Indices - PD and PE should use only the nodes on the tree

[GoogleCodeExporter]

This is a follow-on from issue #319.

Basically, the PD and PE calcs iterate over all the labels in the neighbour 
set, skipping those not in the tree.  It would speed up the calculations if we 
started with a list of those on the tree, as then there would be no need to 
check the validity repeatedly.

The PD_per_taxon index will also be correctly calculated since it will divide 
PD by the nodes actually used, not the full list of species.  

Original issue reported on code.google.com by shawnlaffan on 9 Aug 2012 at 2:55