In dealing with creationists, they will often steer the conversation to something like specificity. They will say that no new information can be added without a designer… with the requirement that they get to define information and then they define it in such a way that only a designer can add information.
One of their constant complaints is that randomness cannot create the specific genes/proteins that do very specific things using very specific interactions with other very specific genes/proteins. The issue is that our bodies require a specific gene to create a specific protein and, if at any point, something changes, then the whole system crashes and the organism probably dies. They say that this deep interaction among very specific systems is evidence of an intelligent designer, since humans are intelligent (generally) and specify very complex things that require very specific systems to run. Go ahead, run 7.5 volts through your USB port instead of the 5 volts required by specification and see what happens.
But most 8th grade students can explain why this objection makes no sense. There are things things called alleles, variations of a gene that has a slightly different result. The HLA family of genes has thousands of alleles.
Ah, but the creationists are saying, those are all just minor variations on the same thing.
Ok, fine. Whatever. But let’s look at this paper. Functional proteins from a random-sequence library.
There are countless variations of amino acid sequences. There are 20 amino acids used in the human body and there are no limits (of which I am aware) of how large a protein can be. Tintin is the largest known human protein. At over 34,000 amino acids in length, it approaches visibility by a good microscope. Now consider how many variations you could find of any combination of 20 choices (feel free to reuse them) and count every length from 2 to 34,000. It’s a pretty big number.
Creationists love to use scary big numbers. That’s their whole shtick in this case. “See, 1 protein out of all those nearly infinite number of combinations.”
But is that really the case?
The thing about proteins is that, unlike creationist claims, there are many, many paths to a protein that works.
In the paper I present, the authors build a random database of 1,000,000,000,000 proteins, each of which is 80 amino acids long. That’s just a fraction of the total number of proteins available in that size range.
They set an experiment (which, I’m given to understand, is very clever and very accurate), in which they could search this huge field of proteins for one very specific function. Binding to ATP. Now, ATP is the energy source of cells. It’s a nucleoside (adenosine) combined with three phosphate groups (Adenosine TriPhosphate – ATP).
I need to talk about the experiment and it’s kind of technical. Basically, the authors had these gel beds that had ATP on them. Proteins that could bind to the ATP would attach themselves to the ATP on the beads. Any proteins not attached where rinsed away. Then a wash was done with ATP in the wash. The proteins would preferentially bind to the free ATP, then be carried out in that wash for the next round of selection. Everything in that second wash was amplified using common practice and run through the sequence again.
Remember that the proteins were totally random. After 8 rounds of the experiment, where they looked for specific activity with ATP and chemically selecting those that bound to ATP, they found that 6% of the proteins had that activity.
We cloned and sequenced 24 individual library members, which showed that the population was now dominated by 4 families of ATP-binding proteins (Fig. 3a). These families show no sequence relationship to each other or to any known biological protein. The members of each family are closely related, indicating that each family is descended from a single ancestral molecule, which was one of the original random sequences.
First of all, that’s all pretty cool and it shoots the idea of specificity pretty much out of the water. In a random collection of proteins, at least one was capable of reacting with ATP. But it’s not the one used by any known organism.
This also shows that there are 4 different methods (at least) of accomplishing that goal, reacting with ATP. So another creationist notion is destroyed. There can be changes to proteins and have the remain functional. In fact, there were so many changes that there were four different families of proteins that all did the same thing.
Here’s where it gets moderately interesting though. The researchers took those proteins and subjected them to mutagenesis. They mutated them, probably by copying them with a low fidelity polymerase. For three more rounds of the experiment. Then six more rounds with out being mutated.
The amount of ATP binding proteins rose to over 30% of the total. Which neatly destroys one of Casey’s complaints about mutation and selection (#4 on this list) that natural selection struggles to fix traits in a population. In 9 rounds of the experiment, the percentage of ATP binding proteins when from 6% to 34%. Interestingly, all of the individual proteins sampled were from the same family. Showing that particular family had a real advantage in the selection process.
In conclusion, we suggest that functional proteins are sufficiently common in protein sequence space (roughly 1 in 1011) that they may be discovered by entirely stochastic means, such as presumably operated when proteins were first used by living organisms
Science is really impressive when you realize that, even obscure papers such as this single-handedly destroy multiple notions and complaints of creationists. It was never the authors purpose to do this. But it shows just how fragile and edifice that the creationists have created. Pick an evolution paper almost at random and it will devastate an argument of the creationists.
I need to add that I wrote this post specifically (get it?) as a counter to a couple of creationists that are pissing me off with their refusal to learn. However, the paper is really cool in and of itself. One thing that the authors suggest is that this may be a method to find proteins to do something that other researchers need a protein to do.
Designing a protein from scratch to specifically accomplish a specific reaction is impossible. There is too much going on in proteins. The sequence is not the only factor. There are multiple levels of folding involved and a host of other things that take super-computers to calculate even for one protein.
Maybe designing proteins from scratch isn’t the best approach. What if, a random search followed by evolution is a viable approach. In probably a couple of weeks, the authors found a highly efficient ATP bonding protein that isn’t already used by an organism (that we’re aware of). This is something that a team of people with supercomputers couldn’t design in a hundred years (probably).
Evolution beats design.
P.S. Many thanks to the commentors at Larry Moran’s Sandwalk blog for help with the biochemistry.
1 In case you’re wondering, yes, that was an attempt to use specify in as many variations and as many times as possible.
2 Keefe, A. & Szostak, J. Functional proteins from a random-sequence library. Nature410, 715–718 (2001).
3 I say chemically selecting those that bound to ATP because the authors did not look at all the molecules and preferentially select those that did what they wanted.