THE COMPUTATION LABORATORY OF HARVARD UNIVERSITY
CAMBRIDGE 38, MASSACHUSETTS

December 21, 196)

 

Professor Joshua Lederberg
Department of Genetics \
Stanford University
Medical Center

Palo Alto, California

Dear Professor Lederberg:

Many thanks for your letter of December 17, as well as for the
enclosed reprint.

I believe that you are right in saying that there is no redundancy
between your work and our graph matching procedures. We have in fact
by-passed the canonical form problem by devising a matching process which
does not require an initial normalization before matching. However, we
might make good use of a canonical form, if one were easily constructible,
as an initial screening device, that is, to eliminate from the matching
process in advance most of those structures which could obviously never
match.

veatjya 4

Another question (other than the input screening) which we have
not properly solved is the Markush group problem which arises with in-
completely specified molecules. This might also cause some problems with

the normalization process. :

I hope that you will continue to keep us acquainted with your work.

Many thanks again,

Sincefely,

1 i

es we) bite.
Gerprd Salton

GS: jm