Evolution News & Views
Eric Bapteste with ten other researchers across Europe and the United States are ready to provide a more “expansive” view of evolution that replaces Darwin’s tree with a “network” of life. Why is this necessary? Because “genetic data are not always tree-like.”
We’ve heard Bapteste criticize the tree of life before (see here and here). His new paper in Trends in Genetics, “Networks: Expanding Evolutionary Thinking” (see the summary at PhysOrg), seeks to “expand” evolutionary thinking by incorporating it within a larger “network” model. But if they replace the tree with a complex set of interconnections, what happens to the notion of universal common descent?
Down with Trees
The pro-network gang finds trees inadequate on several grounds. For one, a tree diagram is too simplistic:
However, many patterns in these data cannot be represented accurately by a tree. The evolution of genes in viruses and prokaryotes, of genomes in all organisms, and the inevitable noise that creeps into phylogenetic estimations, will all create patterns far more complicated than those portrayed by a simple tree diagram. Genetic restructuring and non-vertical transmission are largely overlooked by a methodological preference for phylogenetic trees and a deep-rooted expectation of tree-like evolution. (Emphasis added.)
Interesting: trees are an “expectation” and a “preference” — more on that later. Another problem is that trees, as subsets of networks, stifle thinking:
A way forward is to recognize that, mathematically, tree graphs are a subset of the broader space of general graphs (henceforth: networks). Trees are optimized, pared-down visualizations of often more complex signals. When confined to trees, we overlook additional dimensions of information in the data. By moving beyond the exclusive use of trees, and adopting a routine application of networks to genetic data, we can expand the scope of our evolutionary thinking.
Interesting: they still want “evolutionary thinking,” but what kind without trees? Another problem is that much of the genetic data is not tree-like:
Evolutionary networks today are most often used for population genetics, investigating hybridization in plants, or the lateral transmission of genes, especially in viruses and prokaryotes. However, the more we learn about genomes the less tree-like we find their evolutionary history to be, both in terms of the genetic components of species and occasionally of the species themselves.
Interesting: if “evolutionary history” is not tree-like, does universal common ancestry still hold? They explain that many patterns are mosaic-like rather than tree-like due to a number of non-vertical processes. What predominates are “reticulate” (net-like) relationships. Another problem is that tree diagrams are often inaccurate:
Tree-based genomic analysis is proving to be an accuracy challenge for the evolutionary biology community, and although genome-scale data carry the promise of fascinating insights into treelike processes, non-treelike processes are commonly observed.
Further, tree diagrams are often contradictory:
There are long-standing controversies regarding the evolutionary history of many taxonomic groups, and it has been expected by the community that genome-scale data will end these debates. However, to date none of the controversies has been adequately resolved as an unambiguous tree-like genealogical history using genome data. This is because quantity of data has never been a satisfactory substitute for quality of analysis. Many of the underlying data patterns are not tree-like at all, and the use of a tree model for interpretation will oversimplify a complex reticulate evolutionary process.
Interesting: how does a “reticulate evolutionary process” square with universal common descent? They give examples: the yeast phylogenetic data can only be force-fit into a tree, but then, “a species tree becomes only a mathematical average estimate of evolutionary history, and even if it is supported it suppresses conflicting phylogenetic signals.” It’s misleading, in other words.
Another example is the tree of placental mammals: “a problem that has been difficult to resolve as a bifurcating process because different genetic datasets support different trees.” Wriggling out of the tree-thinking straitjacket can resolve these controversies: “the network provides biological explanations that go beyond what can be accommodated by a simple tree model.”
Up with Networks
The team believes that network theory has matured to the point where it can be a valuable tool for biologists. It also promises job opportunities: “The further improvement of networks for evolutionary biology offers many outstanding opportunities for mathematicians, statisticians, and computer scientists.”
A network can be both a more parsimonious description of the amount of discordance between genes, and a starting point for generating hypotheses to explain that discordance.
Trees, Networks, and Scientific Explanation
The authors recognize that network-thinking is not a panacea. Biologists will still need to “interpret” the findings correctly:
However, biologists must also keep in mind that networks are not yet free of interpretive challenges. One must knowledgeably select from the various types of network methods available to interpret properly such features as internal nodes and the meaning of taxon groupings, which differ in important ways among methods. Furthermore, community standards do not yet exist for network assessment and interpretation. As with tree methods, the responsibility remains with the researcher to understand network methodology, apply it correctly, and make valid inferences.
Philosophers could have fun with this paragraph. It has the potential for investigator bias at each stage. It sounds like Finagle’s First Law: “To study a subject best, understand it thoroughly before you start” — i.e., know what the valid inferences are before you infer anything; know the right methods before you choose which method is right; and if all else fails, trust the consensus (community standards). But that’s predictable; they are, after all, still Darwinian evolutionists. What matters is the extreme paradigm shift this represents.
Calling it “historic,” the authors recognize the extent of the shift they are proposing:
These challenges do not detract from the fact that networks represent an historic juncture in the development of evolutionary biology: it is a shift away from strict tree-thinking to a more expansive view of what is possible in the development of genes, genomes, and organisms through time.
They use “development… through time” as a synonym for evolution. But what kind of evolution? If it is not tree-like, what is it? In a network diagram, common descent gets scrambled if one accepts “random lateral gene transfer” and “hybridization” as key processes, as these authors do. In fact, they say nothing about natural selection. The new picture is of interconnected nodes, with no clear progression from simple to complex. After all, a gene has to already exist to be laterally transferred. Two species must already exist in order to hybridize. There’s nothing here about a beginning and a progression. It’s all about relationships between nodes that could have (avoiding tree-thinking) been in existence all along. The sample network diagram in the PhysOrg article shows lines going up, down, and sideways between nodes. It claims that “Moving from tree-like depictions of evolution to network diagrams is an effective way to amend the Tree of Life without dismissing it,” but the move turns the tree upside down and inside out. The focus is on nodes and relationships — not progression.
Even the strictest creationists allow for “change over time” in terms of new interconnections and horizontal modifications among existing kinds of organisms. There’s nothing really Darwinian about Bapteste’s proposal. It could even be considered ID-friendly: pre-existing intelligently designed organisms change their relationships through time, occasionally sharing genetic information. By “expanding” the tree of life, this team is demolishing it. Bifurcating trees within network diagrams vanish as artifacts, like imagined faces in a bumpy ceiling when one backs away and sees the whole.
Trees as Dogma
Back to the complaint of Bapteste et al. that tree-thinking is an “expectation” and a “preference” – i.e., a set of assumptions chosen before the data has a chance to speak. Their opening paragraph shows that Darwinian evolutionists produce trees because they are trained to produce them:
Ever since Darwin, a phylogenetic tree has been the principal tool for the presentation and study of evolutionary relationship among species. A familiar sight to biologists, the bifurcating tree has been used to provide evidence about the evolutionary history of individual genes as well as about the origin and diversification of many lineages of eukaryotic organisms. Community standards for the selection and assessment of phylogenetic trees are well developed and widely accepted. The tree diagram itself is ingrained in our research culture, our training, and our textbooks. It currently dominates the recognition and interpretation of patterns in genetic data.
What they are saying is that this dominant way of looking at the data is both ingrained as a method, and also used to provide evidence for evolution! That’s circular. They are trained to think in terms of evolutionary trees, and then use evolutionary trees as evidence for evolutionary trees.
One can only welcome this paper’s bold proposal to overturn entrenched dogma and offer a more “expansive” view of “development…through time.” For one thing, if trees are artifacts emerging from expectations, they should be exposed as such. For another, the “network” diagram seems conducive to ID research inasmuch as it calls into question universal common ancestry via natural selection (i.e., neo-Darwinism), and seeks to portray the evidence honestly.
Their paper is the product of a meeting in Leiden last October called “The Future of Phylogenetic Networks.” It’s too soon to tell if Darwin security forces will let this band of independent thinkers gather a following. If nothing else, it shows (notwithstanding the insistences of the National Center for Science Education) that insiders know about the fundamental controversies in evolutionary theory, and are calling for some of the same reforms that advocates of intelligent design do.