MBY134 at sysh.surrey.ac.uk MBY134 at sysh.surrey.ac.uk
Mon Jul 22 12:11:47 EST 1991


Joe Felsenstein's reply to my question regarding the analysis
of RFLP data to derive phylogenetic relationships suggests
that I should use distance methods such as Nei and Li's.

The computational methods such as Nei & Li's assume that RFLPs
are due to base-substitutions in the RE site and use shared
bands to thereby calculate DNA divergence on the basis of the
probability of a certain level of base-substitution hitting a
RE site.

However, I do not believe that the DNA polymorphisms we detect
using a number of random (anonymous) DNA probes, between
related meningococcal strains are due to single base
substitutions. Many, if not most polymorphisms we see are
likely to be the result of recombination/insertion/deletion
events since many probes give polymorphisms on all digests
whereas others probes are never polymorphic. I am therefore
wary of using this type of analysis.

It is for this reason that I've taken an essentially numerical
taxanomic approach, counting a band as an independent
character and derived a similarity matrix that has no
pretensions to relate to eg. DNA divergence. A series of
random (anonymous) probes were used to give a reasonable
sample of the genome for about 100 different strains to give a
100x100 similarity matrix. It is deriving and testing likely
phylogenies from this matrix (rather than inputing the bands
all again!) that I'm interested in.

I'm sure this is a simple matter for numerical taxonomists and
the enzyme electrophoretic typing people who must do similar
types of analysis so I would be keen to discover what programs
they use and where can I get hold of them?

More information about the Mol-evol mailing list