Classification-standards

[tarak]

We classify different variables based on different parameters.The reconstructionof phylogenies through evolution is carried out based on different methods as appropriate with the data acquired.
With respect to the estimation of genetic similarity or dissimilarity using distance methods there are several statistical formulae and they yield different results.Coming to phylogeneies of organisms there are different methods with respect to analysis of molecular sequence information.In this background I would like to know if there is any unified approach to this.Are there any references regarding this.Please do help me.
t.Dhurjati

[MortenS]

When analyzing DNA, RNA or protein data, Maximum Likelihood methods are preferred over the other methods if the dataset is small (less than 30 sequences). Since the ML methods compute the probability of large amount of trees, the method becomes very CPU intensive with many sequences.

One way of resolving this problem, could be to use parsimony methods or distance methods to find candiate trees, then use Maximum Likelihood methods to evaluate these trees, instead of using Maximum Likelihood methods to find the best trees out of all possible trees when you have large datasets.

Some resources:

Phylogenetic Methods Come of Age: Testing
Hypotheses in an Evolutionary Context

The Robustness of Two Phylogenetic Methods: Four-Taxon Simulations
Reveal a Slight Superiority of Maximum Likelihood over Neighbor Joining

Phylogenetic Analysis Using Parsimony and Likelihood Methods