I often get questions about the immune system, because many in the Darwin lobby point to it as an example of “random mutations”. That is kind of a half-truth, but the full truth is so amazing that it needs to be told more often.
Anyway, I’m posting this so I have something to refer people back to in the future.
The immune system is often used by the Darwin lobby to say that random mutations can have great effect. An example of this is here (see Part 2 on “Antibody Genes”) and it was used in Dover (see here). It also appears in several books as well.
There are two important processes that are often discussed. They are both important, and both involve randomness, are both highly designed, and are both well worth knowing. They are:
- V(D)J recombination – the process that generates the initial set of antibodies
- Somatic Hypermutation (SMH) – the process that produces new antibodies in response to a new antigen
We’ll start by considering V(D)J Recombination.
There a millions of different antibodies that you have to assist you. There are essentially two parts to an antibody – the CDR region (the “complementary-determining region”, which sticks to the antigen), and the C (Constant) region (which signals the immune system that it needs to come and take a look). We only have a few different C regions, each for different types of immune responses. However, we need millions of different CDR regions, because there are millions of things to watch out for.
However, we don’t have millions of genes. The total number of human genes is less than 25,000. So what happens? How do we get these millions of different proteins from a handful of genes? The genes for the CDR region are broken up into three different “interchangeable parts” – a V, a D, and a J (they stand for Variable, Diversity, and Joining, but that doesn’t really matter for our discussion). Pretty much any V can be matched with any D then with any J. Each of these “pieces” is marked with a marker that allows the genetic system to cut them at the proper place for recombination, called the RSS (recombination signal sequence). Because the process is combinatorial, you can get the millions of antibody genes needed from just recombining a few hundred gene parts!
Now, each cell recombines a different set of these gene pieces. However, how is the communication handled so that each individual cell knows which combinations have already been tried so it can be sure to do a new one? Answer – they don’t. They don’t have to! In such a process, all you need is to randomize which particular pieces you are trying out, and the net effect will be almost the same as if you communicated about which ones to do, without any communication overhead!
So, it is random in the sense that there is a combinatorial process that includes a random variable, but it is in no way random because the specific pieces which can be recombined are specifically marked, they are recombined in the proper order, and with a C region attached to the end to communicate with the rest of the cell. It is kind of a lego-like building block system – interchangeable parts that are different but with matching attachment sites. It only works if all of the parts match the design (i.e., the right size, flanked by the correct RSS so the immune system knows where to cut, etc.). There is a random component, but it is carefully controlled.
That’s the V(D)J system which lands the initial set of antibodies. But what about new antibodies? That’s where the Somatic Hypermutation system comes into play.
The Somatic Hypermutation (SMH) System
The SMH System is another wonderful system that generates new antibodies when needed. If you have an unknown antigen come into your body that your body doesn’t have a matching antibody for, then it knows it has to generate one. How does it do this?
- It detects that it needs to generate a new antibody.
- It takes an existing, antibody gene (one that is already recombined as above) that almost matches the antigen.
- It mutates JUST THAT GENE. Not only that, it only mutates the CDR of the gene. That is, it skips the C region (why? because it still needs to signal to the immune system that it found something! If it mutated the CDR, the communication would break down.).
- When it detects that it has a match, it stops mutating.
So, where is the randomness? Well, within the CDR, as best as we can figure, the immune system generates random changes. Is that really a random mutation, though? It has excluded 99.99993% of the genome and focused on the 0.00007% that it knows needs changing. What I tell people is that I will be happy to agree with them that this process is 0.00007% random if they will agree with me that it is 99.99993% designed. That usually ends the conversation.
What’s also interesting is that the cell itself starts and stops the process as needed. That is, these aren’t just happenstance mutations – the cell actively knows it needs mutations, and activates a system to produce them, then produces them within the span of 2,000 base pairs that it knows is likely to produce benefit, and then stops once it finds a benefit (it’s actually closer to 500 base pairs that get significant mutations, but there are a few rare mutations farther into the gene).
So, that is the long story of how, yes, there is randomness involved in these processes, but that is not the same thing as a philosophically random (i.e., unconstrained by systems) process. These are very constrained and directed by their systems, but they do allow small amounts of randomization precisely at the points where it is most useful to do so.
I once did a paper discussing the difference between philosophical and statistical randomness and its importance in this discussion, which you can find here.
Here’s a summary of the random-vs-nonrandom pieces of the processes:
|System||Nonrandom Elements||Random Elements||How Randomness is Harnessed|
||The randomness is used to prevent communication overhead for having a massively parallel antibody generation process.|
|Somatic Hypermutation (SMH)||
||The randomness matches the amount of ambiguity that the cell can know about which locations to mutate. This makes use of Bernoulli’s Principle of Insufficient Reason to develop an effective mutation strategy.|
With this table, I have probably left out an inordinate number of ways in which these systems are not random. But, my main point is that they are not random in haphazard ways, but in very specific ways that are nuanced to the needs and patterns of the process at hand.
In other words, there may be some ways that, statistically, there are some properties of randomness which show up, but there is no evidence whatsoever that this statistical randomness implies a philosophical randomness (i.e., that these mutations occur outside of the control systems of the cell). Quite the opposite – the features of these processes which are statistically random are just as tuned as the features which are not, and none of them give any evidence that you could get similar results if you had a mutation process that was not under similar types of control structures.