[phylocom-user] aot varcontr
steve.kembel at gmail.com
Sun Nov 15 11:09:27 PST 2009
VarContr in the aot/aotf output is the observed variance of
The observed variance of standardized contrasts is calculated as
sum(squared standardized contrast values)/(N-1), as described by
Blomberg and Garland.
VarCn.rankLow and VarCn.rankHi are the rank of the observed VarContr
versus the distribution of randomized standardized contrasts. These
can be used to calculate a P-value testing for phylogenetic signal in
the traits. VarCn.rankLow divided by number of runs plus one is a one-
tailed P-value testing for greater phylogenetic signal than expected.
On Nov 15, 2009, at 10:46 AM, Scott Chamberlain wrote:
> I am confused about what varcontr is actually calculating in the aot
> or aotf output.
> Is the varcontr in the aot and aotr output the expected from the
> randomizations, or the observed? I have been taking this as the
> expected value from randomizations.
> In the phylocom manual, varcontr is described as the variance of
> standardized contrasts across the tree. However, calculating the
> observed variance of standardized contrasts is very far from the
> expected varcontr, whereas variance of raw contrasts is closer to
> the expected varcontr. Is the varcontr from the output actually the
> variance of raw contrasts, not standardized contrasts?
> Thanks, Scott Chamberlain
> phylocom-user mailing list
> phylocom-user at lists.phylodiversity.net
More information about the phylocom-user