1977 - 1978
ANNUAL REPORT

RESOURCE-RELATED RESEARCH

COMPUTERS AND CHEMISTRY
Grant No. RR-006I2

BIOTECHNOLOGY RESOURCES PROGRAM
OF THE
NATIONAL INSTITUTES OF HEALTH

February, 1978

COMPUTER SCIENCE DEPARTMENT
STANFORD UNIVERSITY
Resource Related Research - Computers and Chemistry
Stanford University
NIH/BRP Grant RR-00612

Carl Djerassi, Principal Investigator
(Social Security No.

Research Highlights (1977-78)

1. Stereochemistry in Structure Elucidation.

The set of computer programs developed at Stanford as tools
for molecular structure elucidation have been considerably
enhanced by the addition of 3-dimensional structural information.
The programs can now deal. with some basic geometrical properties
of molecules that are essential for understanding their
biological significance. Research progress this year has
resulted in extensions that allow computation of stereoisomers
(alternative structures differing in 3 dimensions but having
identical connections among atoms). Thus geometrical variations
on structural hypotheses can be presented as well as topological
variations.

2. Unified Package for Structure Elucidation.

Significant progress was made in unifying the computer
programs for structure elucidation into a coherent package that
is easily understood and used by chemists for complex
biomolecular structure problems. Powerful tools are now well
integrated for defining problem constraints, producing plausible
solutions to structure problems, reducing the sets of alternative
solutions with information about biosynthetic pathways, testing
the alternatives, and suggesting new tests for further
discrimination. New toolS currently under development will be
integrated into this same package.
1977-78 Annual Report RR-00612

Table of Contents

Section

Subsection
1. OVERVIEW OF RESEARCH ACTIVITIES .
2. STRUCTURE ELUCIDATION PROGRAMS .

2.1 Stereochemistry in CONGEN
2.2 Constraints Interpretation

2.3 Experiment Planning Program

2.4 The Reaction Chemistry Program

2.5 Mass Spectral Prediction and Ranking

2.6 Molecular Ion Determination
2.7 Congen Improvements - .
2.8 CONGEN Efficiency . . .

2.9 CONGEN Reprogramming . .

3. THEORY FORMATION PROGRAMS - Meta-DENDRAL

3.1 Incremental Learning . .

3.2 New Capability To Emphasize
Discriminatory Power .

3.3 Improved Ranking Capability
3.4 Data Selection Program .
3.5 Feedback Loops . . . .

3.6 Program Improvements . .

4. COLLABORATIVE RESEARCH . . . .

Page

Ov

17
25
35
49
53
60

64

70°

70

79
30
80
81
81

83
1977-78 Annual Report RR-00612

4.1 CONGEN Users > S|

4.2 Marine Natural Products . . . . . . « «+ 86

5. Carbon-13 Work eee lel GD
5.1 Rule Formation Results a ee eel 9D

5.2 Adding Stereochemistry to the Rule
Language ee ee

5.3 Structure Elucidation . . .« »« « »« «» « « 94

5.4 Geometric Distortions in Steroids . . . . . 95

6. DATA COLLECTION AND DATA REDUCTION. . . . . « . 95
6.1 DENDRAL GC/MS and MS Work 8 ee ew ee 9S

6.2 Collaborators Receiving the CLEANUP and
BISLIB ProgramS . »~ »« © »2© © «© «© «© « 97

7. APPENDICES ~ ee ele LOD
7.1 Appendix A a ee ee we ee ee «102

7.2 Appendix B eee eee ee 10
References |

ii