Statistical Phylogenetics: An Application to Corbiculate Bees and Geography

1
Statistical Phylogenetics An Application to
Corbiculate Bees and Geography

• Amrish Deshmukh
• Ryan Starski

2
Phylogenetics

• Investigate evolutionary relationships using
  similarity in DNA, behavioral, or physical
  characteristics
• Methods
• Parsimony supports the hypothesis (tree) which
  requires the minimum number of changes and
  homoplasies
• Likelihood supports the hypothesis which is most
  statistically likely given the assumptions of an
  evolutionary model
• Bayesian supports the distribution of hypotheses
  which would most likely produce the data

3
Parsimony

• Edwards and Cavalli-Sforza(1963)
• Evaluation of parsimony via union-intersect
  method
• Limitation long branch attraction
• Application Morphological or behavioral
  characters

4
Likelihood Models

• Molecular evolution is reasonably well modeled by
  finite state continuous time Markov chains.
• The evolution of a single site is modeled over
  the state space of an n-letter alphabet.
• Assuming independence of sites implies the
  likelihood, L, can be put in terms of the
  products of likelihoods for each site
• Or

5
Corbiculate Bees

• Include economically important species
• Form complex eusocial societies
• Family Apidae
• Subfamily Apinae
• Monophyletic group
• Comprised of 4 monophyletic tribes

6
Euglossini Orchid bee

• Pollinate orchids by collecting chemicals on
  their legs
• Do not possess eusocial behavior
• Most are solitary

http//en.wikipedia.org/wiki/ImageMormodes_buccin
ator_-_infl_3.jpg
7
Bombini Bumble bee

• Important for crop and wildflower pollination
• Form small social colonies

http//farm2.static.flickr.com/1317/531377215_44c4
001e7c.jpg?v0
8
Apini Honey bee

• Only 6-11 species
• Appear in the fossil record 30 million years ago
• Highly eusocial

http//en.wikipedia.org/wiki/ImageHoneybee_therma
l_defence01.jpg
9
Meliponini Stingless bee

• Minimal honey production
• Are highly eusocial with large colonies
• Bite, defend in numbers

http//www.innovations-report.com/bilder_neu/30264
_bstingless_bee2.jpg
10
Contradictory Data Sets

• Behavioral\Morphological
• Noll(2000) Analyzed 42 behavioral characters
• Predicted single origin of eusocial behavior
• Placed (Apini, Meliponini) in a monophyletic
  clade
• Molecular
• Cameron(2001) Analyzed 2 mitochondrial and 2
  nuclear data sets
• Predicted a dual origin of behavior
• Kawakita(2008) Investigated 12 nuclear genes.
  Confirmed Camerons results

11
Our Data set

• 12 genes investigated by Kawakita
• 15 taxa
• Outgroup members of the genus Centris

12
Multi-gene analysis

• Concatenation
• Completely links all considered genes
• Consensus
• Completely unlinks all considered genes
• Assumes a prior distribution of gene trees which
  is independent of the species tree, contradictory
  to the purpose of using the genes
• Coalescence

13
Coalescence

• A concern when molecular evolution is used to
  investigate species relationships
• The gene tree obtained by sampling individuals
  within a species may be inconsistent with the
  species tree
• A large ancestral population weakens the
  correspondence between gene and species
  coalescent times.

14
A three-species phylogeny

• Gene tree predicts the
• Incorrect species tree if
• 1. A and B do not coalesce
• before the common ancestor
• with C
• 2. A and B are not the first to
• coalesce
• The net probability

15
Bayesian Estimation of Species Trees
16
BEST Assumptions

• Random mating
• No gene transfer after species divergence
• No recombination within a locus

17
Methods

• Use simple parsimony analysis to investigate
  individual behavioral characters
• Use Mr. Bayes to investigate the 12 genes
  individually
• Concatenation
• B.E.S.T.

18
Results Behavioral Analysis
19
Consensus
20
Partitions

• Each star represents a taxon
• Each row represents a partition
• Each combination in rows a percentage of
  occurances

21
Result Kawakita Comparisons
22
Result Concatenation Analysis
23
ResultBEST Analysis
24
Limitations and Future Directions

• Complete genome data
• Multiple Individual data
• Social structure may affect mating

25
Phylogeography Hepatitis

• Complete Genome, 58 Taxa with spatiotemporal data
• Phylogenetics with relation to geographic disease
  spread.
• PoY and Supramap (Dr. Janies)

26
Phylogeography Variola

• A nuclear gene coding for NTPase
• 48 taxa, with spatiotemporal data

27
Thanks

• The MBI
• Dr. Pearl
• Dr. Pan
• Dr. Janies
• Dr. Wenzel

28
References

• Noll, Fernando B. "Behavioral Phylogeny of
  Corbiculate Apidae (Hymenoptera Apinae), with
  Special Reference to Social Behavior." Cladistics
  18 (2002) 137-153.
• Winston, Mark L. and Charles D. Michener. "Dual
  origin of highly social behavior among bees."
  Proc. National Academy of Science 74.3 (Mar
  1977) 1135-1137.
• Cameron, Sydney A. and Patrick Mardulyn.
  "Multiple Molecular Data Sets Suggest Independent
  Origins of Highly Eusocial Behavior in Bees
  (HymenopteraApinae)." Systematic Biology 50.2
  (Apr 2001) 194-214.
• Felsenstein, Joseph. Inferring Phylogenies. 2004.
• Kawakita, Atsushi, John Ascher, and Teiji Sota.
  "Phylogenetic analysis of the corbiculate bee
  tribes based on 12 nuclear protein-coding genes."
  Apidologie 39(2008) 163-175.