Can this tree be replanted? II

 

Troubles with the Tree of Life

Mike Keas

Paul Nelson

Claims to know that an evolutionary Tree of Life (TOL) existed are increasingly problematic. The TOL is a picturesque way of imagining a branching pattern of universal common descent (UCD) — the alleged evolution of all current life by descent with modification from earlier life forms on earth, with all organisms tracing back to the Last Universal Common Ancestor (abbreviated as LUCA). We will update you on TOL troubles by analyzing a new paper authored by a group of biologists associated with the largest university in Latin America: the National Autonomous University of Mexico. 

Amadeo Estrada and his colleagues show how the huge literature devoted to constructing the TOL by means of large data sets of molecular sequencing (including many complete genomes of many organisms) is fraught with debilitating problems. Following up on earlier stern warnings from top TOL critics such as W. Ford Doolittle, Estrada et al. survey a wide field of molecular-based phylogenetic chaos — a bunch of mutually inconsistent accounts of the alleged branching pattern of evolution. They observe: “The strictly statistical approach [to molecular-based phylogenetic studies] … has resulted in divergent and even contradictory evolutionary hypotheses unsupported by independent evidence, between different research groups, and at times in single research groups.”1 Put more plainly, such studies have produced a large number of deeply inconsistent pictures of UCD, which undermines confidence about UCD itself. 

How Deep? How Severe?

Just how deep are these inconsistencies in the popular story of UCD? Molecular studies have produced radically different answers to what lies near the base of the TOL — that is, these studies have created confusion about LUCA. To grasp the severity of the situation, consider the wildly divergent range of recent opinion about LUCA; in the words of Estrada et al.(their key points numbered by us, with minor edits to English):

The LUCA has been characterized as

1. close to the origin of life (Koonin 2003; Weiss et al. 2016a), or as being far away from the origin of life (Mirkin et al. 2003; Delaye et al. 2005)
2. having a small genome (Koonin 2003), or as having a genome similar in size to many free living bacteria today (Kyrpides et al. 1999; Harris et al. 2003; Mirkin et al. 2003; Delaye et al. 2005; Yang et al. 2005; Ouzounis et al. 2006; Ranea et al. 2006; Becerra et al. 2014)
3. being autotrophic (Martin et al. 2008; Weiss et al. 2016a), or as being heterotrophic (Delaye et al. 2005; Becerra et al 2014, Muñoz-Velasco et al. 2018)
4. being hyperthermophilic (Woese 1987; Weiss et al. 2016a); or as being mesophilic (Galtier et al. 1999; Groussin et al. 2013; Cantine and Fournier 2018)
5. constituted by an RNA genome (Mushegian and Koonin 1996; Koonin 2003), or as having a DNA genome (Ouzounis et al. 2006; Delaye et al. 2005; Becerra et al. 2014)
6. being a simple cell (Koonin 2003), or as having a complex cell, similar to today’s bacteria (Kyrpides et al. 1999; Harris et al. 2003; Mirkin et al. 2003; Delaye et al. 2005; Yang et al. 2005; Ouzounis et al. 2006; Ranea et al. 2006; Becerra et al. 2014).2

Obviously, the evolution of the TOL could not have occurred in all of these mutually inconsistent ways (contradictory inferred stages of evolution near the base of the TOL). These evolutionary inferences are all over the biological map.Nonetheless, Estrada et al. do not extend their skepticism to the entire TOL-UCD paradigm. However, they do note:

New findings and changes in what we think about certain subjects are common in science. Nevertheless, we think that these extreme divergences between and even inside some researchers’ characterizations of the LCA are linked to the fact of relying on statistical approaches only without other kinds of data outside the sequence comparisons methods. In doing so, researchers can become subjects of contradictory algorithm results.3

As they hint here, Estrada et al. (building on Doolittle and others) propose their own way out of this mess, but admit that even their revised approach has its own additional troubles — though these are lesser troubles in their estimation (more on that below).

Scaling Back Claims

This research group recommends making estimations of LUCA that are less detail-rich, by scaling back claims of what we can legitimately know from molecular studies. They also urge fellow evolutionists to take into account more data beyond the confines of comparative molecular sequencing. Let’s explore the first prong of their dual revisionist research program first. 

On the one hand, they acknowledge that without molecular (especially genetic) sequence data, “there is no possibility of phylogenetic [TOL] reconstruction.” On the other hand, they point out the “serious epistemic disadvantages” of these studies “for the reconstruction of early forms of life, despite being rewarded in scientific practice.”4 Put bluntly, many scientists have advanced their careers by churning out shiploads of TOL claims. The relative ease these days of molecular sequencing and computer-aided statistical analysis make this bioinformatics research program hard to resist. When, however, this results in “contradictory hypotheses even within the same team and in consecutive publications, with no recognition of their divergent conclusions,”5 the internal coherence and logical consistency of the TOL / LUCA model inevitably suffer.

As Estrada et al. observe, over the past two decades Doolittle and others have attributed some of the molecular phylogenetic confusion to lateral (horizontal) gene transfer (LGT). By multiplying the possible modes of genetic transmission beyond so-called “vertical” inheritance, LGT greatly complicates tracing organismal lineages through subsequent generations. But most evolutionists do not think that LGT severely downgrades the historical signals that they use to determine the shape of the TOL. But, Estrada et al. note, “the problem is that there are major difficulties to measure LGT, not the least because the statistical criteria and bioinformatic tools used to estimate it share the same methodological constraints that plague phylogenetic reconstructions (Cortez et al. 2009).”6

A Candid Confession

That is a refreshingly candid confession. LGT is often cited as part of the reason why we get very differently shaped candidate TOLs (and different candidate LUCAs). But we should not doubt the overall TOL-UCD story in the face of such conflicting TOL reconstructions, we are told, because LGT is partly to blame for this situation.

This does not solve, however, the severe LUCA retrodictive inconsistencies that the authors bemoan in the large block quote above, listing six major evolutionary contradictions. Why? Because LGT estimates themselves depend upon, and are epistemically limited by, the “same methodological constraints that plague phylogenetic [TOL] reconstructions.”

We now turn to the second prong of this Mexican research group’s revisionist recommendations: the call to take into account more data beyond the confines of comparative molecular sequencing. “Any hypothesis of the LCA must be confronted with current empirical knowledge from the Earth sciences, as well as what scientists know about biochemistry and metabolic pathways….” This sounds wise, but then they immediately admit the severe limitations of this recommendation due to the “scarcity of biochemical and geochemical knowledge surrounding the early stages of life,” which “poses a severe epistemic constraint” on LCA theories.7

In an attempt to remedy this situation, they advocate scaling back LUCA retrodictions to what they call “a slimmer LCA.” This means that phylogenetic reconstructions should be “aiming at a lower resolution” — claiming we know a lot less — so that our claims are less likely to be contradictory or to be falsified by stubborn data. A skeptical philosopher of science could have said this about many branches of evolutionary biology. 

Kuhn’s Diagnosis

Sixty years ago, the historian and philosopher of science Thomas Kuhn listed what he described as the “symptoms” of a research field undergoing destabilizing change. Kuhn’s diagnosis stands as acutely relevant today as when he first offered it — especially the first symptom, which we have set in bold:

The proliferation of competing articulations, the willingness to try anything, the expression of explicit discontent, the recourse to philosophy and to debate over fundamentals, all these are symptoms of a transition from normal to extraordinary research.8

There is only one true history of life. (If you doubt this, ask yourself if you have, somewhere, an unknown set of biological parents with an equally valid claim to being your actual physical ancestors, when compared with the familiar names on your birth certificate.) Estrada et al. identify the competing historical articulations, only one of which can be the case, now current in evolutionary theory. A mature science converges on a single answer. A science in trouble? Not so much.

Notes

1. Amadeo  Estrada, Edna Suarez-Diaz, and Arturo Becerra, “Reconstructing the Last Common Ancestor: Epistemological and Empirical Challenges.” Acta Biotheoretica 70, no. 2 (2022): 1-18, p. 3.
2. Ibid., 3.
3. Ibid., 3.
4. Ibid., 4.
5. Ibid., 9.
6. Ibid., 6.
7. Ibid., 10.
8. T. S. Kuhn, The Structure of Scientific Revolutions (Chicago: University of Chicago Press, 2nd ed., 1970), p. 91.