Notes 20100720 113343 CS 798 Mohommad's Project Talk
From SnOwy - Ed's Wiki Notebook
Study different objective functions
- branch lengths
- correct topologies
- accuracy
- if the sequences being studied are short, the distances cannot be measured very well
- objective function used: the known actual distance versus the guessed distance given the method
- multiplicative noise model
- assume in nature, error is proportional to distance
ROSE
- random model of sequence evolution
- implements a probabilistic model of the evolution of sequences
- returns correct phylogeny and MSA
- old -- classic, still used
Noise independent of topology?
- additive noise model
- internal nodes manage to retain the noise of its child taxa
- in the end: which objective function can give better scores for correct topology given noise