Mats.Sundvall@bmc.uu.se (03/18/90)
Does somebody out there have some comments on how to compare evolutionary trees calculated by maximum parsimony algorithms. If I generate one tree that is based upon one gene and another tree based upon another gene, how can I tell if they are significantly different from each other. One method may be to impose the structure of both trees one one gene and calculate how many changes it takes in each position. Then doing a sign test or something similar may give me a some info on how good they are. But I suppose something must have been done in this area before. Maybe somebody can ngive me some hints. Sincerly, -- Mats Sundvall Biomedical Center +46/18174583 University of Uppsala Mats.Sundvall@BMC.UU.SE Sweden psi%24020019700620::MATS
lloyd@bruce.OZ (lloyd allison) (03/19/90)
Mats.Sundvall@bmc.uu.se writes: >Does somebody out there have some comments on how to compare evolutionary >trees calculated by maximum parsimony algorithms. If I generate one The best approach I know of as yet is the maximum likelihood method of Joseph Felsenstein Dept of Genetics, Uni of Washington, Seattle eg "Inferring Evolutionary Trees from DNA Structures" Statistical Analysis of DNA Sequence Data (ed) B. S. Weir, Marcel Dekker inc 1983 p133-150 This has excellent statistical validity. It may not be practical if your tree is very big. JF distributes a package containing several tree building methods for free. This includes a parsimony method and a M.L. method. Lloyd Allison, Computer Science, Monash University, AUSTRALIA 3168 zzz
lloyd@bruce.OZ (lloyd allison) (03/19/90)
Mats.Sundvall@bmc.uu.se writes: >Does somebody out there have some comments on how to compare evolutionary >trees calculated by maximum parsimony algorithms. If I generate one The best method I know for comparing phylogenetic trees is Joseph Felsenstein's Maximum Likelihood method. It has excellent statistical credentials. It may not be practical if the data set is very large. You also have to give it the n-way alignment as input. J.F. distributes a package with several different tree methods on, including max parsimony and max likelihood for DNA and protein: Dept of Genetics, University of Washington, Seattle, Washington. see J. Felsenstein, Inferring Evolutionary Trees from DNA Sequences in Statistical Analysis of DNA Sequnece Data ed B.S.Weir, Marcel Dekker inc. p 133-150, 1983 L. Allison, Computer Science, Monash University, AUSTRALIA 3168 zzz
lloyd@bruce.OZ (lloyd allison) (03/19/90)
Mats.Sundvall@bmc.uu.se writes: >Does somebody out there have some comments on how to compare evolutionary >trees calculated by maximum parsimony algorithms. If I generate one The best method I know for inferring phylogenetic trees is Joseph Felsenstein's Maximum Likelihood method. It has excellent statistical credentials. You have to give it the n-way alignment as input. It may not be practical if the data set is very large. On the other hand if you want to choose between a few given trees it would be much faster. J.F. distributes a package with several different tree methods on, including max parsimony and max likelihood for DNA and protein: Dept of Genetics, University of Washington, Seattle, Washington. see J. Felsenstein, Inferring Evolutionary Trees from DNA Sequences in Statistical Analysis of DNA Sequence Data ed B.S.Weir, Marcel Dekker inc. p 133-150, 1983 L. Allison, Computer Science, Monash University, AUSTRALIA 3168 zzz