New genomic estimate misses monotreme-marsupial split by 43 million years

Summary for those in a hurry:
Fossils provide hard evidence. Deep time gene studies provide estimates and false positives too often to trust them.

Zhou et al. 2021 report:
“Our phylogenomic reconstruction shows that monotremes diverged from therians around 187 million years ago, and the two monotremes diverged around 55 million years ago. This estimate provides a date for the monotreme–therian split that is earlier than previous estimates (about 21 million years ago, but agrees with recent analyses of few genes and fossil evidence.”

Let’s stop putting our faith in estimates derived from genomic deep time studies that have proven themselves to be wrong too many times. Here, the Zhou et al. estimate is at least 43 million years too late (Fig. 2) based on Brasilitherium (Fig. 3) fossils and the tree topology recovered by the LRT (Fig. 1).

Figure 7. Subset of the LRT focusing on Metatheria (marsupials) including Paedotherium and Adalatherium.

Figure 1. Subset of the LRT focusing on Metatheria (marsupials) including Paedotherium and Adalatherium.

By contrast with Zhou et al. 
Morganucodon (Late Triassic, 205mya, Fig. 4) is a basal marsupial in the large reptile tree (LRT, 1790+ taxa; subset Fig. 1) based on phenomic (= trait) analysis that includes fossil taxa. Genomic tests are infamous for false positives when dealing with deep time taxa.

Figure 1. Select basal cynodonts and mammals set chronologically. The divergence times for placentals (Eutheria), marsupials (Metatheria) and monotremes (Mammalia) are estimated here.

Figure 2. Select basal cynodonts and mammals set chronologically. The divergence times for placentals (Eutheria), marsupials (Metatheria) and monotremes (Mammalia) are estimated here.

Brasilitherium,
(Figs. 3, 4) from the Early Norian, Late Triassic, 225mya, is a derived monotreme in the LRT. That means it lived AFTER the monotreme-therian split which must have occurred at least 230mya.

Figure 1. Brasilodon compared to Kuehneotherium, Akidolestes and Ornithorhynchus, the living platypus.

Figure 3. Brasilodon compared to Kuehneotherium, Akidolestes and Ornithorhynchus, the living platypus.

As everyone knows
the platypus and echidna are highly derived monotremes. Megazostrodon (Fig. 4) is the last common ancestor (LCA) of all monotremes and all mammals. Megazostrodon was a Late Jurassic late survivor of that earlier (Middle Triassic?) radiation.

Figure 5. Basal mammals and their proximal ancestors. Here taxa below Megazostrodon are mammals. Those above are not. Hadrocodium is uniquely reduced, but this occurs within the Mammalia.  The dual jaw joint was tentatively present in Pachygenelus.

Figure 4. Basal mammals and their proximal ancestors. Here taxa below Megazostrodon are mammals. Those above are not. Hadrocodium is uniquely reduced, but this occurs within the Mammalia.  The dual jaw joint was tentatively present in Pachygenelus.

According to the LRT,
there was no gradual ascent of monotremes leading to marsupials. Rather the monotreme-marsupial split occurred at the origin of mammals and monotremes. How this affects the genes for lactation discussed in the Zhou et al. paper is beyond the scope of this blogpost.

The purpose here
is to emphasize the importance of a broad, proper and valid phylogenetic context before proceeding to the narrow focus of your interests. 42 co-authors using cutting edge genomic techniques hobbled their otherwise excellent and in-depth report by skipping step number one.


References
Zhou Z et al. (41 co-authors) 2021. Platypus and echidna genomes reveal mammalian biology and evolution. Nature https://doi.org/10.1038/s41586-020-03039-0