA from parent to offspring.

The other segment of geneticists dealt with the majority of traits that clearly did not 'segregate' from parent to offspring, the quantitative traits  like stature, milk-yield, grain nutrient properties and the like, from which the field of quantitative genetics developed.  It was more pragmatic and said that a quantitative or complex trait was due to the inheritance of many genes: we might not be able to identify them, but jointly they were responsible for traits in organisms.  The similarities between parents and offspring for complex traits was consistent with this view as well.

Cajanus Cajan; Wikimedia Commons
Thus, sight-unseen, inheritance of traits was the  underlying theory, equated to the inheritance of genes, even when we couldn't identify the genes.  When the DNA sequencing age gradually developed, we got the idea that by tracking down genetic variance we are tracking down inheritance variance--the extension of the Mendelian illusion that genes were the same as traits!  This has led to the problems that we so widely see (but are so widely waved away) in studies like GWAS: we are not really finding the genes that 'cause' our favorite traits (including most diseases).

Of course all of this is manifestly a Grand Illusion!  Even a fertilized pea ovule does not have peas, wrinkled, green, or otherwise!  Once the connection between a gene and a trait becomes less than 100%, or once many genes contribute information about a trait, we see how obviously Mendelian ideas were a badly misleading mistake.  They were great for providing ways to set up experiments that isolated genetic effects and led to an understanding of genetic inheritance.  But they were, from the beginning, very misleading about trait inheritance.

In the next installments of this series, we'll examine the idea of dominance and genetic effects further, and will eventually ask whether, surprisingly,  there really is such thing as dominance in the first place!


John R. Vokey said...

I like this thread: keep it coming!

Ken Weiss said...

We'll try to be well-reasoning, to point out that here is a fundamental issue with proliferating consequences for science (and society). Further, it we think illustrates the grip that an assumed theory can have, even when the facts don't support it--or have to be modified or even twisted to make a fit.

Rosie Redfield said...

I've just reworked our Introductory Genetics course, trying to push dominance out of its pre-eminent position and give students a more rational view of the relationship between diploid genotype and phenotype. Not easy; all the available textbooks have drunk the Mendelian Kool-ade.

Ken Weiss said...

Yes, and not just textbooks, but most geneticists, the biomedical and pharmaceutical communities, and others. There is much more reason to want to believe in the Cool-Aid fairy than the truth fairy.

Anonymous said...

Just a small correction regarding Mendel's "native MOLDAVIA". Mendel was born in Heinzendorf bei Odrau, Austrian Silesia, Austrian Empire (now Hyncice, MORAVIA, Czech Republic).

Ken Weiss said...

Yes, of course, and thanks very much for spotting the blunder!