Ox (A N ‘ ene ‘

DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE

PUBLIC HEALTH SERVICE
NATIONAL INSTITUTES OF HEALTH
BETHESDA, MARYLAND 20014

December 14, 1973

Dr. Thomas C. Rindfleisch

Stanford University School
of Medicine

Stanford, California 94305

Dear Dr.

eisch:

The following are my comments on the DENDRAL draft renewal application
sent to me on November 30, 1973.

l.

I find it difficult to see a strong connection between
the Djerassi mass spectrometry laboratory (GC-HRMS) and
the AI effort. Looking at the pros and cons, it is

much easier to separate your proposal as it now stands
into the two components allowing each to sink or swim on
its own merit. The bridge between the two, DENDRAL's
need for MS data and the MS resource's need for DENDRAL,
is insufficient justification, in my opinion to combine

both into one project. If we consider AI as the primary

focus, then a case for on-line MS data acquisition must
be strengthened considerably. A question that comes to
mind immediately is this - why must DENDRAL have its own
MS capability? Can't other laboratories (such as
Burlingame's) and the literature. satisfy your needs?

If we turn things around and use MS as the driving force
behind DENDRAL, we are confronted with another set of
problems. Now it is necessary to justify DENDRAL in terms
of the research community's need for much improved methods
of interpreting MS data and showing how DENDRAL can meet
this need. Also, we must face the possibility that by the
time DENDRAL reaches the point of being generally useful,
technology might be developed to the point that MS is no
longer a useful analytical tool or other methods of dealing
with MS could be far superior to DENDRAL. The doubts
raised by these considerations are important when a priority

NY) XS
Dr. Thomas C. Rindfleisch — page 2

must be placed upon investing substantial sums of money
into a long-term project.

3. In order to support the need for direct involvement with
a MS laboratory, you should cite cases in which it was
necessary to go to the lab for additional data obtained
under controlled conditions and requiring quick turn
around.

4. Upgrading the Djerassi MS resources in the present proposal
must be justified on the basis of DENDRAL's needs. If
the resulting capability exceeds that required by DENDRAL,
then it is logical to make full utilization by offering
specialized services to the biomedical research community.

I hope you find these comments useful. Call on me if you desire further
explanation,

Sincerely yours,

Michael A. Oxman, Ph.D.
Program Director
Biomolecular Characterization Resources

Biotechnology Resources Branch
Division of Research Resources