## Bootstrapping

*A short "as I understand it" description of
bootstraping.*

This page assumes you've read the previous links on
phylogenetics.

- Bootstrapping is a technique for estimating the reliability of
an internal branch of a phylogenetic tree by resampling the
original data set. With DNA sequences the bases at each position
are randomly sampled then returned to the pool so that they may be
resampled again. The bootstrap value for a branch is the percentage
of such resamplings (typically 500 to 1000) that recover the
branch.

## What does the definition mean?

As I understand it from the above definition and various examples
on the web (I never found one that I thought was entirely clear so
I haven't provided a link to any), bootstraping is randomly
modifying bases in the sequences and seeing how this affects the
phylogentic tree produced.

+---+---+ -------------------
| | | A | B | C |
| | A| \ | 20 | 30 |
| +---+---+ |------------------
| | | B| 20 | \ | 15 |
| | | |------------------
A B C C| 30 | 15 | \ |
-------------------

Figure 1 - Phylogentic Tree and table of
evolutionary distances

Assume you have a phylogenetic tree and table of evolutionary
distances as in Figure 1. During bootstrapping, you would randomly
modify bases in sequences A, B and C and see what effect this had
on the phylogenetic tree. To do this, you would recalculate the
distance matrix with the randomly modified sequences and rebuild
the tree. If the tree now branches differently then there is a
lesser probability of this being the correct tree.

This process would be repeated some number of times (say 1000)
and probabilities would be assigned to each branch based on the
number of times that branch occured over the 1000 iterations
(Figure 2). The idea is to find out which branches are questionable
because the sequences involved are so closely related that some
small number of random changes will alter the branching order.

.70+---+---+ -------------------
| | | A | B | C |
| |.87 A| \ | 20 | 30 |
| +---+---+ |------------------
| | | B| 20 | \ | 15 |
| | | |------------------
A B C C| 30 | 15 | \ |
-------------------

Figure 2 - Phylogenetic Tree with weighted
branches