Please enable JavaScript.

Coggle requires JavaScript to display documents.

Trees and Networks (Networks (character conflict (homoplasy, recombination…

- - - - split decompostion
        
        antagonistic signals
        
        genetic distance data
    - - Minimum spanning
        
        Median joining
        
        Based on the limited introduction of likely ancestral sequences into a minimum span
        
        links between all taxa seperated by single steps
        
        mutational steps of increasing size are considered until most parsimonious tree is made
      - Median networks
        
        Reduced median networks
        
        When median are too reticulalated
        
        Factors
        
        weight of the character
        
        frequency of haplotypes
        
        K dimensional hypercube
        
        all MP trees
        
        Method
        
        variant sites within squences converted to binary
        
        sites showing perfectly correlated variation - combined
        
        for each 3 triplets in turn a median is calculated
        
        median and observed haplotypes linked by single step network
        
        used with homoplasies
- - - - distance
        
        clustering
        
        Algorithim
        
        UPGMA
        
        additive tree
        
        2 taxon with the least genetic distance grouped
        
        Used with little reason now
        
        Neighbor Joining
        
        allows for variation in evolutionary rates along branches
        
        not ultrametric
        
        low impact of multiple hits
        
        finds the minimal value for the shortest sum of branches
        
        minimal evolution
      - characters
        
        Searching
        
        optimality
        
        Parsimony
        
        sensitive to LBA
        
        works directly with character data
        
        simplest tree - smallest number of evolutionary changes
        
        Length is a function of:
        
        number of branches
        
        number of characters
        
        weighting of characters e.g. transitions
        
        Maximum likelihood
        
        looks at the tree that maximises the likelihood of observing the data
        
        evaluates a hypothesis
        
        elements:
        
        data- tree that has the best chance of explaining it
        
        unknowns
        
        mutation model
        
        tree and model parameters must be varied
        
        Simplest model is the Jukes cantor
        
        Need to account for #
        
        4 more items...
        
        Assumes equal changes
        
        Bayesian
    - - robust/efficient/consistent/powerful/falsifiable
  - - - percentage
      - Same node was constructed by the other datasets
      - not easily interpretible
  - - - branch and bound- takes away not optimal areas
      - exhaustive search
    - - hill climbing
      - try from different points