quickly, unicity scoring is done by counting at each node of the gene tree how many times species are represented in the subtree.
This score is used (by me at least) to define good subtrees to cut to do sub-families and to perform Prunier (http://pbil.univ-lyon1.fr/software/prunier/, a software for detection of phylogenetic incongruences / gene transfers ) searches on unicopy subtrees made by different combination of sampling species only once in the subtree
for instance, in a subtree (((A1, B1),(C,D)),((A2, B2),(G,H))) which unicity score=4, following unicopy sampling will be tested:
((A1, B1),(C,D)),(G,H))
((B1,(C,D)),(A2,(G,H)))
((C,D),((A2, B2),(G,H))
((A1,(C,D)),(B2,(G,H)))
Actually, the unicity score = product(number of times a species is represented in the subtree) and corresponds to the number of different hypothesis to test.
But in this very example, one could test only two hypothesis for dectecting phylogenetic incongruence / transfers where A and B are together:
((A1, B1),(C,D)),(G,H))
((C,D),((A2, B2),(G,H))
could that be taken into account when calculating the unicity score?
what do you think?
I can be more specific in a sequel post.
Hi, I saw people speak english here, so will I.
quickly, unicity scoring is done by counting at each node of the gene tree how many times species are represented in the subtree. This score is used (by me at least) to define good subtrees to cut to do sub-families and to perform Prunier (http://pbil.univ-lyon1.fr/software/prunier/, a software for detection of phylogenetic incongruences / gene transfers ) searches on unicopy subtrees made by different combination of sampling species only once in the subtree
for instance, in a subtree (((A1, B1),(C,D)),((A2, B2),(G,H))) which unicity score=4, following unicopy sampling will be tested: ((A1, B1),(C,D)),(G,H)) ((B1,(C,D)),(A2,(G,H))) ((C,D),((A2, B2),(G,H)) ((A1,(C,D)),(B2,(G,H)))
Actually, the unicity score = product(number of times a species is represented in the subtree) and corresponds to the number of different hypothesis to test.
But in this very example, one could test only two hypothesis for dectecting phylogenetic incongruence / transfers where A and B are together: ((A1, B1),(C,D)),(G,H)) ((C,D),((A2, B2),(G,H))
could that be taken into account when calculating the unicity score?
what do you think? I can be more specific in a sequel post.
Bye
Florent