Can natural selection create?

6. Endless fitness valleys

Evolutionists agree that to create a new gene requires a great deal of mutational "experimentation". During the "construction phase" of developing a new trait or a new gene, we have to expect a period of time when the experiment reduces a species' fitness. This is called a fitness valley. A half-completed gene is neither beneficial nor neutral - it is going to be deleterious. So in a sense, the species has to get worse before it can get better. It is easy to imagine a species surviving fitness valleys if they are brief, and if they are rare.

However, long deep fitness valleys are likely to lead to extinction, not evolution. The extreme rarity of desired mutations and the extreme slowness resulting from Haldane's dilemma, should make fitness valleys indefinitely long and deep. If evolutionary innovation was continuous (as is widely claimed), then a species' fitness should just keep going down - life would be just one fitness valley upon another upon another. Life's super-highway of evolution would always be under construction, and total fitness would always be declining rather than increasing. The concept of a species passing through fitness valleys makes evolutionary sense only when individual traits are considered. However, when the whole genome is considered, the concept of indefinitely numerous, and indefinitely long, fitness valleys argues strongly against the evolution scenario.

7. Poly-constrained DNA - Most DNA sequences are poly-functional, and so must also be poly-constrained. This means that when a DNA sequence has meaning on several different levels (polyfunctional), each level of meaning limits possible future change (poly-constrained). For example, imagine a sentence which has a very specific message in its normal form, but has an equally coherent message when read backwards. Now let's suppose that it also has a third message when reading every other letter, and a fourth message when a simple encryption program is used to translate it. Such a message would be poly-functional and polyconstrained. We know that misspellings in a normal sentence will not normally improve the message - but at least this would be possible. However, a poly-constrained message is fascinating, in that it cannot be improved - it can only degenerate (see Figure 12). Any misspellings which might possibly improve the normal sentence form - will be disruptive to the other levels of information. Any change at all will diminish total information - with absolute certainty.

There is abundant evidence that most DNA sequences are polyfunctional, and therefore are poly-constrained. This fact has been extensively demonstrated by Trifonov (1989). For example, most human coding sequences encode for two different RNAs, read in opposite directions (i.e. both DNA strands are transcribed - Yelin et al., 2003). Some sequences encode for different proteins depending on where translation is initiated and where the reading frame begins (i.e. read-through proteins). Some sequences encode for different proteins based upon alternate mRNA splicing. Some sequences serve simultaneously for protein-encoding and also serve as internal transcriptional promoters. Some sequences encode for both a protein coding region, and a protein-binding region. Alu elements and origins-of-replication can be found within functional promoters and within exons. Basically all DNA sequences are constrained by isochore requirements (regional GC content), "word" content (species-specific profiles of di-, tri-, and tetranucleotide frequencies), and nucleosome binding sites (i.e. all DNA must condense). Selective condensation is clearly implicated in gene regulation, and selective nucleosome binding is controlled by specific DNA sequence patterns - which must permeate the entire genome. Lastly, probably all sequences do what they do, even as they also affect general spacing and DNA-folding/architecture - which is clearly sequence dependent. To explain the incredible amount of information which must somehow be packed into the genome (given that extreme complexity of life), we really have to assume that there are even higher levels of organization and information encrypted within the genome. For example, we know there is another whole level of organization at the epigenetic level (Gibbs, 2003). There also appears to be extensive sequence-dependent three- dimensional organization within chromosomes and the whole nucleus (Manelidis, 1990; Gardiner, 1995; Flam, 1994). Trifonov (1989), has shown that probably all DNA sequences in the genome encrypt multiple "codes" (up to 12 codes). In computer science, this type of "data compression" can only result from the highest level of information design, and results in maximal information density. These higher levels of genomic organization/information content, greatly multiply the problem of poly-constrained DNA. Every nucleotide interacts with many other nucleotides, and everything in the genome seems to be context-dependent. The problem of ubiquitous, genome-wide, poly-constrained DNA seems absolutely overwhelming for evolutionary theory. Changing anything seems to potentially change everything! The poly-constrained nature of DNA serves as strong evidence that higher genomes cannot evolve via mutation/selection - except on a trivial level. Logically, all poly-constrained DNA had to be designed.

8. Irreducible complexity - The problem of irreducible complexity has been brilliantly presented by Behe (1996). He has illustrated the concept of irreducible complexity in various systems that have multiple components, such as a mousetrap design which requires 5 independent parts, or a flagellum having perhaps 10-20 component parts. His idea is that each part has no value except within the context of the whole functional unit, and so irreducible systems have to come together all at once, and cannot arise one piece at a time. In the case of a mousetrap - all the pieces may have been sitting next to each other on the inventor's workbench - but they would not have come together by chance, or by any realistic evolutionary progression. They came together as a synthesis, simultaneously, in the mind of the inventor. It is in the realm of mind that deep complexity comes together and becomes integrated. In our example of the evolution of transportation technology, the simplest first improvement we might imagine might be the occurrence of misspellings that would convert our red wagon into a blue tricycle. It is indeed easy to imagine a misspelling that might cause the paint code to be changed (although the blue paint would have to already be available, and coded). Likewise, a misspelling could certainly cause a wheel to fall off. However, a three-wheeled wagon is not a tricycle - it is a broken wagon. To convert a wagon to a trike would require extensive reworking of the instruction manual and radical changes in most of the manufactured component parts. There would be no intermediate functional steps to accomplish these complex changes, and so no prospect for our quality control agent to selectively help the process along - in fact he would be selecting against all our desired misspellings and changes. So the correct combination of misspellings would have to arise simultaneously by chance, all at the same time - which would never ever happen. Obviously, a trike could only arise from a wagon by way of intelligent and extensive reworking of the design, and a thorough re-writing of the instruction manual (see Figure 13).

Although a wagon or trike may have dozens of component parts, even the simplest protein is a much more complex machine - having hundreds of component parts, and thus representing irreducible complexity profoundly greater than that illustrated by our wagon analogy. As the number of components of a design increases linearly, the number of interactions (hence the complexity) increases exponentially.

As complex as proteins are, underlying every protein is a genetic system comprising even higher levels of irreproducible complexity.

The molecular machinery underlying the coding, transcription, and translation of a protein is phenomenal. Ignoring all the other accessory proteins involved, just the design of the DNA/RNA sequence is mind-boggling. Although a simple protein has a few hundred component parts, the underlying gene that produces it has thousands of component parts. All of these parts are interacting and mutually-defining. Each nucleotide has meaning only in the context of all the others. Each nucleotide is polyfunctional - interacting with many other nucleotides. The DNA sequence defines regional 3-D chromatin structure, local protein binding, uncoiling, transcription, and also defines one or more RNA sequences. The RNA sequence defines RNA stability, RNA variable splicing, RNA processing, RNA transport, transcription efficiency, and protein sequence.

We do not yet really understand how any single gene from a higher life form really works - not in its entirety. Not in the context of everything else that is happening in the cell. A single gene with all its interactions is still way too complex for us. When we consider the full complexity of a gene, including its regulatory and architectural elements, a single gene has about 50,000 component parts. I presume that this is more component parts than are found in a modern automobile. There is no simple linear path that leads car components to spontaneously become a functional car - mind is obviously required (actually, many brilliant minds). In the same way, there is no linear path of selection that can build a single gene from its individual nucleotides - a mind is likewise required. Yet a single gene is just a microscopic speck of irreducible complexity, within the universe of irreducible complexity that comprises a single cell. Life is itself the very essence of irreducible complexity - which is why we cannot even begin to think of creating life ourselves. Life is layer upon layer upon layer of irreducible complexity. Our best biochemical flow charts, of which we are so proud, are just childish cartoons of true biological complexity - which is something we cannot even comprehend. It is a tribute to the mind of man that we have started to understand how even a single gene works, and that we can now design and build very small artificial genes. But we still cannot design a new gene for a new and unknown protein, which could then precisely integrate into the complexity of a higher life form. If we cannot do this - why would we think that random mutations, combined with a very limited amount of reproductive sieving, could accomplish this? For the reader's interest I have attempted to expand upon the concept of irreducible complexity - with the concept of Intergrated Complexity (see Appendix 3).

9. Almost all beneficial mutations must be near-neutral. We have already discussed at length the difficulty of -selecting against near-neutral deleterious mutations, and this problem is begrudgingly acknowledged by most geneticists. However, there is a flip side to this problem, which is even more important, but which I have never heard acknowledged. As we have already discussed in Figure 3d, the problem of near-neutrality is much more severe for beneficial mutations than for deleterious mutations. Essentially every beneficial mutation must fall within Kimura's "no selection zone". All such mutations can never be selected for. This problem multiplies all of the problems I have already outlined above. Our hoped-for new gene will certainly have a few nucleotides that have major effects - for example the ones that specify the active site of an enzyme. But such nucleotides can only have major effects within the context of the whole protein and the whole gene sequence. The whole protein/gene is constructed primarily with components that individually have only a small impact on the whole unit, and have only a miniscule impact on the fitness of the whole individual. In combination, these nucleotides contain most of the information contained within the gene - without them the "important nucleotides" are meaningless. Yet they are all individually un-selectable. So how can we establish them and keep them in their respective places, during gene construction? The answer is obviously that we simply cannot. And apart from these "insignificant masses" of nucleotides the elite "important nucleotides" cannot be selected for either. Because of the near-neutral problem, we cannot even get to first base in terms of building our hoped-for new gene. The entire framework of the new gene is defined by the near-neutrals - but there is no way to either put them or hold them in place. The near-neutral nature of beneficial mutations is strong evidence that every gene had to be designed, and that there is simply no conceivable way to build a gene one nucleotide at a time, via selection.

10. Putting bad mutations back in the picture. We have briefly considered a variety of powerful arguments about why progressive mutation/selection must be very limited in its scope. These arguments have temporarily excluded from consideration all deleterious mutations. However, in reality, progressive selection must occur in the real world, where deleterious mutations outweigh beneficial mutations by perhaps a million to one. To be honest, we must now re-introduce deleterious mutations.

a) Muller's Ratchet - As I have mentioned earlier, when we study the human genome, we see that large blocks of DNA have essentially no historical evidence of recombination (Gabriel et al. 2002, Tishkoff and Verrelli, 2003). Recombination appears to be primarily between genes rather than between nucleotides. So within any limited gene sequence there is essentially no recombination. Any such block of DNA that does not have recombination is subject to "Muller's ratchet" (Muller, 1964). This means that the good mutations and the bad mutations cannot be separated. Since we know that the bad mutations overwhelmingly outnumber the good, we can be certain that any such stretch of DNA must degenerate. The hordes of bad mutations will always drag the rare good mutations down with them. While we are waiting for a rare beneficial mutation, bad mutations are piling up throughout the region. Even if we could succeed in accumulating perhaps a hundred "good" mutations within a region, and were waiting for the next one to come along - we would start to see many of our good mutations start to back-mutate into the bad. Time is our enemy in this situation - the more time, the less information. Muller's ratchet will kill a new gene long before it can take shape.

b) Too much selective cost - In previous chapters we have discussed the cost of selection. Haldane's dilemma only considers progressive selection. But we can only afford to "fund" progressive selection for beneficial mutations after we have paid for all other reproductive costs - including all costs associated with eliminating bad mutations. As we have already seen, there are so many bad mutations we cannot afford even to pay just the reproductive cost of eliminating them. Since we cannot afford to stop degeneration - we obviously have nothing left over to fund progressive selection. There is just one way around this. In the short run, we can fund progressive selection for a very limited number of traits - if we borrow "selection dollars" from our long-term struggle against bad mutations. However, we need to understand that this means that any short-term progress in terms of specific beneficial mutations is paid for by faster genomic degeneration in the long run.

c) Non-random mutation - As it turns out, mutations are not entirely random. Can this help us to create new genes? No, it makes our problem much worse! For example, we now know that some nucleotide positions are much more likely to mutate than others ("hotspots"), and that certain nucleotides are favored in substitutions. Mutational "hot spots" will give us the mutant we want sooner in that location, but while we then wait for the complementary mutations within the "cold spots", the hotspots will proceed to back-mutate again. We are forced to keep re-selecting our good mutations within the hot spots, while we wait for even the first good mutation to occur within the cold spots. This makes things worse, rather than better. The greater tendency to mutate to a certain nucleotide, (let's say T), will help us in positions where T is desired, but it will slow us down whenever G, C, or A is desired. Therefore, 75% of the time the bias toward T mutations will slow down progressive selection. "Non-random mutation" sounds good from the point of view of building information, but unfortunately we are not talking about the non-randomness of design - rather we are talking about a type of non-randomness which (ironically) is antithetical to information building.

We have reviewed compelling evidence that even when ignoring deleterious mutations, mutation/selection cannot create a single gene - not within the human evolutionary timescale. When deleterious mutations are factored back in, we see that mutation/selection cannot create a single gene - ever. This is overwhelming evidence against the Primary Axiom. In my opinion this constitutes what is essentially a formal proof that the Primary Axiom is false.

In conclusion, the genome must have been designed, and could not have evolved. Yet we all know that "micro-evolution" (adaptive selection) does happen, correct? How can this be? To use the terminology of our earlier chapters, mutations are the dings, scratches, and broken parts of life. Therefore, I believe most useful variation must be designed. When we see adaptive selection occurring, we are usually witnessing segregation and recombination of useful variants of genes and gene components - which were designed to segregate and recombine in the first place. We are not usually seeing the result of random mutations - which are consistently deleterious. Selection operates to eliminate the worst of mutations, while favoring the most desirable recombinants and segregants of designed variation. For example, a single human couple, if they contained designed and functional heterozygousity at only a tiny fraction of their nucleotides, would produce (via recombination and segregation) an essentially unlimited range of useful diversity. It is this type of designed diversity that natural selection can act upon most effectively. All such designed variants would be expected to be created within useful linkage groups, and would have originated at high allelic frequencies. For example, in the case of a single human couple, there could be only four initial sets of chromosomes - so all initial nucleotide frequencies would be at least 25%. Functional linkage groups and high allele frequency allow for very rapid responsiveness to selection, and thus rapid local adaptation. Like an ordered deck of cards, the net information in such a scenario would be greatest at the beginning, but diversity would be greatest only after many hands had been played out. Except at the beginning, no new information would be required.

Hopepage

If you have any comments or questions
contact: bhaktivedanta_108@yahoo.com