Transcript Phylogeny
EEOB 400: Lecture 13 Phylogeny Outgroup Species A Species B Species C AAGCTTCATAGGAGCAACCATTCTAATAATAAGCCTCATAAAGCC AAGCTTCACCGGCGCAGTTATCCTCATAATATGCCTCATAATGCC GTGCTTCACCGACGCAGTTGTCCTCATAATGTGCCTCACTATGCC GTGCTTCACCGACGCAGTTGCCCTCATGATGAGCCTCACTATGCA Extra credit Using genetic markers to examine spatial and temporal variation in Ohio Canada Goose harvest composition Dr. Kristin Mylecraine The Ohio State University and Ohio Division of Natural Resources Thursday November 9 12:00 PM, Lazenby 21 Tree of life Why is phylogeny important? Understanding and classifying the diversity of life on Earth Testing evolutionary hypotheses: - trait evolution - coevolution - mode and pattern of speciation - correlated trait evolution - biogeography - geographic origins - age of different taxa - nature of molecular evolution - disease epidemiology …and many more applications! Phylogeny What is a phylogeny? Branching diagram showing relationships between species (or higher taxa) based on their shared common ancestors Species: A B C D E F Time E A B C F D Time A and B are most closely related because they share a common ancestor ( call the ancestor “E”) that C and D do not share A+B+C are more closely related to each other than to D because they share a common ancestor (“F”) that D does not share Phylogeny Terminal nodes = contemporary taxa Internal nodes = ancestral taxa Phylogeny and classification Hierarchy All taxonomic classifications are hierarchical – how does phylogeny differ? Class Order Order Family Genus Species 1 Species 2 Species 3 Species 4 Genus Species 1 Species 2 Species 3 Family Genus Species 1 Species 2 Family Genus Species 1 Species 2 Species 3 Species 4 Species 5 Species 6 Species 7 Species 8 Species 9 Genus Species 1 Species 2 Genus Species 1 Genus Species 1 Species 2 Species 3 Phylogeny and classification Hierarchy Phylogenetic (cladistic) classification reflects evolutionary history The only objective form of classification – organisms share a true evolutionary history regardless of our arbitrary decisions of how to classify them Phylogeny Classification Genus Family Genus Order Genus Family Genus Class Genus Family Genus Order Genus Family Genus Phylogeny and classification Classification Note that taxa are nested on the basis of shared common ancestors e.g., All tetrapods share a common ancestor with legs, but other chordates outside of Tetrapoda do not share this common ancestor The traits mapped onto the phylogeny are synapomorphies – we will return to them later Phylogeny and classification Monophyletic group Paraphyletic group Polyphyletic group Includes an ancestor all of its descendants Includes ancestor and some, but not all of its descendants Includes two convergent descendants but not their common ancestor A B C D How could this happen? A B C D Taxon A is highly derived and looks very different from B, C, and ancestor A B C D Taxon A and C share similar traits through convergent evolution Only monophyletic groups (clades) are recognized in cladistic classification Phylogeny and classification Monophyly Each of the colored lineages in this echinoderm phylogeny is a good monophyletic group Asteroidea Ophiuroidea Echinoidea Holothuroidea Crinoidea Each group shares a common ancestor that is not shared by any members of another group Paraphyletic groups Foxes Paraphyly “Foxes” are paraphyletic with respect to dogs, wolves, jackals, coyotes, etc. This is a trivial example because “fox” and “dog” are not formal taxonomic units, but it does show that a dog or a wolf is just a derived fox in the phylogenetic sense Lindblad-Toh et al. (2005) Nature 438: 803-819 Paraphyletic groups Canids Monophyly Note that canids are still a good monophyletic clade within Mammalia Each of the colored lineages within canids is also a monophyletic clade Lindblad-Toh et al. (2005) Nature 438: 803-819 Paraphyletic groups Lizards Paraphyly “Lizards” (Sauria) are paraphyletic with respect to snakes (Serpentes) Serpentes is a monophyletic clade within lizards Squamata (lizards + snakes) is a monophyletic clade sister to sphenodontida Snakes are just derived, limbless lizards Fry et al. (2006) Nature 439: 584-588 Paraphyletic groups Reptilia Paraphyly Birds are more closely related to crocodilians than to other extant vertebrates Archosauria = Birds + Crocs We think of reptiles as turtles, lizards, snakes, and crocodiles But Reptilia is a paraphyletic group unless it includes Aves What does this mean? It means that “reptiles” don’t exist! No, it means that you’re one of us! What it means is that “reptile” is only a valid clade if it includes birds Birds are still birds, but Aves cannot be considered a “Class” equivalent to Class Reptilia because it is evolutionarily nested within Reptilia Reptilia Aves (birds) Turtles Crocodiles Lizards and snakes Tuataras Testing evolutionary hypotheses Mapping evolutionary transitions Some horned lizards squirt blood from their eyes when attacked by canids How many times has blood-squirting evolved? Blood squirting? No Yes Testing evolutionary hypotheses Mapping evolutionary transitions Some horned lizards squirt blood from their eyes when attacked by canids How many times has blood-squirting evolved? This phylogeny suggests a single evolutioary gain and a single loss of blood squirting Blood squirting? No Yes Testing evolutionary hypotheses Mapping evolutionary transitions But a new phylogeny using multiple characters suggests that blood squirting has been lost many times in the evolution of this group Our interpretation of these evolutionary scenarios depends on phylogeny Leaché and McGuire. Molecular Phylogenetics and Evolution 39: 628-644 Testing evolutionary hypotheses Reconstructing ancestral characters This phylogeny also shows how we can use data from living species to infer character states in ancestral taxa ? ? Ancestral state could be blue, purple, or intermediate…outgroup comparison indicates blue is most parsimonious Leaché and McGuire. Molecular Phylogenetics and Evolution 39: 628-644 Testing evolutionary hypotheses Mapping evolutionary transitions How many times has venom evolved in squamate reptiles? Once in the large “venom clade” Groups within this clade then evolved different venom types e.g., different proteins found in Snakes versus Gila monsters Even non-venomous lizards in this clade (Iguania) share ancestral toxins Fry et al. (2006) Nature 439: 584-588 Testing evolutionary hypotheses Convergence and modes of speciation What can this phylogeny tell us about homology/analogy and speciation? Lake Tanganyika 1. Similarities between each pair are the result of convergence 2. Sympatric speciation more likely than allopatric speciation Lake Malawi Testing evolutionary hypotheses Coevolution Aphids and bacteria are symbiotic Given this close relationship, we might expect that speciation in an aphid would cause parallel speciation in the bacteria When comparing phylogenies for each group we see evidence for reciprocal cladogenesis (but also contradictions) Clark et al. (2000) Testing evolutionary hypotheses Geographic origins A Where did domestic corn (Zea mays maize) originate? Populations from Highland Mexico are at the base of each maize clade B Matsuoka et al. (2002) Testing evolutionary hypotheses Geographic origins Where did humans originate? Each tip is one of 135 different mitochondrial DNA types found among 189 individual humans African mtDNA types are clearly basal on the tree, with the nonAfrican types derived Suggests that humans originated in Africa Vigilant et al. (1991) Science Reconstructing evolutionary history Phenetic methods Based on overall difference between taxa = “distance” methods Only considers shared characters; not shared, derived characters Suppose you use DNA hybridization to compare DNA of 4 species A differs from B by 4% A differs from C by 10% A differs from D by 10%…for all pairs Use algorithm to find shortest tree “Quick and dirty” method Distance method will often recover trees that are similar to cladistic trees, but it requires constant rate of evolution or it will give erroneous groupings Reconstructing evolutionary history Cladistic methods (Willi Hennig 1966) Based on shared, derived characters = synapomorphies Similarity is not enough – requires similarity reflecting descent with modification Requires characters that can be assigned a particular character state Characters and character states Character: Molecular Characters eye color Character states: blue, brown, green mammary glands present, absent number of legs 0, 2, 4, 6, 8, etc. nucleotide bases A, C, T, G amino acid codons ACC, CGT, GAT, etc. Terminology Plesiomorphy Character state found in ancestor of group Apomorphy Derived character state in descendants of group Symplesiomorphy Shared, ancestral character state Synapomorphy Shared, derived character state (indicates homology) Polarity Distinguishing ancestral (0) from derived (1) = assigning polarity - polarity can be assessed by outgroup comparison A “Blue” and “square” are plesiomorphic “Small size” is an apomorphy for A “Red” is a synapomorphy for A + B “Circle” is a synapomorphy for A + B + C …but a symplesiomorphy for A + B B C D Synapomorphy Synapomorphy Each character shown in pink is a synapomorphy Shared - by all descendants in the clade e.g., all chordates share a notochord Derived – not present in ancestral taxa e.g., ancestral deuterostome lacks a notochord Any clade must share at least one synapomorphy Synapomorphy How can we tell how well a clade is supported? In part, by the number of synapomorphies Few synapomorphies = weaker support Many synapomorphies = stronger support Homoplasy Homoplasy Taxa share a character, but not by descent from a common ancestor Equivalent to analogy, homoplasy is a product of convergent evolution Homoplasy gives the impression of homology (synapomorphy) and therefore misleads phylogenetic analyses by supporting polyphyletic taxa Recovered phylogeny True phylogeny Homoplasy Homoplasies that look like homologies: Stripes Spotted caudal fin Yellow color Recovered phylogeny Lake Tanganyika True phylogeny: Malawi cichlids monophyletic Lake Malawi Morphological characters Examples Skull structure in cetaceans Genitalia in ants Morphological characters Constructing a character matrix Suppose we want to know the phylogeny of cichlids A, B, C using an Outgroup First, we need characters that are variable within this group Character: Pattern Caudal Caudal Pattern Shape Forehead Bulge? Striped Spot No Out A Round Synapomorphies Barred None Forked No Barred None Forked No B C Barred None Round Yes Apomorphy Parsimony How do we decide the “best” phylogeny? Parsimony – the simplest explanation is preferred (Occam’s razor) A trivial example (much more complicated with real datasets) Most parsimonious: Requires 5 steps Requires only 4 steps Round forked tail Round forked tail Round forked tail Stripe barred Spot plain tail Stripe barred Spot plain tail No bump forehead bump No bump forehead bump Molecular characters Outgroup Species A Species B Species C AAGCTTCATAGGAGCAACCATTCTAATAATAAGCCTCATAAAGCC AAGCTTCACCGGCGCAGTTATCCTCATAATATGCCTCATAATGCC GTGCTTCACCGACGCAGTTGTCCTCATAATGTGCCTCACTATGCC GTGCTTCACCGACGCAGTTGCCCTCATGATGAGCCTCACTATGCA 2. Sequence 1. Extract 3. Align Molecular characters Outgroup Species A Species B Species C Out AAGCTTCATA GAGCTTCACA GTGCTTCACG GTGCTTCACG Invariable sites These are not useful phylogenetic characters Out A A B B C C Molecular characters AAGCTTCATA GAGCTTCACA GTGCTTCACG GTGCTTCACG Outgroup Species A Species B Species C Out Synapomorphies supporting A+B+C Out A A B AG C TC Any mutations at this time would affect A, B and C because they have not yet diverged B C Molecular characters AAGCTTCATA GAGCTTCACA GTGCTTCACG GTGCCTCACG Outgroup Species A Species B Species C Synapomorphies supporting A+B+C Synapomorphies supporting B+C Out Out A A B AG B TC C AT AG Any mutations at this time would affect A and B C Molecular characters AAGCTTCATA GAGCTTCACA GTGCTTCACG GTGCCTCACG Outgroup Species A Species B Species C Synapomorphies supporting A+B+C Synapomorphies supporting B+C Apomorphy for C Out Out A A B AG B TC C AT AG C Any mutations at this time would only affect C TC Molecular characters Homoplasy is still a problem There are only 4 possible character states for nucleotides: A G C T Homoplasy arises when nucleotide mutates back to ancestral state: Out AAGCTTCATA GAGCTTCACA GTGCTTCACG GTGCTTCACG AG TC A AAGCTTCATA GAGCTTCACA GTGCTTCACG GTGCTTCACG AT Back-mutation “erases” synapomorphy and produces homoplasy AG B AAGCTTCATA GAGCTTCACA GTGCTTCACG GAGCTTCACG C TA ATA Molecular characters Homoplasy is still a problem There are only 4 possible character states for nucleotides: A G C T Homoplasy arises when nucleotide mutates back to ancestral state: ATA Out AAGCTTCATA GAGCTTCACA GTGCTTCACG GTGCTTCACG AG TC CA A AAGCTTCATA GAGCTTCACA GTGCTTCACG GTGCTTCACG AT Back-mutation “erases” synapomorphy and produces homoplasy AG B Homoplasy can also reflect convergent mutations AAGCTTCATA GAGCTTAACA GTGCTTCACG GAGCTTAACG C TA CA Morphology vs molecules Morphology Nucleotides Homoplasy can be assessed from structure, development, etc. PRO Homoplasy can’t be assessed directly (an “A” is an “A”) CON Characters may be subject to selection = convergence = homoplasy CON Characters may or may not be subject to selection – depends on the site ? Takes lots of time to identify and code characters for analysis CON Sequencing yields lots of characters if gene is sufficiently variable PRO Requires parsimony analysis ? Can use either parsimony or likelihood analysis – stronger inference PRO Only someone familiar with taxon can identify good characters PRO & CON Any idiot can get sequence data PRO & CON With either approach, it all comes down to successfully identifying synapomorphies and distinguishing them from homoplasies Character conflict Character conflict With either morphology or molecules, some characters will not “agree” on the most parsimonious phylogeny Some characters support monophyletic Reptilia exclusive of birds These are not synapomorphies for “reptiles”, they are ancestral traits Feathers, two legs, and endothermy are apomorphic in birds Other characters reflect synapomorphy and recover the true relationships But in many cases it is more difficult to resolve character conflict Consensus When multiple phylogenies are supported… A consensus tree shows only those relationships common to all trees The lower tree is a “compromise” between conflicting upper phylogenies Examples: - two equally parsimonious trees - two trees from different genes - morphological vs. molecular tree - parsimony vs. likelihood tree Consensus trees will always have at least one polytomy - a branching event that is not a bifurcation Better to have an incompletely resolved tree than an incorrect tree Consensus An example of a consensus tree for loons The middle tree is a “compromise” between conflicting left and right trees Polytomy Are polytomies real? Usually not - they reflect inability to reconstruct the true bifurcating phylogeny We often encounter polytomies in cases of rapid speciation when an ancestor rapidly diverged into many new forms A B C D True phylogeny E F G Inferred phylogeny H I J K = 1 million years = Change in character state We can only recover those branches on which we “see” characters change Different genes for different questions Molecular stopwatch Deepest root: 35 mya (use mtRNA) Molecular hourglass 600 mya (use nuclear rRNA) Different genes, different trees Gene 1 Species A Species B Gene 2 Species C Species A Species B Species C Red and blue indicate different alleles for a particular gene (gene 1 or 2) A B C A B C Incorrect Correct Because genes are inherited as a single unit, all of the nucleotides in a gene can support the same phylogeny, and it could still not reflect true speciation sequence