Silent Sites: A Creationist View

Silent sites in DNA are base pairs that do not affect the protein coded for. Evolutionary biologists explain the patterns observed in the silent sites by the gradual accumulation of mutations over long time periods, with organisms that are more closely related presumably having a closer agreement, and organisms more distantly related having less agreement in the silent sites. Another pattern that is observed is that the silent sites typically differ in two organisms by much more than the remaining sites. This can be explained in evolutionary terms by the fact that mutations to non-silent sites are more likely to be harmful, and so the non-silent sites change more slowly in response to the accumulation of mutations over long time periods. But how can a creationist explain the observed patterns in the silent sites? And what about the patterns observed between organisms in the non-silent sites?

This seems like a rather unusual subject even to be considering, but it has been raised in a discussion with a biologist. Let me propose an answer from a creationist view. The Creator will, I suppose, choose the features of each organism to optimize its function, whatever that may be. So if there is some purpose P for a hypothetical organism to be created, there will be a function f_P associated with it, measuring how closely a possible organism fulfils purpose P. Thus if X is a possible organism, f_P(X) is a real number measuring how closely organism X fulfils purpose P. We can assume that f_P is a function from organisms to real numbers such that the higher the value of the function f_P(X), the more closely the organism X meets the purpose P. So if the Creator has a purpose P in mind, He will presumably create the organism X that maximizes the associated function f_P(X). Since we are considering the DNA, we can consider f_P as a function from the DNA sequence of the organism to the real numbers. For this, it is convenient to express the DNA sequence as a sequence of bits (zeroes and ones). We can express each base by two bit positions. One bit position will specify whether the base is a purine or a pyrimidine, and the other bit position will specify the base. Thus, with two bits, we get four possibilities, corresponding to the four DNA bases. We can then represent n bases (base pairs of DNA) by 2n bit positions, each bit being either a 0 or a 1.

Suppose there are two purposes P and P' which are very similar. Now, suppose the functionality f_P(X) of this sequence for a purpose P can be represented by a linear function c₁b₁ + c₂b₂ + ... + c_nb_n, where the c_i are real coefficients and the b_i are either 0 or 1. This is an oversimplification, but it will illustrate the ideas. Suppose the functionality f_P'(X) of this sequence for purpose P' can be represented by the linear function c'₁b₁ + c'₂b₂ + ... + c'_nb_n. We can assume that the Creator would choose the b_i to optimize (maximize) this value. Thus if c_i is positive, b_i would be one, and if c_i is negative, b_i would be zero. If the purposes P and P' are very similar, c_i and c'_i would also be very close. Let A and A' be the organisms maximizing f_P(A) and f_P'(A'), respectively. Thus A is the organism that best fulfils purpose P, and A' is the organism that best fulfils purpose P'. Then we would expect A and A' to be similar organisms, because the purposes P and P' are similar.

Silent sites (typically third codon positions) may have some function, even though they do not affect the amino acid coded for. For third codon positions, therefore, the coefficients c_i would be very small, corresponding to the fact that these positions typically do not have much of an effect. Thus the "noise" or random difference between organisms A and A' is likely to have a large effect on these c_i, so we would expect many of the third codon positions to differ between A and A' because the signs of c_i and c'_i differ. However, for non-silent sites, we would expect most of the c_i to have values much different from zero, so the optimum value of b_i would usually be the same in both organisms. This explains why the silent sites would tend to differ more than the non-silent sites. However, sometimes we might find the silent sites nearly the same in organisms that are fairly similar.

This also explains the observed pattern of amino acid sequences in similar proteins between different organisms. Organisms that are more similar tend to have proteins (hemoglobin, cytochrome C, etc.) that are similar, and more distantly related organisms have greater differences in the proteins, as a rule. Some proteins, such as fibrinogen peptide, vary a lot between different organisms; this can be explained by the fact that the choice of amino acid has very little effect, and thus it is likely that an optimum choice will be very different for two different organisms.

There may also be some third codon positions that have no effect at all. For these, the choice may be determined by artifacts of the creation process, by esthetic considerations, or by other means (such as providing a type of "serial number").

Since the Creation, mutations have accumulated, so some of the observed patterns are not due to created differences, but have arisen more recently. In general, however, this latter contribution should be relatively small.

We hope that readers will not find this discussion to be too far out!

Back to home page.