Signal conflict in mitochondrial genes of land plants

Using a codon-naive maximum likelihood (ML) and non-parametric bootstrap (BS) analysis of the data set used by Sousa et al. (PeerJ 8: e8995; 2020) in combination with consensus networks one can explore and visualise the character and gene support for the found Septaphyta clade.

As outlined in the related post, Can we dig too deep? (a copy is provided as PDF here), the inferred Septaphyta clade is probably a model-induced branching artefact ("short-branch culling") rejected by most genes and poorly supported by character splits in a few other genes. The earlier alternative, mosses as first-diverging lineage of land plants, likely represents a data/signal artefact related to the primitive sequence structure of their mitochondrial genes.

Overall, the analysis shows that mitochondrial genes provide little signal to resolve any deep splits in surviving land plants.

The xlsx spread sheet file tabulates the support of competing splits per gene. All files to reproduce the analysis and check the results are included in the 7z-archive.