GCG 2017 - overview of the big tree

From systematics

Jump to: navigation, search

OVERVIEW OF THE PHYLOGENY Santiago and Pedro introduce the 1300-tip phylogeny from most recent analyses


Previous GCG Phylogeny


  • 2150 individuals
  • 996 species
  • ETS, ITS, matK


Recovered 5 main clades, but not the caricoid clade Major subclades found in previous studies were recovered Taxonomic Disparity Index (TDI) used to discover mislabelled, misidentified, misplaced samples Non-monophyly of many sections largely due to homoplasy of selected diagnostic characters in Carex taxonomy

New GCG phylogeny


We are now to a total of 1510 species, 75% of all recognized species -- not all are seuqenced yet The most poorly sampled area are palaearctic, New Zealand, SE asia.

GCG releases

  1. Release 1: 2016 publication 960 spp
  2. Release 2: 224 spp (Pedro sequencing)
  3. Release 3: 250 spp (in progress)

Combining NCBI and GCG data

  1. GCG: ca. 8000 specimens
  2. NCBI: ca. 3400 specimens


Combine sequences for each marker

  1. Divided into three matrices to make alignment easier: Caricoid + OG + Siderostictae; core Carex; Vignea
  2. Align with muscle, manually revise
  3. Remerge using profile-provile option
  4. All-data matrix is the "all-data matrix", ca. 5000 sequences
  5. 62% missing data
  6. Used a scaffolding approach
    1. build a matrix using only accessions that have both ITS and ETS to build a reference tree
    2. add in additional sequences to build a "query tree" based on the reference tree, using the Evolutionary Placement Algorithm (EPA; Berger et al. 2011).


  1. round 1: examine 5001 accessions to get to 3897 accessions and :
    1. delete those with conflicting phylogenetic placement due to misID or lack of markers
    2. delete repeated accessions due to genBank duplicates of GCG sequences
    3. Rename misnamed accessions
    4. select best accession for each taxon with the one tip per species tree
    5. remaining questions marked in red, for review by experts
  2. round 2: pruned to 1 individual per species: 1238 species, all sections

Next steps

  1. propose a revised sectional classification
  2. reconstruct the evolution of important diagnostic morphological characters
  3. Calibrate tree
  4. reconstruct biogeographic evolution

Walk through the tree

  • Carex satsumensis -- placement next to Vignea makes a lot of sense. It looks like a unispicate indocarex with sessile lateral spikes, like a Vignea. Not only that, but C. gibba has fertile profiles, and so does C. satsumensis. [J.Starr] most reconstructions place it sister to Vignea [P.Jimenez]
  • C. bostrychostigma and C. dissitiflora (fertile prophylls on the latter) fall all off by themselves. They fall next to each other, which makes sense [P.Jimenez, Starr]

General conversation: we have a lot of clades with high degree of inflorescence variation, and we are probably going to have to get over the idea that sections are homogeneous in inflorescence structure.

Siderosticta clade

Problem is unresolved position of C. tsiangii (Surculosae) -- there are not a lot of Surculosae to begin with, and this sample also lacks data. Not clear how this section is separable from Siderostictae... a vegetative character? Maybe not a clear character to begin with.

Carex esculeriana is in Siderostictae but probably is actually a core Carex -- not sampled in our tree.

"Basal blush": Red plant bases only known from core Carex. But there is a basal red, different hue, a sort of pinky red, that occurs in Siderostictae. Carex kucyniakii has this character as well... the pinky red appears to be diagnostic for the Siderostictae (but not for Hemiscaposae). It can be miniscule, but somewhere down near the rhizomes you find it.

Schoenoxiphium clade

Mainly S.ern hemisphere, with one Mediterranean / S. African clade. Huge morphological variation in the clade.



Aciculares / Junciformes (Leucoglochin)

Unispicate clade

unresolved at the base: C. curticeps, C. himalaica, C. prainii

Kobresia clade

Well behaved unispicate clade




Filifoliae + Nardinae


Problematic characters:

  1. simple vs complex inflorescences
  2. creeping rhizomes


This clade is all east Asian... Carex unisexualis was placed in Holarrhenae only because it has long-creeping rhizomes.

Potosina + Carex rosea group of Phaestoglochin group

Perigynium adaxial surface cleanly divided into two zones

corrugated inner band of the leaf sheath clade

This clade also has the reddish spots, coppery inner surface of the leaf sheath... A number of spp in this clade are Mexican, and the sectional boundaries break down there as well.


Seems to fall wtih C. elongata, which it also does in Bruce's 2006 paper C. foetida only sampled in this tree


Good... we don't know about C. iljinii. C. balfourii looks to be a fine Ovales

Personal tools