Johnjoe McFadden asks if there is a way to take a similarity matrix
generated from +/- data from RFLP patterns and use it as the bases for input
to phylogeny programs. Of course there are methods that can use the +/-
data directly (parsimony, for example) but I assume that distance matrix
methods were intended.
Nei and Li (PNAS 1979) proposed a distance measure for such cases, which is
basically a variant on -ln(S) where S is the fraction of fragments shared.
It is not hard to write a program to calculate it, and perhaps one could
find a way to do it with a spreadsheet program.
Joyce Miller, who can be reached through the Department of Agronomy,
University of Wisconsin, Madison, Wisconsin 53706 (last I heard) has a
package, RESTSITE, for PCDOS that computes this measure.
It is a limitation on my own package that I don't happen to have put a
program in to compute this measure. I will in the next release (3.5, due
who knows when. But it is really not hard to write a program in, say,
BASIC that reads the data files and writes a distance matrix compatible with
existing packages.
-----
Joe Felsenstein, Dept. of Genetics, Univ. of Washington, Seattle, WA 98195
Internet: joe at genetics.washington.edu (IP No. 128.95.12.41)
Bitnet/EARN: felsenst at uwavm
UUCP: ... uw-beaver!evolution.genetics!joe
Let's all hope the human-chimp hybrid discussion goes to the
POPULATION-BIOLOGY group and does not come here (it's not molecular anyway)!
