Julius B. Lucks/Meetings and Notes/SMBE2007/lateral gene transfer the ring web or network of life

James McInerny : What does it mean when we make a Tree of life from genomes?

Wed Jun 27 12:36:22 EDT 2007

Words to Look Up

• Robertson Foulds distance (tree distance method)

Talk

• early 80s - Carl Woese added archea via 16/18S rRNA trees
• 1995 - H. Influenzae sequenced - shortly later 2 strains of E. coli
  • 2 strains of E. coli - same species - 25% different genome sizes
• differences in phylogenetic trees when use 16S rRNA or pseudouridine tRNA
  • lots of lateral transfer of genes
• Ford Doolittle - Nov 1999 - Science
  • 'most archael and bact genes contain genes from multiple sources'
  • 'if lgt can be limited to special categories of gene, then no heirarchical classification can be taken as natural'
• Darwin - 'a classification based on any single character, however important that may be, has always failed.'
• look at all the trees - do they agree or not?
• supertree approach
  • amalgamate information from disparate data sources (extension of consensus approach)
  • explore congruence across datasets
  • identify datasets w/ unusual features
  • Clann software
    • input trees derived from orthologs - diff sets of leaves
    • candidate supertree
    • tree to tree distance between input tree and supertree
      • prune supertree to have same set of leaves until has same set of leaves as input, then look at distance
    • score (sum distances) supertree (0 if perf compatible with all input trees, > 0 otherwise)
    • search over supertrees
    • do a permutation to create random dataset - permutation test
• gamma proteobacteria - compatibility of each of input trees better than random
  • can have a meaningful supertree
• deep-level prokaryotic - compatibility of each of input trees NOT better than random
  • input trees could be very bad
  • bad ortholog detection
  • horizontal transfer so pervasive that phylogenetic relationship completely eroded
• Choi, PNAS, 104, 4489, 2007 - Global extent of horizontal gene transfer.
• diff methods for same data give diff trees - method problem
• see where eukaryote fitted on prokaryote tree - use supertree method
  • eukaryotes went with cyanobacteria
    • cyanobacteria endosymbiosis more recent than mitochondrial
  • removed - went with alpha-proteobacteria (mitochondrial endosymbiosis)
  • removed - went in middle of archaea (thermoplasmatalis)
• bacteria and archaea not monopdylecit
• eukaryo are monophyletic, but have 3 parents (cyanobacteria, alpho-proteobacteria, archaea)

Ford Doolittle : Not much - testing the tree of life hypothesis

Wed Jun 27 13:00:29 EDT 2007

Talk

• 1st slide is a tree stump
• new title: 'The practice of classification and the theory of evolution, and what the demise of Darwins Tree of Life hypothesis means for both of them'
• outline
  • classification
  • evolution as a theory
  • TOL specific version of an evol theory - how TOL hypoth untrue
  • implications of this for practices of classification
• classification
  • natural and useful heman practice
  • 'our job as sytematists is not to discover wahat an organism truly is, but rather to decide what we should call it' - Baptiste and Doolittle
  • inclusive heirarchies
    • arbitrary and negotiable - dewey decimal system
    • taxanomy - we think natural and non-negotiable - not something we decide, but something we discover
      • Why this difference?
        • 2000 years of western philosophy (essentialist - things really are things)
        • species are real things
        • Darwin's TOL hypothesis - spec form of evolution that relates to classification
• evolution as gen theory
  • 1858 William Paley - watch - intelligent design
  • 2 part theory claiming past explains present
    • (1) adaptation and (2) classification have historical, not theistic or metaphysical purpose
  • Darwin's theory - specific model how past processes affect current
    • 'from the 1st dawn of life all org beings class in groups onto groubs, not arb - something we dont decide, but discover'
    • organisms appear to be related because they are related
  • TOL - classification tree like because process of creation is tree like - one of successive branching
  • TOL - ladder which helped community to adopt larger theory of evolution
  • 1859 - tree-like class gotten credibility by association with TOL hypoth (which got further credibility)
  • Alec Panchen - 'Clossification, Evolution and the Nature of Biology', 1992
    • TOL circular argument
  • Zuckerlandl and Pauling, 1965 - basic theory of molec phylogeny
• 1987- Woese - Bacteriological Reviews - summary of his work
  • considered lateral gene transfer
• Claim TOL hypothesis untrue
  • Bergey's Manual (1st ed 1923) of Bacteriology - made some claims obout phylogenetic relationships
  • Welch, 2002 - 3 sequenced E. coli strains sequenced - only 40% genes overlay
    • species - many of these genes used in a species classification
  • Walsh, 2007 - more than 1/2 genes of archeae for metab have best hits in bacterio
    • phenotypes not being det by process not underlined by branching
  • eukaryotes as a whole - more a sister group to bacteria than archaea for more than 1/2 the genes
  • universal core of genes shared by all genomes - something around 30
• implications of this untruthfulness for classification and evolution
  • many of prokaryote properties not inhereted in branching classification
  • helps to disentangle evolution and classification (used to be completely intertwined)
    • can still do phylogenies, but treat them as classifications - no 'really is' pressure anymore
      • multiple proper places for organimsm in classification scheme
  • general theory - present happened spontaneusly from the past - through general evolution
  • but evolution not a branching process

Peter Gogarten : Ancient Gene Transfer and Phylogenetic Reconstruction: Friends or Foes

Wed Jun 27 13:00:54 EDT 2007

Talk

• can use hgt events as tool to learn more like evol history
• alternative hypoth to observations other than hgt
  • genome duplication followed by differential gene loss