Data from: Complex but clear allopolyploid pattern of subtribe Tussilagininae (Asteraceae: Senecioneae) revealed by robust phylogenomic evidence, with development of a novel homeolog-sorting pipeline

Polyploidy is a significant mechanism in eukaryotic evolution and is
particularly prevalent in the plant kingdom. However, our knowledge about
this phenomenon and its effects on evolution remains limited. A major
obstacle to the study of polyploidy is the great difficulty in untangling
the origins of allopolyploids. Due to the drastic genome changes and the
erosion of allopolyploidy signals caused by the combined effects of
hybridization and complex post-polyploid diploidization processes,
resolving the origins of allopolyploids has long been a challenging task.
Here we revisit this issue with the interesting case of subtribe
Tussilagininae (Asteraceae: Senecioneae) and by developing HomeoSorter, a
new pipeline for network inferences by phasing homeologs to parental
subgenomes. The pipeline is based on the basic idea of a previous study
but with major changes to address the scaling problem and implement some
new functions. With simulated data, we demonstrate that HomeoSorter works
efficiently on genome-scale data and has high accuracy in identifying
polyploid patterns and assigning homeologs. Using HomeoSorter, the maximum
pseudo-likelihood model of Phylonet, and genome-scale data, we further
address the complex origin of Tussilagininae, a speciose group (ca. 45
genera and 710 species) characterized by having high base chromosome
numbers (mainly x = 30, 40). In particular, the inferred patterns are
strongly supported by the chromosomal evidence. Tussilagininae is revealed
to comprise two large groups with successive allopolyploid origins:
Tussilagininae s.s. (mainly x = 30) and the Gynoxyoid group (x = 40). Two
allopolyploidy events first give rise to Tussilagininae s.s., with the
first event occurring between the ancestor of subtribe Senecioninae (x =
10) and a lineage (highly probably with x = 10) related to the
Brachyglottis alliance, and the resulting hybrid lineage crossing with the
ancestor of Chersodoma (x = 10) and leading to Tussilagininae s.s. Then,
after early diversification, the Central American group (mainly x = 30) of
Tussilagininae s.s., is involved in a third allopolyploidy event with,
again, the Chersodoma lineage and produces the Gynoxyoid group. Our study
highlights the value of HomeoSorter and the homeolog-sorting approach in
polyploid phylogenetics. With rich species diversity and clear
evolutionary patterns, Tussilagininae s.s. and the Gynoxyoid group are
also excellent models for future investigations of polyploidy.