Lebanonian Cedar Some additional inquiry information
Left arrow Back to the help page Back to the description of this program Left arrow
Right arrow WPGMA

As any cluster method, UPGMA (unweighted pair-group method of analysis) performs the structure of a set of some homogeneous elements (in our case alignment is a set of amino acid sequences). Like UPGMA (carrying out an unweighted clusterisation), this method is based on consecutive joining of elements into clusters and then clusters and (or) separate elements together. Usually cluster methods are used for constructing phylogenetic trees. But WPGMA has its own specificity: working with clusters, it estimates weights of every cluster. Moreover, weights of all sequences in these clusters are also calculated. This statement is also right for the last cluster (containing all objects of the set) - thus, there is an opportunity of using WPGMA for weighting sequences in alignment.
Here, sequence weights are important in two cases. Firstly, when we calculate correlations between distance matrices on the whole alignment and on separate position (for selecting "diagnostic" positions) - we should be sure that these data are not distorted by multiple repeating of one sequence. Secondly, weights are important for defining branch supporting positions - by the same reason. Anyway, several copies of one sequence (more commonly - copies of one object in a given set) should give the same contribution as if there were only one copy of this object.

One copy of every sequence Polytens
If some sequences are repeated several times, they can mark out wrong signs (like gap in positions 2 and 3 or valine in position 9, aspartic acid in position 13 etc.), define variable positions as some kind of diagnostic (positions 14, 15, 17, 20) or hide real sign (position 7 - its cysteine is not noticeable behind multiple copies of alanines and lysines which are non-specific for this position; position 11 - though it contained positive acids and there were only two cases of negative glutamic acid, after repeating this acid remains in majority)

While gathering objects, WPGMA constructs a tree. This tree represents phylogeny rather reliably (WPGMA is based on the same principles with UPGMA, which is widely used for phylogeny reconstruction). SVETKA can analyse alignments together with any assigned tree (of course, sets presented in the tree and in the alignment should coincide - but the topology of this tree can be arbitrary). Despite this, assignment of tree topology (by citing tree bracket scheme) is not necessary. In the case when only alignment is given, the program uses WPGMA-constructed tree (as default). The tree is then presented in all diagrams ane web-pages, further calculations (definition of supporing positions, calculations of branch parameters) are made for this tree.

Arrow upstairs
When mistakes or interesting facts are found, please tell!