The Last Universal Common Ancestor

Reconstructions of LUCA disagree by an order of magnitude, from 355 gene families to 2,657 proteins, and a survey of eight studies found none resembling any other. What all of them agree on is that LUCA already had a genetic code, ribosomes, a membrane and ATP synthase, which makes it a poor stand-in for a first cell.

Draft:  September 2026

"None of them showed a high or even moderate level of similarity with any other individual study." — a 2022 audit of eight LUCA reconstructions.1

A complex LUCA "provides strong evidence that this stage of evolutionary history was far removed from the origin of life." — Goldman and Becerra, stating the concession both camps have to make.2

Summary

LUCA is the organism from which every living thing descends.  It is not the first cell, and treating it as one is the commonest error made about it, by both camps.

  • The reconstructions disagree by an order of magnitude. 355 gene families in one study, 2,657 proteins in another, and a survey of eight studies found raw counts from 80 to 571 with "none of them showed a high or even moderate level of similarity with any other individual study."1
  • The disagreement is itself the finding, and an article that quotes only the largest LUCA is cherry-picking.
  • The methods have named weaknesses. Adding more homologs "rejects a simple LUCA hypothesis… for a majority of the previously identified LUCA proteins," about 82%.3 Peter Gogarten names a specific error in one reconstruction, where a known gene transfer between domains was read as LUCA's own gene.4
  • "Progenote" is misused in both directions. Gogarten, who has spent his career on this, argues LUCA was not a progenote: the progenote is "an earlier phase preceding LUCA."4
  • What all reconstructions agree on is the part that matters. LUCA had a genetic code, ribosomes, aminoacyl-tRNA synthetases, ATP synthase and a membrane. Whatever its gene count, it was a cell running the machinery this series is about.
  • And that cuts against using LUCA at all. If LUCA was complex, then a great deal of evolution stands between it and any prebiotic chemistry, and Goldman and Becerra say so in print.

The verdict.  LUCA marks the far end of the origin-of-life gap and nothing more.  The apologetics argument from LUCA's complexity mistakes it for a first cell, and the popular-science argument from an early complex LUCA quietly assumes a timeline that its own authors reached by discarding a different constraint.

Reconstructions of the last universal common ancestor are the other way to measure the far end, and the disagreement between them is itself the finding.

Weiss and colleagues got a spare 355-gene LUCA in 2016.5  Moody and colleagues got 2,657 proteins in 2024.6  A 2022 survey of eight studies found raw counts from 80 to 571 and concluded that "none of them showed a high or even moderate level of similarity with any other individual study."1

An article that quotes only the largest LUCA is cherry-picking.  This one reports the spread.

A specific caution about the word "progenote," because it is routinely misused in both directions.  Peter Gogarten — who has spent his career on this — does not argue that LUCA was a progenote.  He argues the reverse:  the modern view "recognizes that LUCA was a complex organism with an established coupling between geno- and phenotype, and depicts the progenote as an earlier phase preceding LUCA."4  He also names an error in the Weiss 2016 reconstruction — its phenylalanyl-tRNA synthetase "LUCA branch" reflects a known archaeal-to-Spirochete gene transfer — and states the methodological floor that limits all of this work:  "At present, reliable inference of genes to be present in LUCA is limited to genes with nearly universal distribution."4

There is also a concession that has to be printed:  Goldman and Becerra read the same data as evidence that LUCA is "far removed from the origin of life."2  LUCA is not the first cell.  Whatever it was, a great deal of evolution stands between it and any prebiotic chemistry, and arguments that treat LUCA's complexity as the first cell's complexity are wrong.

What remains after all those concessions is still the gap.  Whatever LUCA was, it already had a genetic code, ribosomes, a membrane and ATP synthase.  Between a 45-nucleotide ribozyme at 0.2% yield and a free-living cell with a megabase-scale genome, nobody can currently name a chemical route.  That is the claim this article defends, and it is deliberately smaller than the claim usually made.

The numbers, side by side

Study LUCA's size What it rests on
Weiss et al. 2016 355 protein families, about 0.1% of 286,514 clusters phylogenetic criteria on protein clusters
Crapitto et al. 2022, surveying eight studies raw predictions from 80 to 571; consensus 366 families re-analysis of the published sets
Moody et al. 2024 2.75 Mb encoding 2,657 proteins gene-tree/species-tree reconciliation with relaxed clocks

Set those in one table and the disagreement is not a nuance.  It is an order of magnitude on the central quantity, between studies published in strong journals within eight years of each other.

The dates are wider than the disagreement they are used to settle.  Moody's headline age for LUCA is about 4.2 billion years, from a preferred range of 4.09 to 4.33.  Their own composite estimate across models runs "from 3.94 Ga to 4.52 Ga," a spread of 580 million years — wider than the window the figure is normally quoted to constrain.6  A 2026 survey of the method calls molecular clock dating "challenging and highly uncertain."

What the same study also found is worth reporting, because it cuts against a simple picture in a different way.  Moody's LUCA carries "19 class 1 CRISPR–Cas effector protein families," which the authors read as evidence that "cellular life was already involved in an arms race with viruses at the time of LUCA."6  Whatever LUCA was, it was not the beginning of anything.

What every reconstruction agrees on

Strip out the disagreements and a shared core remains, and it is the part that matters for this series.  LUCA had a complete ribosome, the canonical genetic code, aminoacyl-tRNA synthetases, ATP synthase and a membrane.  It was, in the field's summary, a cellular organism with a full genetic system and complex metabolism.7

Two anomalies in that consensus deserve stating, because they are the parts a simple story cannot absorb.  There is no universal replicative DNA polymerase, which is why some workers infer an RNA or hybrid genome at that stage.  And there is no conserved phospholipid metabolism across the two domains, which is the long-standing lipid divide between bacteria and archaea.7

The claims, audited

Claim Verdict
"LUCA was the first cell" False, and both camps do it. A great deal of evolution separates them
"LUCA had 2,657 proteins" One study of several. The published range runs from 80 to 571 raw predictions
"The reconstructions converge" False. A survey of eight found none resembling any other
"LUCA was a progenote" Not what the field's specialist on the question says. He puts the progenote before LUCA
"LUCA's complexity shows life cannot have arisen naturally" Does not follow. It shows LUCA is a poor stand-in for a first cell
"LUCA was already running the code and the ribosome" Sound, and agreed across every reconstruction

Objections

1. You are using disagreement between studies as an argument

Only for the claim it supports:  that no specific gene count for LUCA should be quoted as settled.  The disagreement is documented by a survey inside the field, not asserted here.  What the reconstructions agree on is stated just as prominently.

2. A complex LUCA is good evidence for design

It is evidence that the ancestor of all life was already a working cell, which nobody disputes.  The design inference needs LUCA to be close to the origin, and the mainstream reading is the opposite.  The complexity of LUCA is an argument about how much unobserved evolution preceded it, not about whether that evolution happened.

3. An early LUCA compresses the time available for chemistry

That is the strongest version, and it has an internal problem.  The reconstruction giving the earliest date reaches it by rejecting the bombardment constraint that the compression argument usually also relies on.  Using both at once argues against itself, which is why the earliest-life article handles the timing question separately.

4. Molecular clocks are unreliable, so the dates mean nothing

The uncertainty is real and stated by the people who publish the dates.  What the clocks are used for here is a window rather than a date, and the window is wide.

What this article claims, and what it does not

It claims that LUCA reconstructions disagree by an order of magnitude, that the disagreement is documented inside the field, that the methods have named weaknesses, that "progenote" is misused, and that every reconstruction agrees LUCA already had the code, the ribosome, a membrane and ATP synthase.

It does not claim any particular gene count, or that LUCA's complexity supports design.  Weiss and colleagues' reply to Gogarten is paywalled past the first paragraph, so the attack is reported here without the defence.  The supplementary notes that would say which ribosomal and synthetase components the largest reconstruction fails to place in LUCA were not read.  Those gaps are listed in the literature review.

Sources

  1. 1.

    Crapitto, Andrew J. et al.  “A consensus view of the proteome of the last universal common ancestor.”  Ecology and Evolution 12:e8930.  2022.  “while some studies performed well on the inter-rater tests… none of them showed a high or even moderate level of similarity with any other individual study.”  A 2022 comparison of eight published reconstructions of LUCA's proteome, whose raw predicted counts run from 80 to 571, against a consensus of 366 protein families.  The authors measured how far each study agrees with the others and with that consensus.  Some studies agree well with the consensus, but no single study shares a high or even moderate degree of similarity with any other, and the authors conclude that the consensus gives the more accurate picture of LUCA's core proteome.  The article uses it for the disagreement between reconstructions and quotes its verdict on that point. ↩ ↩ ↩

  2. 2.

    Goldman, Aaron D., and Arturo Becerra.  “A New View of the Last Universal Common Ancestor.”  Journal of Molecular Evolution 92:659–661.  2024.  Page 660:  “This depiction of a physiologically and ecologically complex LUCA provides strong evidence that this stage of evolutionary history was far removed from the origin of life.”  The same commentary supplies the strongest version of the window argument, and it is not a creationist one:  Page 659:  “the 4.2 Ga dating, if correct, suggests that early evolution, from the origin of life to the LUCA, may have occurred within only 100–200 million years.”  On the reconstructions themselves, it says:  “despite this general consensus, the specific predictions of individual studies often diverge, rendering them somewhat unreliable (Crapitto et al. 2022).”  A short commentary, written for a special issue on LUCA, that discusses the 2024 study by Moody and colleagues.  It calls that study a significantly more rigorous approach to LUCA proteome reconstruction, and it reads the complex LUCA described there as far removed from the origin of life.  The article quotes it for that concession: a complex LUCA is not the first cell. ↩ ↩

  3. 3.

    Berkemer, Sarah J., and Shawn E. McGlynn.  “A New Analysis of Archaea-Bacteria Domain Separation: Variable Phylogenetic Distance and the Tempo of Early Evolution.”  Molecular Biology and Evolution 37(8):2332–2340.  2020.  Section Abstract:  “we report new research that suggests the clustering approaches used in this previous study undersampled protein families, resulting in incomplete phylogenetic trees which do not reflect protein family evolution.  Phylogenetic analysis of protein families which include more sequence homologs rejects a simple LUCA hypothesis based on phylogenetic separation of the bacterial and archaeal domains for a majority of the previously identified LUCA proteins (∼82%).”  A reanalysis of the protein families that Weiss and colleagues tied to LUCA in 2016.  It compares phylogenetic distances within and between the bacterial and archaeal domains for thousands of orthologous groups, and it argues that the earlier clustering undersampled protein families, which left the trees incomplete.  The article uses it as a methodological attack on one reconstruction, from inside the field, which is why it quotes no single gene count as settled. ↩

  4. 4.

    Gogarten, Johann Peter.  “The progenote, LUCA, and the root of the cellular tree of life.”  Handbook of Astrobiology.  CRC Press.  Edited by Vera M. Kolb.  2019.  Chapter 8.1, section 8.1.3 The Progenote Concept:  Woese and Fox's 1977 definition of a progenote is a stage.  “before a strict coupling between geno- and phenotype had emerged.”  Figures 8.1.1, panel A:  Panel B is contrasted with panel A in the caption to Figure 8.1.1, which calls panel A.  “the view popularized through the work of Carl Woese and Otto Kandler”.  In that view the three domains evolve independently from the progenote.  Figures 8.1.1, panel B:  Panel B depicts the progenote as.  “an earlier phase preceding LUCA”.  A chapter in a reference handbook of astrobiology, in which the author explains where the root of the tree of life lies and what the "progenote" means.  The chapter treats the progenote as an earlier phase preceding LUCA, a complex organism.  The article uses it to show that "progenote" is misused in both directions, because the author does not argue that LUCA was a progenote. ↩ ↩

    1. 4a.

      Section 8.1.4 Inferring LUCA's Properties:  On the Weiss 2016 result, the chapter describes the branch.  “identified in Weiss et al. for these gene families corresponds to a known transfer of the archaeal version of this enzyme to Spirochetes, whereas most of the bacterial pheRS genes were not included in the analysis.”  The general limit is stated as follows.  “At present, reliable inference of genes to be present in LUCA is limited to genes with nearly universal distribution.  Analysis of less conserved genes encoding functions in metabolism faces hurdles, such as the decay of phylogenetic information and difficulties to identify gene transfer events.”  That constraint applies to every LUCA reconstruction this article cites, including the ones it uses. ↩ ↩

  5. 5.

    Weiss, Madeline C. et al.  “The physiology and habitat of the last universal common ancestor.”  Nature Microbiology 1:16116.  2016.  “Among 286,514 protein clusters, we identified 355 protein families (∼0.1%) that trace to LUCA by phylogenetic criteria.”  The methodological attack on this paper is severe and comes from inside the field.  Berkemer and McGlynn, in Molecular Biology and Evolution 37 (2020): 2332–2340, report that adding more sequence homologs.  “rejects a simple LUCA hypothesis based on phylogenetic separation of the bacterial and archaeal domains for a majority of the previously identified LUCA proteins (∼82%).”  A genome survey of 6.1 million protein-coding genes from sequenced bacteria and archaea, sorted into 286,514 protein clusters, to find the genes that trace back to LUCA by phylogenetic criteria.  Its results describe LUCA as an anaerobic, CO2-fixing, hydrogen-dependent microbe with a Wood–Ljungdahl pathway, living in a hydrothermal setting.  The article uses its count of 355 gene families for the 2016 reconstruction of LUCA. ↩

  6. 6.

    Moody, Edmund R. R. et al.  “The nature of the last universal common ancestor and its impact on the early Earth system.”  Nature Ecology & Evolution 8:1654–1666.  2024.  “support for the presence of 19 class 1 CRISPR–Cas effector protein families in the genome of LUCA…  This is consistent with the idea that cellular life was already involved in an arms race with viruses at the time of LUCA.”  The honest caveat on molecular clocks is the authors' own.  Their composite age estimate for LUCA runs.  “from 3.94 Ga to 4.52 Ga”.  That is a 580-Myr spread, wider than the window the estimate is used to constrain.  The 4.09–4.33 Ga figure is the range from their preferred relaxed-clock models, not the total uncertainty.  A recent survey by Elizabeth A. Bell and Gregory P. Fournier, in Astrobiology 26 (2026): 633–643, calls the method.  “challenging and highly uncertain”.  A research article that dates LUCA with molecular clocks, using gene duplicates that arose before LUCA, and then reconstructs its genome by reconciling gene trees with a species tree.  It puts LUCA at about 4.2 billion years ago (4.09–4.33 Ga), with a genome of 2.75 Mb (million DNA bases; range 2.49–2.99 Mb) encoding 2,657 proteins (2,451–2,855) and a Wood–Ljungdahl carbon-fixation pathway.  The article uses its age, its protein count and its CRISPR evidence of an early immune system, and it uses the authors' own caveats about dating. ↩ ↩ ↩

  7. 7.

    Goldman, Aaron D. et al.  “Universal paralogs provide a window into evolution before the last universal common ancestor.”  Cell Genomics 6:101140.  2026.  A perspective, which is a review reporting two earlier LUCA reconstructions, not new results of its own.  It surveys universal paralogs, protein families duplicated before LUCA that keep traces of the evolution before it, and how new computing tools use them.  The article uses it for the shared core of LUCA's features, and for the two anomalies in that core: no universal replicative DNA polymerase and no conserved phospholipid metabolism across the two domains. ↩ ↩