Skip to main content

substmodel_belief_propagation.tppl

function compute_postorder_message​

Compute postorder messages on the observed tree

function compute_postorder_message(tree: GTR_Tree, data: Int[][], q: Matrix[Real]) => Matrix[Real][]

function get_leaf_message​

Get message from leaves for each site

function get_leaf_message(seq: Int) => Matrix[Real]

function get_log_likes​

Compute log likelihood for each site

function get_log_likes(msg: Matrix[Real], pi_col: Matrix[Real]) => Real

function gtr​

A help function to compute the scaled rate matrix for GTR

function gtr(pi: Real[], r: Real[]) => Matrix[Real]

type GTR_Tree​

type GTR_Tree =
  | Leaf {age: Real, index: Int}
  | Node {age: Real, left: GTR_Tree, right: GTR_Tree}

function message​

A simple help function to compute a message for a branch in the tree in a postorder traversal towards the root. We assume standard conventions here for the structure of q with respect to time (rows = from-states)

function message(start_msg: Matrix[Real][], q: Matrix[Real], time: Real) => Matrix[Real][]

function myModel​

MODEL clock_rate not needed in this version

model function myModel(data: Int[][], tree: GTR_Tree) => TheReturn

type TheReturn​

type TheReturn = TheReturn{stationary_probs: Real[], exchangeability_rates: Real[]}