New placental genome study fails again and again

Gupta, Mirarab and Turakhia 2025
just published a genomic study of placental taxa in which the presumed, traditional and tested outgroup clade, Marsupialia = Metatheria is not mentioned.

Even the term ‘outgroup’ is not found in the text.

Like other genomic studies
this one recovered clades not recovered in the large reptile tree (LRT, 2338 taxa) a trait-based study that importantly includes fossil taxa omitted from the GM&T study.

If genomic results are to be judged as accurate, they should reflect standard trait-based studies. This one does not. Omitting fossils is an inherent sin in taxonomy.

GM&T had high hopes.
They wrote, “These genomes hold the promise to resolve long-standing questions surrounding the evolutionary relationships of species (species trees) and illuminate differences in evolutionary history across the genome (gene trees).”

Alas, the authors have no interest in the results (Fig 1, see below), other than to publish them without critical analysis.

Alas, the authors bought into the hope and promise of deep time genes, not caring that their results were too often untenable (Fig 1). Judge for yourself, as always. Run your own tests comparing traits to genes.

Lacking fossil and marsupial outgroup taxa
the authors had no idea that now four origins for placental taxa arise from various marsupial outgroup taxa.

Genomic taxonomy has to come to grips with this reality before proceeding with the next study. They are creating and promoting myths.

The cladogram from GM&T
(Fig 1) bears almost no resemblance to the LRT, which includes fossils and outgroup taxa.

From GM&T caption figure 2b here rotated to fit a narrow column: “Order-level trees of 240 placental mammals estimated by ROADIES (on the Right) and the reference tree from the Zoonomia consortium (8) (on the Left). Dashed branches show the differences between the two trees.” ” data-image-caption=”

From GM&T caption figure 2b here rotated to fit a narrow column: “Order-level trees of 240 placental mammals estimated by ROADIES (on the Right) and the reference tree from the Zoonomia consortium (8) (on the Left). Dashed branches show the differences between the two trees.”

” data-medium-file=”https://pterosaurheresies.wordpress.com/wp-content/uploads/2025/05/gmtcladogram.genomic588.jpg?w=174″ data-large-file=”https://pterosaurheresies.wordpress.com/wp-content/uploads/2025/05/gmtcladogram.genomic588.jpg?w=584″ class=”size-full wp-image-92733″ src=”https://pterosaurheresies.wordpress.com/wp-content/uploads/2025/05/gmtcladogram.genomic588.jpg” alt=”From GM&T caption figure 2b here rotated to fit a narrow column: “Order-level trees of 240 placental mammals estimated by ROADIES (on the Right) and the reference tree from the Zoonomia consortium (8) (on the Left). Dashed branches show the differences between the two trees.”" width=”584″ height=”1007″ srcset=”https://pterosaurheresies.wordpress.com/wp-content/uploads/2025/05/gmtcladogram.genomic588.jpg?w=584&h=1007 584w, https://pterosaurheresies.wordpress.com/wp-content/uploads/2025/05/gmtcladogram.genomic588.jpg?w=87&h=150 87w, https://pterosaurheresies.wordpress.com/wp-content/uploads/2025/05/gmtcladogram.genomic588.jpg?w=174&h=300 174w, https://pterosaurheresies.wordpress.com/wp-content/uploads/2025/05/gmtcladogram.genomic588.jpg 588w” sizes=”(max-width: 584px) 100vw, 584px” />

From GM&T caption figure 2b here rotated to fit a narrow column: “Order-level trees of 240 placental mammals estimated by ROADIES (on the Right – now top) and the reference tree from the Zoonomia consortium (8) (on the Left – now bottom). Dashed branches show the differences between the two trees.”

From left to right  the GM&T cladogram
(Fig 1) nests toothless pangolins with cats, seals and bears. And these with horses and rhinos. Those pairings are not supported by the LRT.

The next step to the right
nests whales (without splitting mysticetes from odontocetes) with artiodactyls. Mysticetes do nest with hippos, so THAT is supported by the LRT… and these with horses and rhinos. THAT is also supported by the LRT, but as highly derive taxa, not basal to cats, pangolins, etc.

Odontocetes arise from tenrecs in the LRT, so they have to be considered separately.

The next step to the right
nests bats basal to all the above taxa. That is NOT supported in the LRT. In the LRT bats are primates.

Basal to this clade = the next step to the right
nests shrews, moles and hedgehogs basal to all the above taxa. In the LRT these taxa are basal only to odontocetes via sengis, tenrecs and pakicetids = giant tenrecs.

The next clade to the right
nests colugos with primates. THAT is supported by the LRT where colugos are primates.

The next clade to the right
nests tree shrews with primates. That is NOT supported by the LRT where these two clades arise from different marsupials.

The next clade to the right
nests rabbits with rodents. THAT is supported by the LRT. Tree shrews are related to these two.

The next clade to the right starts over with
golden moles nesting with tenrecs. This is NOT supported by the LRT.

The next clade to the right
nests aardvarks with sengis. THAT is supported by the LRT and not far from tenrecs.

The next clade to the right
nests elephants with rock hyraxes. This is NOT supported by the LRT.

The final clade to the far right
nests sloths + anteaters with armadillos. In the LRT anteaters nest with armadillos but apart from sloths and glyptodonts. Anteaters and armadillos have tube snouts that nest them with tube-snouted pangolins on the far left of the GM&T cladogram.

Note the lack of an outgroup marsupial or marsupials (Fig 1) essential to anchor these several placental clades.

No one knows why
gene studies don’t more closely match trait studies (other than omitting fossils). Let’s work on that. Meanwhile, let’s eliminate deep time genomic studies from publication until they do match. Promoting inaccurate models for evolution is not good science. Sorry. That’s just the way it is at present.

References
Gupta A, Mirarab S and Turakhia Y 2025. Accurate, scalable, and fully automated inference of species trees from raw genome assemblies using ROADIES. PNAS Biophysics and Computational Biology 122(19): e2500553122