Friday, October 11, 2013

revised 599nts and 999nts Raxml vs Clustering using Spherical Phylogenetic Tree


Dataset

1. revised 599nts
This work follows the previous work on http://salsafungiphy.blogspot.com/2013/09/pairwise-distances-from-msa-vs-pairwise.html
This dataset has a total number of 831 sequences.
2. 999nts
This work follows the previous work on http://salsafungiphy.blogspot.com/2013/06/pairwise-distances-from-multiple.html
This dataset has a total number of 1306 sequences.

Alignment

The clustering was done based on
1) SWG, using EDNAFULL scoring matrix, with gap open = -16 and gap extension = -4
2) Multiple sequence alignment, PID.

The Raxml was done based on multiple sequence alignment,
1) revised 599nts, newick file
2) 999nts, newwick file

Spherical Tree

The information of spherical tree can be found on previous work of http://salsafungiphy.blogspot.com/2012/11/phylogenetic-tree-generation-for.html and http://salsafungiphy.blogspot.com/2012/11/phylogenetic-tree-mega-table.html

Dimension Reduction

Manxcat SMACOF:
The pviz file for SWG clustering result with revised 599nts is here, for 999nts is here.
For MSA clustering result with revised 599nts is here, for 999nts is here.
WDA-SMACOF:
alpha set to 0.95
The pviz file for SWG clustering result with revised 599nts is here, for 999nts is here.
For MSA clustering result with revised 599nts is here, for 999nts is here.

Sum of branch lengths (edge sum)

(note that the difference of edge sum between SWG and MSA should due to the difference of the original distances of the clustering plot)
Manxcat SMACOFWDA-SMACOF
revised 599nts999ntsrevised 599nts999nts
SWG19.3719.8916.0112.74
MSA16.6216.3615.2013.65
1) revised 599nts spherical tree

MSA Result

SWG Result


2) 999nts spherical tree

MSA Result
SWG Result