Dear phylogeneticists,
While perusing some recent Sys Bio articles I noticed some strict
consensus trees that were reportedly based on tens of thousands of
most-parsimonious trees (mpts). In one case the strict consensus was
based on 45,000 mpts. What was odd is that upon examining the strict
consensus tree it seemed unbelievable that one could derive 45,000 unique
mpts from that consensus tree (because it had very few polytomies). I
have noticed this in my own analyses- a run that produced 8,000 mpts
yielded a strict consensus with only a few polytomies.
I then examined the tree filtering options in PAUP* and noticed that one
can filter out "non-binary" trees. Upon doing so the number of mpts
drops dramatically (in one case from 4,000 mpts to 1,000 mpts) but the
strict consensus of both sets (with and without the non-binaries) are
IDENTICAL.
What does this mean? Does it mean that when we publish a strict
consensus based on 10,000 mpts these are on fact including thousands of
ambiguous (non-binary) trees that actually don't contradict each other?
And thus the actual number of contradictory (i.e. UNIQUE) mpts is far
lower?
Why would this be unknown to us? One would think that when we publish a
count of mpts we would know what we were doing (i.e. whether that number
represents only unique trees or includes a huge proportion of redundant
trees...) And why would anyone care about counting redundant trees????
I'd like to hear what folks think about this.
Yours,
Derek Sikes
++++++++++++++++++++++++++++++++++++++++++++++
Derek Sikes
Dept. of Ecology and Evol. Biology U-43
University of Connecticut
Storrs, CT 06269 USA
dss95002 at uconnvm.uconn.eduhttp://www.eeb.uconn.edu/grads/sikes.htm
"Remember that Truth alone is the matter you are in Search after; and if
you have been mistaken, let no Vanity reduce you to persist in your
mistake." Henry Baker, London, 1785
++++++++++++++++++++++++++++++++++++++++++++++
