Distance-based Phylogeny Inference
See Notes below for instructions on using this application, including formatting guidelines.
Enter dissimilarity matrix
Label branch lengths
Draw branches proportional to
Draw branches proportional to
See Notes below for instructions on using this application, including formatting guidelines.
This application is meant to infer a phylogenetic tree using UPGMA or Neighbor-joining, given a dissimilarity matrix. In case the tree structure is already known, it can directly render a tree given its Newick notation.
In Dissimilarity Matrix mode, this application will accept an input and the mode of inference to infer a phylogenetic tree. The format of the dissimilarity matrix must adhere to the following guidelines:
n2 through n+1 contains the taxa names; this is requiredn lines contain the rows of the matrix, in the same order (in both dimensions) as the taxa in the n lines beforeThe dissimilarity matrix may be given in full, or as an upper-triangular or lower-triangular matrix. A full matrix is assumed to be symmetric about the diagonal. The application will infer which format is given. The following three valid inputs are logically identical:
3
L. williamsii
O. elata
L. usitatissimum
0 10 20
10 0 20
20 20 0
Full matrix
3
L. williamsii
O. elata
L. usitatissimum
0
10 0
20 20 0
Lower-triangular matrix
3
L. williamsii
O. elata
L. usitatissimum
0 10 20
0 20
0
Upper-triangular matrix
If the tree structure is already known, then Newick Tree mode will accept the tree and render it directly. Note that even if internal nodes in the tree string are named, the rendered tree will not display internal node names for space considerations. The rendered tree will be rooted.
If no branch lengths are provided, then the option to draw branches proportionally cannot be selected.
The string provided must be a single valid tree in Newick notation. For example, a tree that would correspond to the above dissimilarity matrices, inferred using UPGMA, would be:
(L. usitatissimum:10.0,(L. williamsii:5.0,O. elata:5.0):5.0);
A few options are available for tree aesthetics.
Branches can be labelled with their length. For space, the length is automatically rounded to 0.1 precision. Very large numbers may not render very nicely in the space given. It is advisable to input dissimilarities and lengths that remain in the [0,1000) range.
Branches can also be drawn to reflect their length in the tree. If this option is selected, the branch may be drawn proportional to its raw length or its log length. The former may make the image very large in order to maintain proper formatting. This is especially true if the branch lengths vary by more than an order of magnitude. In this case, log proportions may be more desirable. If this option is not selected, then the branches will be drawn such that every branch is of equal rendered length.
When inferring from a dissimilarity matrix, note that UPGMA will result in a rooted tree, whereas Neighbor Joining will result in an unrooted tree (as this is the nature of their algorithms). Here is an example of what each would look like, using data from Sarich (1969):
The tree inferred by Neighbor Joining can be rooted if an outgroup is specified. The outgroup must be a single taxon, and must have the exact same name as the one provided in the input above. In the process of re-rooting, the new root will be placed at the center of the pendant edge of the outgroup taxon.
In addition to the SVG image, the tree's Newick string can also be downloaded. An inferred tree does not have named internal nodes, so the downloaded string will not include them. Additionally, note that the downloaded Newick string will have two lines. The first line will be a single word, "rooted" or "unrooted", to reflect the tree's structure (this also corresponds to UPGMA or Neighbor Joining as the inference method, respectively). The second line is the Newick tree itself.