Ohnologs arising from whole-genome duplication contributed to vertebrate brain cell-type evolution
Background: Duplicated genes and the evolution of brain complexity
A study published in Nature reports that duplicated genes derived from whole-genome duplication (WGD) early in vertebrate evolution contributed to the diversification of brain cell types. The researchers compared brain transcriptomes from humans, mice, lizards and lampreys with those of amphioxus, an invertebrate chordate. Their analysis supports the interpretation that these duplicated genes, known as ohnologs, helped shape cell-type identities and continued to be used in cell-type evolution long after the duplications occurred.
Genes can increase in number not only when an entire genome is duplicated, but also when smaller genomic regions are duplicated. Retained copies may divide the original roles between them or acquire new roles. However, how much each type of duplication contributed to vertebrate brain complexity remained unclear. A key challenge was also how to match cell types across distantly related species.
Key findings: Division of existing roles rather than acquisition of new functions
The researchers reanalyzed single-cell and single-nucleus RNA data from juvenile or adult brains of four vertebrate species and compared them with newly generated amphioxus brain data. The amphioxus analysis included 3,217 neurons and 2,351 non-neuronal cells. They used the cross-species mapping tool SAMap and conserved expression patterns of transcription factors, which regulate gene expression, to examine correspondences among cell-type groups.
Among vertebrates, transcription-factor programs defining major cell-type groups were conserved. By contrast, most amphioxus neurons did not correspond one-to-one with a specific vertebrate cell-type group. In all four vertebrate species, ohnologs were significantly overrepresented among marker genes distinguishing cell-type groups, whereas small-scale duplicate genes showed the opposite trend. Analyses accounting for differences in duplication timing also found a stronger association between ohnologs and marker genes.
The researchers also analyzed how expression patterns diverged. Compared with inferred ancestral states, about 78% of changes in marker use among the duplicate-gene groups analyzed were classified as subfunctionalization, in which duplicate copies divide existing roles, and about 21% as acquisition of new roles. These percentages do not refer specifically to ohnologs alone and are not direct measurements of biochemical function. The tendency for a particular copy to maintain predominant expression across multiple cell types was consistent with selection acting on the balance of gene-expression levels.
Mechanistic analyses focused on macroglia, which support neurons. In amphioxus glial cells, SoxE and Olig family transcription factors, which distinguish vertebrate macroglial subtypes, were co-expressed. The researchers interpreted this as evidence supporting a model in which an ancestral cellular regulatory system was partitioned after gene duplication. They also performed SoxE mutation experiments in amphioxus, but none of the expression comparisons involving the relevant transcription factors and glial markers shown in Figure 3h were statistically significant. These comparisons alone do not establish a functional effect of SoxE.
Implications and outlook: Continued use of ancient duplicates and limits on causal inference
Additional analyses of excitatory neurons in the cerebellar nuclei of humans, mice and chickens also found an association between ohnologs and cell-type-specific markers. Thirteen of 20 previously experimentally validated markers were ohnologs. This supports the interpretation that copies created by ancient gene duplications were also used in the differentiation of cell types that emerged much later.
However, an association with cell identity does not prove that ohnologs were required for the emergence of new cell types. Ancestral expression states were inferred from living species, and the methods may detect subfunctionalization more readily than acquisition of new functions. The researchers also considered that whole-genome duplication does not produce new cell types in every lineage and that other changes, including changes in gene regulatory networks, likely contributed as well.
The organisms studied, analyses, experimental results and interpretive limitations are described in the original Nature article.
Nature, Published online: 10 June 2026; doi:10.1038/s41586-026-10629-x Analyses of brain single-cell transcriptomes from human, mouse, lizard, lamprey and amphioxus reveal that duplicated genes (ohnologues) played a pivotal part in early vertebrate cell-type diversification.
Comparative neurobiology researchers can examine conserved transcription-factor programs and cell-type hierarchies together when matching cells across species, rather than relying on a single shared marker. For example, when comparing glial cells in mice and amphioxus, they could design analyses to examine whether differences in expression of a particular gene reflect different cell identities or division of roles between duplicated genes. The marker lists and analysis code that the paper indicates are publicly available can serve as starting points for revisiting candidate gene groups. Whether the repository can be run needs to be checked separately. This study did not demonstrate improvements in patient diagnostic accuracy, treatment efficacy or drug-development success rates; its immediate application is to the design of comparative analyses of cell-type evolution.