Phylogenetic Trees
A phylogenetic tree is a diagram of how sequences, and the organisms they come from, are related by descent. The tips are the sequences you compared, each internal node is a hypothetical common ancestor, and the branching pattern shows who is closest to whom.
How a tree is built
This tool uses a distance method. It aligns the sequences, measures how different each pair is, converts that to an estimate of evolutionary distance (correcting for positions that mutated more than once), and then repeatedly joins the two closest groups. The result is a tree.
Distances. For DNA the simple fraction of differing positions p is corrected with the Jukes–Cantor formula, d = −¾ ln(1 − 4p/3). For proteins the Poisson correction d = −ln(1 − p) is used. Positions with a gap or an ambiguity code are left out.
UPGMA (unweighted pair group method with arithmetic mean) joins the closest pair, replaces them with their average distance (weighted by group size) and repeats. It produces a rooted tree in which all tips end at the same height.
What it assumes
UPGMA assumes a molecular clock: every lineage evolves at the same rate. If some lineages evolve faster, UPGMA can group the wrong sequences together, and neighbor-joining or maximum-likelihood methods are more reliable. Distance methods also reduce each pair of sequences to one number, which loses information that likelihood and Bayesian methods use.
How Aaron's Archive does it
| Item | What the tool does |
|---|---|
| Alignment | Each sequence is aligned to a central reference sequence, then the pairwise distances are calculated. |
| Model | Jukes–Cantor for DNA, Poisson for protein, then UPGMA. The same rule for DNA versus protein is applied across the whole site, so DNA containing ambiguity codes such as R or Y is still treated as DNA. |
| Check | Tree distances were compared with SciPy's UPGMA on over a hundred simulated datasets and match. |
| Output | A drawn tree, the distance matrix and the tree in Newick format. |
The Newick format
Newick writes a tree as nested brackets with branch lengths, for example ((A:0.1,B:0.1):0.2,C:0.3);. Nearly every tree viewer (FigTree, iTOL, MEGA) opens it.
Pitfalls
A tree is a hypothesis, not a result. This tool does not calculate bootstrap support, so it cannot say how confident any branch is. Use it to explore and for teaching; for published work use a dedicated package.
Common questions about phylogenetic trees
What is the difference between UPGMA and neighbor-joining?
UPGMA assumes every lineage changes at the same rate, which gives a tree with all tips at the same distance from the root. Neighbor-joining does not assume that, so it is usually preferred when rates differ. UPGMA is easier to follow step by step.
What do branch lengths mean?
In this tool they are distances in substitutions per site, corrected for multiple changes at the same site (Jukes-Cantor for DNA).
What is a Newick string?
A text format for trees: brackets group the branches and numbers after colons give branch lengths. Most tree viewers can open it.
Where can I build a tree?
Use the phylogenetic tree maker: paste 2 to 100 FASTA sequences.