| Abstract Detail
Systematics Section / ASPT Rai, Hardeep [1], Graham, Sean W. [1]. Inference and misinference of the overall pattern of seed-plant phylogeny from plastid data. We do not know what the extant sister group is for angiosperms, conifers, cycads, Ginkgo or Gnetales; multiple well-supported solutions are recovered using different gene and genomic combinations, diverse taxon samplings, and so on. Here we evaluate these major hypotheses, by focusing on the position of Gnetales, and by using real and simulated data derived from a large and data-dense plastid matrix (15-17 conservative protein coding genes from 30 seed plants, including 19 exemplar conifers, and multiple outgroups). For different data subpartitions we infer strongly conflicting relationships, but mostly we see Gnetales as sister to other seed plants, or infer a clade of conifers and Gnetales in some combination. No straightforward sets of patterns are observed when data are partitioned into ‘fast’ vs. ‘slow’ predicted rates. For example, ML analyses of the two fastest and five slowest of nine rate categories provide strong to moderate support for Gnetales as sister to other seed plants, but two moderately rapid rate classes strongly support a clade with Gnetales and conifers. The ‘GnePine hypothesis’ is inferred in several cases, but is rarely well supported (an exception is MP analysis of the first two codon positions). We never recover the ‘anthophyte hypothesis’ (angiosperms plus Gnetales). Our simulations predict that removal of the fastest rate categories improves the accuracy of MP but not ML. They also show that ML is generally better able to deal with the negative impact of inclusion of the third codon position, but that both inference methods are strongly biased using this position alone. In contrast to the other putative relationships considered, we predict that the anthophyte hypothesis, if correct, should often lead to ML inference of Gnetales as sister to other seed plants (i.e., an hypothesis actually observed in many of our analyses of real plastid data). Log in to add this item to your schedule
1 - University of British Columbia, Botanical Garden And Centre For Plant Research, 6804 Sw Marine Drive, Vancouver, British Columbia, V6T 1Z4, Canada
Keywords: Seed plant Conifers Gnetales GnePine Anthophyte Rate classes Simulations plastid genome.
Presentation Type: Oral Paper:Papers for Sections Session: CP42 Location: Boulevard C/Hilton Date: Wednesday, July 11th, 2007 Time: 8:15 AM Number: CP42002 Abstract ID:2254 |