One of the greatest discovery of 20th century science was Watson and Crick's model of the structure and function of the DNA molecule. A rapid and exciting burst of research around this discovery formed the basis for the genetic revolution that we have been discussing in previous lectures.
By now, we've all grown up on the idea that many of our physical and behavioral (or "phenotypic") traits come from our genes which lie in our DNA and that one can describe these genes by describing the structure of a DNA molecule. Today we're going to discuss in more detail what the connection is between the two.
Today, we're going to discuss some of the philosophical questions lying at the very foundation of this discovery. In particular, we will be analysing the meaning of a central term in this theory, the notion of information and information transfer among the molecular building blocks of life.
The long-standing idea before Watson and Crick's discovery was that proteins give rise to our phenotypic traits, and in particular the 3D structure that proteins have.
Pauling thought that these structures served as templates for themselves, and thus, that the 3D structure of proteins was what's required to produce more proteins, by way of a "template" mechanism.
The 3D structure of proteins is what determines most of our phenotypic traits. These structures take various forms, including hormones, antibodies, enzymes, and many other things. For this reason, Pauling's program focused on these shaped.
They even went so far as to posit that the components making up the proteins, the amino acids, play no role in the shape of the protein itself. So, to Pauling, there was no "genetic code" of the type we are now familiar with.
However, Pauling was relying on just a few pieces of experimental evidence, which he thought showed protein shape does not depend on its amino-acid makeup. This evidence turned out to be faulty, and when it was overturned, this made way for the rise of the Francis and Crick revolution in molecular biology.
According to Crick, there were two main assumptions underlying his discovery. Both served to considerably simplify the things they were describing.
The first was the sequential hypothesis, that it was the components that make up proteins that cause it to fold up into the appropriate shape. This was in direct contradiction with the Pauling approach to molecular biology. And in a sense, it reduces the problem from one of three dimensional shapes to one-dimensional orderings of amino acids.
The second was called the central dogma, that once the genetic information makes it to the proteins, it cannot get out again. So, proteins cannot make DNA or RNA. That information has been too distorted by the time it gets there.
A further simplifying assumption, which you see in the Crick article itself, is that of all the nucleic acids that you see in a protein, there is actually an alphabet of only 20 that actually matter for the protein structure.
His coauthor Watson tended to oversimplify this description. In an influential textbook, Watson wrote that DNA creates RNA which creates protein. This is certainly usually the case, but not always, and the discovery of retroviruses in the 1960's proved that RNA can sometimes code for DNA. The HIV virus is an example of this.
Many thought that this amounted to overthrowing Crick's central dogma. His article shows that it does not really, and was nowhere denied in Crick's careful formulation of his central dogma.
On the Francis and Crick picture, these structures are encoded by the linear structure of DNA. We saw last time that a DNA molecule can be viewed as a sequence of nucleotides, G, T, C, A. This sequence determines the sequence of amino acids that make up proteins, which in turn determines the protein's 3D structure. In this sense, DNA encodes a protein's 3D structure. Here is a video that summarizes the process.
There are more details in this video.
Godfrey Smith's thesis is to deny the encoding of phenotypic traits in DNA.
He says that he accepts all the empirical claims. He even thinks that DNA encodes information about proteins. But he doesn't think it encodes information about our traits.
This is a really radical claim, in that it denies a standard picture of molecular biology, in which DNA is a "recipe for constructing a person," as geneticist George Beadle colorfully put it in 1957.
First, note that Godfrey-Smith accepts that DNA is correlated with phenotypic traits. He even accepts that DNA causes phenotypic traits. But he notes: the "environment" does that too, and in exactly the same sense.
What he's concerned with is a sense of information that "goes beyond" correlation and causation. He doesn't like it when we say that a sequence of nucleotides in a DNA molecule "encodes" our traits.
It's a compelling picture he's asking us to give up. You have sequences of letters A, C, T, and G, and appropriate combinations allow us to predict with incredible accuracy whether not someone has Huntington's, blue eyes, or any number of different traits. It sure looks like these sequences of letters encode phenotypic traits like Huntington's and blue eyes in some sense.
But Godfrey-Smith denies this.
Godfrey Smith's thesis is that genes do not encode phenotypic traits. Let me highlight two of his main arguments for this.
First he describes the way DNA "causes" phenotypic traits in the following way. To begin, DNA creates mRNA, which creates proteins. The DNA, he says, cleaerly codes for those proteins. But these proteins then enter into a complicated causal spaghetti in which various environmental factors plus facts about the proteins give rise to our ultimate traits.
In particular, although the protein structure may restrict causal structure, like to only have a some shade of blue eyes, the exact phenotypic trait depends on environmental factors.
Second, he says that if the Pauling protein-centric theory were true, then we wouldn't really be able to tell the difference. And that theory deals only with proteins and the environment.
So, he says, DNA isn't really needed in the description of phenotypic traits, and so in that sense can't possibly be what "codes" informationally for phenotypic traits. It plays no "useful theoretical role" in coding our traits. The proteins and their environment are what do that.
It is important to think carefully about Godfrey-Smith's argument. Should we so easily give up the famous picture that genes encode phenotypic traits? Do either of his two main arguments establish his thesis conclusively? (Question: Have I encoded information about the number that displays when you press the button below?)