# Distance methods. UPGMA: similar to hierarchical clustering but not additive Neighbor-joining: more sophisticated and additive What is additivity?

## Presentation on theme: "Distance methods. UPGMA: similar to hierarchical clustering but not additive Neighbor-joining: more sophisticated and additive What is additivity?"— Presentation transcript:

Distance methods

UPGMA UPGMA is not additive but works for ultrametric trees. Takes O(n^3) time ADC B 3 3 3 10 A B C D A B CD 626 6 3

1.Initialize n clusters where each cluster i contains the sequence i 2.Find closest pair of clusters i, j, using distances in matrix D 3.Make them neighbors in the tree by adding new node (ij), and set distance from (ij) to i and j as D ij /2 4.Update distance matrix D: for all clusters k do the following (ni and nj are size of clusters i and j respectively) 5.Delete columns and rows for i and j in D and add new ones corresponding to cluster (ij) with distances as computed above 6.Goto step 2 until only one cluster is left UPGMA

A B C D A B CD 632 6 ADC B 3 3 3 13 3

UPGMA Doesn’t work (in general) for non ultrametric trees A D CB 10 3 33 3 A B C D A B CD 131926 1219 13

UPGMA UPGMA constructs incorrect tree here A B C D A B CD 13 19 26 1219 13 BDA C 6 7.25 6

UPGMA Bipartition (BC,AD) is not in true tree BDA C 6 7.25 6 A D CB 10 3 33 3 True treeUPGMA tree

Neighbor joining 1.Additive and O(n^3) time 2.Initialization: same as UPGMA 3.For each species compute 4.Select i and j for which is minimum 5.Make them neighbors in the tree by adding new node (ij), and set distance from (ij) to i and j as

Neighbor joining 6.Update distance matrix D: for all clusters k do the following 7.Delete columns and rows for i and j in D and add new ones corresponding to cluster (ij) with distances as computed above 8.Go to 3 until two nodes/clusters are left

NJ NJ constructs the correct tree for additive matrices A D CB 10 3 33 3 A B C D A B CD 131926 1219 13

Simulation studies

The true evolutionary tree is never known in practice. Simulation allows us to study accuracy of methods under biologically realistic scenarios Mathematics behind the phylogenetics is often complex and challenging. Simulation allows us to study algorithms when not possible theoretically and also examine algorithm performance under various conditions such as different evolutionary rates, sequence lengths, or numbers of taxa