Icons of ID: Avida and Common descent

| 10 Comments

The evidence for common descent, which is a logical prediction from Darwinian descent is quite extensive.

Douglas Theobald presents an excellent outline of the argument at 29+ Evidences for Macroevolution The Scientific Case for Common Descent

I will show how simple experiment with digital and real DNA both recapture the expected nested hierarchies and phylogenies. I hope this may resolve some of the confusions about Darwinian evolution, common descent, phylogenetic reconstruction based on DNA and nested hierarchies. So in other words, while Darwinian evolution leads inevitably to a nested hierarchy, such phylogeny may not always be recoverable using DNA.

Using Avida to Test the Effects of Natural Selection on Phylogenetic Reconstruction Methods

We find that if the organisms undergo natural selection between branch points, the methods will be successful even on very large time scales. However these algorithms often falter when selection is absent.

Saturation

If one were to graph the number of DNA differences as function of time since speciation, it will eventually show a flattening as DNA differences lose their phylogenetic relevance because when saturation is reached the DNA differences reflect as much common ancestry as it does chance convergence. Neutral DNA will saturate faster than non-neutral DNA so the question is how long does it take for neutral DNA to become saturated?

So for recovering ancient common ancestry we need a long, slowly evolving gene since saturation is both a function of mutation rate and size of the genome involved. But such genes are unsuitable for measuring recent events. So just like radiometric dating we need genes of different sizes and different mutation rates.

Saturation can interfere with the ability to reconstruct a reliable phylogeny and one can estimate the time required for saturation to affect the phylogenetic reconstruction. For short genomes and high mutation rates, it is not surprising that phylogenetic reconstruction becomes an issue. Note that there is still a nested hierarchy just that this cannot be reconstructed from phylogenetic data. Some have confused the inability to use phylogenetic reconstruction with an actual absence of nested hierarchies.

saturation_lg.jpg

In Experimental Phylogeny of Neutrally Evolving DNA Sequences Generated by a Bifurcate Series of Nested Polymerase Chain Reactions Sanson et al describe an interesting experiment. They create a known phylogeny using a four step method leading to 16 terminal sequences. They then applied phylogenetic reconstruction methods to see how accurately they could reconstruct the true phylogeny.

Click for larger image

Fig. 1.–Evolution of DNA sequences by a series of bifurcate PCRs. An ancestor SSU rDNA cloned in pBluescript was used as template for series 1 of 70 nested PCR cycles with M13 primers. After the initial 35 cycles, reaction products were diluted 1:1,000 and used as templates for the subsequent 35 cycles, with rDNA primers RIBA and RIBB. After 70 cycles amplicons were cloned, and two clones were picked randomly and used as templates for the next series of nested PCR cycles. Lineages are propagated at random, and therefore the evolution is neutral and behaves as a stochastic process. Tree nodes T1 to T16 indicate terminal sequences, and 1.1 to 4.8, internal ancestors

The following picture describes the reconstruction and shows how the reconstructed phylogeny and the actual phylogenetic tree are identical.

Click for larger image

Fig. 3.–Comparison of real phylogeny with inferred maximum likelihood phylogeny (Felsenstein 1981 ; Posada and Crandall 1998 ; Swofford 1998 ). The serial PCR in vitro evolution resulted in the topology depicted (A) with varying branch lengths whose ancestors (1.1 to 4.8, circled) and terminal sequences (T1 to T16) were sequenced in full length. Scale bar indicates the number of cycles between tree internodes and nodes. The inferred phylogeny (B) has a topology identical to the real tree (A) and 9 out of 30 branch lengths were estimated correctly. Boxed numbers indicate branch lengths (number of substitutions), numbers in italics represent the percentage of a given cluster in 100 bootstrap replicates, with reestimation of parameters at each bootstrap replicate (top), and without reestimation at each replicate (bottom). Numbers below arrows indicate the estimated divergence (cycles ago), with the low-high confidence interval range (in parenthesis) as calculated by maximum likelihood quartet analysis (Rambaut and Bromham 1998 ). Numbers of substitutions, with corresponding standard errors, in the inferred tree (B) were calculated by multiplying the branch lengths (in substitutions per site) by the total number of positions (2,238 bp)

The authors then go on to reconstruct the ancestral states and present this interesting picture

Click for larger image

Fig. 4.–Comparison between ancestor sequences in the real phylogeny (fig. 3A ) and maximum likelihood reconstruction of ancestral states, HA. Only polymorphic positions are shown. Dots indicate residues identical to sequence 1.1. Numbers above the alignment indicate the position numbers. Numbers in parenthesis indicate the differences between the HA and their corresponding real ancestor sequences

Phylogenetic Inferences from Molecular Sequences: Review and Critique provides a good overview of the issues surrounding phylogenetic reconstruction and its problems and assumptions.

Another good paper is “EVOLUTION, WEIGHTING, AND PHYLOGENETIC UTILITY OF MITOCHONDRIAL GENE-SEQUENCES AND A COMPILATION OF CONSERVED POLYMERASE CHAIN-REACTION PRIMERS” by SIMON C, FRATI F, BECKENBACH A, CRESPI B, LIU H, FLOOK P in ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA

Abstract: DNA-sequence data from the mitochondrial genome are being used with increasing frequency to estimate phylogenetic relationships among animal taxa. The advantage to using DNA-sequence data is that many of the processes governing the evolution and inheritance of DNA are already understood. DNA data, however, do not guarantee the correct phylogenetic tree because of problems associated with shared ancestral polymorphisms and multiple substitutions at single nucleotide sites. Knowledge of evolutionary processes can be used to improve estimates of patterns of relationships and can help to assess the phylogenetic usefulness of individual genes and nucleotides. This article reviews molecular processes, discusses the correction of genetic distances and the weighting of DNA data, and provides an assessment of the phylogenetic usefulness of specific mitochondrial genes. The Appendix presents a compilation of conserved polymerase chain reaction primers that can be used to amplify virtually any gene in the mitochondrial genome. DNA data sets vary tremendously in degree of phylogenetic usefulness. Correction or weighting (or both) of DNA-sequence data based on level of variability can improve results in some cases. Gene choice is of critical importance. For studies of relationships among closely related species, the use of ribosomal genes can be problematic, whereas unconstrained sites in protein coding genes appear to have fewer problems. In addition, information from studies of amino acid substitutions in rapidly evolving genes may help to decipher close relationships. For intermediate levels of divergence where silent sites contain many multiple hits, amino acid changes can be useful for construction phylogenetic relationships. For deep levels of divergence, protein coding genes may be saturated at the amino acid level and highly conserved regions of ribosomal RNA and transfer RNA genes may be useful. Because of the arbitrariness of taxonomic categories, no sweeping generalizations can be made about the taxonomic rank at which particular genes are useful. As more DNA-sequence data accumulate, we will be able to gain an even better understanding of the way in which genes and species evolve.

10 Comments

On ARN Salvador argued: In other words, one can see a hierarchy even when one can’t infer a phylogeny. Well done, PvM. because I stated that

Note that there is still a nested hierarchy just that this cannot be reconstructed from phylogenetic data.

Let me explain: Phylogenetic reconstruction tries to reconstruct the history of the tree using statistical methods and models for evolutionary transition probabilities. There is however a risk that the amount of evolution has erased much hope of recovering any tree. This does of course not mean that such a tree does not exist, after all Darwinian processes inevitably lead to a tree structure, it means that statistical methods cannot reliably reconstruct the tree.

I hope that this resolved one of the many confusions Salvador seems to have about common descent, nested hierarchies and Darwinian evolution.

Biotic message of Nested hierarchies is the link to the ARN forum as is What is the twin nested hierarchy?

Salvador again misses the point

This is PvM obfuscation. What it means.

1. If evolution did not constrain the amount of variation, we could not see pylogenies and maybe not even hierarchies in the end.

2. When there is no selection pressure, the likelihood of not seeing a phylogeny or hierarchy in the end is even more remote.

In other words if we aren’t a little lucky, we won’t see a much of a pattern at all. There nothing in Darwinian evolution that requires any of the above listed conditions which PvM laid out above to ever happen.

The only thing it says is “if evolution evolves under the right conditions we can construct a phylogeny or a hierarchy, otherwise not.” Tautology!

‘under the right conditions’ means that the size of the genome studied and the mutation rates need to be reasonable. However it seems that evolution is indeed ‘slow’ enough for reconstruction to reliably happen although we do see evidence fo saturation as well. Selection pressure will alleviate some of these problems, luckily selection pressures are hardly that uncommon in evolution :-)

Of course, hierarchies would still exist, it’s just that we would be unable to reconstruct them from present day genetic data. If however we had someone record the history, we would clearly see a nested hierarchy. Which is not surprising given that nested hierarchies are expected from evolutionary principles.

PvM,

Can you clarify your meaning of hierarchies. We need to be sure we’re talking the same thing.

In creationist and most ID literature hierarchies are accepted to exist. That is, at the morphological level, there are always deep divisions in nature. We could classify things.

In my terminology, a phylogeny is not the same as a hierarchy. Is that the same as your terminology?

Respectfully, Salvador

Hi Salvardor,

A good question. A nested hierarchy for instance would be a family tree. Similarly a phylogeny forms a nested hierarchy. I am not sure what creationists mean by hierarchies but nested hierarchies are not an automatic outcome of a intelligently guided process, unless the process can be shown to be Markov-like for instance. Of course an intelligent designer could always make anything look like a nested hierarchy but that would seem an aweful lot of interventions and effort or reliance on the natural processes of evolution which can achieve a similar effect. Another example of hierarchies would be object based programming which inherits a class and extends or redefines particular routines.

Some examples

From Douglas J. Futuyma’s Science on Trial, (1982, 1995) p. 53:

“One of the many reasons for believing that organisms have a common evolutionary history is that their characteristics are often hierarchically arranged. Since evolution is supposed to proceed by a series of sequential splitting events, a new characteristic that evolves in one particular branch of the tree of life is likely to be passed on to all the descendants of that branch. Within this group, another new characteristic evolves, and is then passed on to the descendants of that particular species. Thus, for example, the four-legged condition evolved in amphibians, and is retained by most of their descendants. Among these, the ancestors of the mammals evolved a single-boned lower jaw. Among some of their descendants, the rodents developed gnawing incisors, and so on. There is a nesting of groups within groups, as a consequence of common ancestry. Objects like minerals that are not descended from common ancestors cannot be arranged in this way.”

From the website of the formidable New Mexicans for Science and Reason

or Hierarchies

or Lecture 31- Classification & Phylogeny

Some ID ‘interpretations’ of hierarchies without scientific evidence can be found

Interview with Ashby Camp

The point is that the hierarchical nature of life can be accommodated by creation theory as readily as by evolution.

or this obfucation Intelligent Design

The similarity of so many species, and ability to organize them in a reasonable way, is claimed as evidence of common descent. But we have already seen that communication systems and vehicles can be classified in a hierarchical order according to similarity. Therefore, similarity might just as well be the result of common design techniques as it is of a common ancestor.

sigh.…

What does Salvador mean by hierarchies?

A quick comment. On ARN Salvador asked if Avida would form nested hierarchies. RBH suggested that there are instance where Avida would not lead to nested hierarchies. It is important to differentiate between two separate issues: 1) One can reconstruct a nested hierarchy in Avida by following every single offspring 2) When looking at the leafs however it may not always be possible to reliably recover the phylogeny due to saturation or other processes which interfere with our ability to reconstuct the phylogeny.

In other words Avida still would generate a nested hierarchy but due to the lack of data we may not be able to reconstruct it without more detailed knowledge.

In this thread Salvador continues his confusion about Avida and nested hierarchies. Salvador basically confuses the concept of reconstruction (the ability to reconstruct a hierarchy) with the existence of such hierarchy.

Under certain circumstances Darwinian processes can saturate the data making a reconstruction impossible. For Avida due to the high mutation rates, this can be quite quickly but lets look at some commonly used genes:

Cytochrome c: theoretical lookback time 5000 Mya Pseudogenes: theoretical lookback time 45 Mya

Based on Doolittle 1986

Link

Enjoy

PvM,

A nested hierarchy for instance would be a family tree

Your definition of hierachy is not the same as the one I’m using. We are talking past each other.

A hierarchy in my usage is simply way of structuring existing objects such that there is a hierarchical relationship based on morphological traits, molecular sequence data, whatever. I don’t intend to make reference to any sort of ancestry when I make an observation of a hierarchy.

In my threads I point out that a hierachy (of existing creatures) is not the same as a phylogeny.

Creationists such as Linnaeus simply classified creatures with no assumption of phylogeny. He saw hierarchies the same way I do and the way Michael Denton does.

My definition of hierarchy is not the same as yours, PvM.

With that, I have little to add to your thread. that I haven’t already posted at ARN.

You’ll basically go uncontested on this thread from this point onward now that you clarified where you stand.

Respectfully, Salvador

salvador Wrote:

In my threads I point out that a hierachy (of existing creatures) is not the same as a phylogeny.

Creationists such as Linnaeus simply classified creatures with no assumption of phylogeny. He saw hierarchies the same way I do and the way Michael Denton does.

But it is consistent with phylogeny

You’ll basically go uncontested on this thread from this point onward now that you clarified where you stand.

How predictable, when Salvador is confronted with data he cannot stand he runs. Good job my friend. Looking for excuses to avoid dealing with the real issues.

Sigh

salvador Wrote:

In my threads I point out that a hierachy (of existing creatures) is not the same as a phylogeny.

Creationists such as Linnaeus simply classified creatures with no assumption of phylogeny. He saw hierarchies the same way I do and the way Michael Denton does.

You mean that your view of hierarchies is as erroneous as Denton’s misinterpretation of cytochrome-c data :-)

Back to hiding at ARN I guess where RBH can help you understand your flaws in reasoning.

Funny how ID proponents are so unable to defend their views and positions. Of course Salvador has an added problem, he is a (tentative) YEC-er… So much evidence that prove him wrong.

I admit I’m confused by the terminology as well. I understood a phylogeny to be the path of ancestry of an organism (or organ). In other words, that phylogenies can be used to construct cladograms. The important point is that the tree formed by this approach contains nothing but species, showing the branch points where species split into two or more.

The Linnaean heirarchy is something entirely different, referring to levels of similarity in organization. In the heirarcy, every level above the organism isn’t the ancestor of that organism, but rather a category including multiple organisms similar in some common characteristic(s).

So in a phlogeny, a father is an individual whose sons are branching events; a father is NOT simply a combinatoric category to conveniently combine all the sons into a single group.

What am I missing here?

About this Entry

This page contains a single entry by PvM published on July 24, 2004 12:04 AM.

Kennewick Man update was the previous entry in this blog.

Icons of ID: Markov processes and nested hierarchies is the next entry in this blog.

Find recent content on the main index or look in the archives to find all content.

Categories

Archives

Author Archives

Powered by Movable Type 4.381

Site Meter