Chemical Structure Access 
& Representation

Response to OMG LSR RFP 8






A Proposal
Submitted by:

 

751 Malena Dr.,
Ann Arbor, MI 48103
 
Bioscience
9880 Campus Point Drive 
San Diego, CA 92121


October 27, 2003October 25, 2003



Table of Contents
A Proposal	1
1	Administrative	8
1.1	Submission Contact Point	9
1.2	Overview to the Material in the Submission	9
1.3	Statement of Proof of Concept	10
2	Chemical Markup Language within the Context of CSAR	12
2.1	Introduction	12
2.2	CML the Chemistry XML	14
2.2.1	The Semantic Web of Chemistry	15
2.2.2	Creating, reading and editing XML/CML documents	15
2.3	Mapping of Models	18
2.3.1	Platform-Independent and Platform-Specific UML Models	18
2.3.2	A formal approach to describing objects	19
2.4	Class Diagram: cml / Overview	21
2.5	Resolution of RFP Mandatory and Optional Requirements	22
2.5.1	Mandatory Requirements	22
2.5.2	Optional Requirements	24
2.6	Relationship to existing or emerging OMG Specifications	24
2.7	[Chemical} Elements	25
2.7.1	Molecule	25
2.7.2	Properties	25
2.7.3	Associations	25
2.7.4	Operations	25
2.7.5	MoleculeFactory	26
2.7.6	Properties	26
2.7.7	Associations	26
2.7.8	Operations	26
2.8	MoleculeUtil	27
2.8.1	Properties	27
2.8.2	Associations	27
2.8.3	Operations	27
2.9	Atom	27
2.9.1	Properties	27
2.9.2	Associations	28
2.9.3	Operations	28
2.10	AtomFactory	28
2.10.1	Properties	28
2.10.2	Associations	28
2.10.3	Operations	28
2.11	AtomParity	28
2.11.1	Properties	29
2.11.2	Associations	29
2.11.3	Operations	29
2.12	Bond	29
2.12.1	Properties	29
2.12.2	Associations	29
2.12.3	Operations	29
2.13	Electron	30
2.13.1	Properties	30
2.13.2	Associations	30
2.13.3	Operations	30
2.14	Isotope	30
2.14.1	Properties	30
2.14.2	Associations	30
2.14.3	Operations	30
2.15	IsotopeFactory	30
2.15.1	Properties	30
2.15.2	Associations	30
2.15.3	Operations	30
2.16	AtomFactory	31
2.16.1	Properties	31
2.16.2	Associations	31
2.16.3	Operations	31
2.17	BondFactory	31
2.17.1	Properties	31
2.17.2	Associations	31
2.17.3	Operations	31
2.18	NumericAtomParity	31
2.18.1	Properties	31
2.18.2	Associations	32
2.18.3	Operations	32
2.19	StringAtomParity	32
2.19.1	Properties	32
2.19.2	Associations	32
2.19.3	Operations	32
2.20	ChemicalElement	33
2.20.1	Properties	33
2.20.2	Associations	33
2.20.3	Operations	33
2.21	BondStereo (new properties only)	33
2.21.1	Properties	33
2.22	Bond (new properties only…)	34
2.22.1	Properties	34
2.23	ChemicalElementFactory	34
2.23.1	Properties	34
2.23.2	Associations	34
2.23.3	Operations	34
2.24	IsotopeSet	34
2.24.1	Properties	34
2.24.2	Associations	35
2.24.3	Operations	35
2.25	IsotopeSetFactory	35
2.25.1	Properties	35
2.25.2	Associations	35
2.25.3	Operations	35
2.26	PeriodicTable	35
2.26.1	Properties	35
2.26.2	Associations	35
2.26.3	Operations	35
2.27	PeriodicTableFactory	35
2.27.1	Properties	36
2.27.2	Associations	36
2.27.3	Operations	36
2.28	BondStereo	36
2.28.1	Properties	36
2.28.2	Associations	36
2.28.3	Operations	36
2.29	Coordinate2	36
2.29.1	Properties	36
2.29.2	Associations	36
2.29.3	Operations	36
2.30	Coordinate3	36
2.30.1	Properties	37
2.30.2	Associations	37
2.30.3	Operations	37
2.31	AbstractAngle	37
2.31.1	Properties	37
2.31.2	Associations	37
2.31.3	Operations	37
2.32	Formula	37
2.32.1	Properties	37
2.32.2	Associations	37
2.32.3	Operations	37
2.33	FormulaElement	38
2.33.1	Properties	38
2.33.2	Associations	38
2.33.3	Operations	38
2.34	Crystal	38
2.34.1	Properties	38
2.34.2	Associations	39
2.34.3	Operations	39
2.35	Query and Collection	39
2.36	Class Diagrams	39
2.37	Recommendation	51
2.37.1	Details	53
3	Glossary	56
Appendix A: UML Use Cases	60
Actor Catalogue	60
Master Use Case	61
Detailed Use Case Scenarios	62
Appendix B: Java Code	67
Appendix C: Use Cases for Chemistry	69
A Proposal	1
1	Administrative	128
1.1	Submission Contact Point	139
1.2	Overview to the Material in the Submission	139
1.3	Statement of Proof of Concept	1410
2	Chemical Markup Language within the Context of CSAR	1612
2.1	Introduction	1612
2.2	CML the Chemistry XML	1814
2.2.1	The Semantic Web of Chemistry	1915
2.2.2	Creating, reading and editing XML/CML documents	1915
2.3	Mapping of Models	2218
2.3.1	Platform-Independent and Platform-Specific UML Models	2218
2.3.2	A formal approach to describing objects	2420
2.4	Class Diagram: cml / Overview	2622
2.5	Resolution of RFP Mandatory and Optional Requirements	2723
2.5.1	Mandatory Requirements	2723
2.5.2	Optional Requirements	2925
2.6	Relationship to existing or emerging OMG Specifications	2925
2.7	[Chemical} Elements	3026
2.7.1	Molecule	3026
2.7.2	Properties	3026
2.7.3	Associations	3026
2.7.4	Operations	3026
2.7.5	MoleculeFactory	3127
2.7.6	Properties	3127
2.7.7	Associations	3127
2.7.8	Operations	3127
2.8	MoleculeUtil	3228
2.8.1	Properties	3228
2.8.2	Associations	3228
2.8.3	Operations	3228
2.9	Atom	3228
2.9.1	Properties	3228
2.9.2	Associations	3329
2.9.3	Operations	3329
2.10	AtomFactory	3329
2.10.1	Properties	3329
2.10.2	Associations	3329
2.10.3	Operations	3329
2.11	AtomParity	3329
2.11.1	Properties	3430
2.11.2	Associations	3430
2.11.3	Operations	3430
2.12	Bond	3430
2.12.1	Properties	3430
2.12.2	Associations	3430
2.12.3	Operations	3430
2.13	Electron	3531
2.13.1	Properties	3531
2.13.2	Associations	3531
2.13.3	Operations	3531
2.14	Isotope	3531
2.14.1	Properties	3531
2.14.2	Associations	3531
2.14.3	Operations	3531
2.15	IsotopeFactory	3531
2.15.1	Properties	3531
2.15.2	Associations	3531
2.15.3	Operations	3531
2.16	AtomFactory	3632
2.16.1	Properties	3632
2.16.2	Associations	3632
2.16.3	Operations	3632
2.17	BondFactory	3632
2.17.1	Properties	3632
2.17.2	Associations	3632
2.17.3	Operations	3632
2.18	NumericAtomParity	3632
2.18.1	Properties	3632
2.18.2	Associations	3733
2.18.3	Operations	3733
2.19	StringAtomParity	3733
2.19.1	Properties	3733
2.19.2	Associations	3733
2.19.3	Operations	3733
2.20	ChemicalElement	3834
2.20.1	Properties	3834
2.20.2	Associations	3834
2.20.3	Operations	3834
2.21	BondStereo (new properties only)	3834
2.21.1	Properties	3834
2.22	Bond (new properties only…)	3935
2.22.1	Properties	3935
2.23	ChemicalElementFactory	3935
2.23.1	Properties	3935
2.23.2	Associations	3935
2.23.3	Operations	3935
2.24	IsotopeSet	3935
2.24.1	Properties	3935
2.24.2	Associations	4036
2.24.3	Operations	4036
2.25	IsotopeSetFactory	4036
2.25.1	Properties	4036
2.25.2	Associations	4036
2.25.3	Operations	4036
2.26	PeriodicTable	4036
2.26.1	Properties	4036
2.26.2	Associations	4036
2.26.3	Operations	4036
2.27	PeriodicTableFactory	4036
2.27.1	Properties	4137
2.27.2	Associations	4137
2.27.3	Operations	4137
2.28	BondStereo	4137
2.28.1	Properties	4137
2.28.2	Associations	4137
2.28.3	Operations	4137
2.29	Coordinate2	4137
2.29.1	Properties	4137
2.29.2	Associations	4137
2.29.3	Operations	4137
2.30	Coordinate3	4137
2.30.1	Properties	4238
2.30.2	Associations	4238
2.30.3	Operations	4238
2.31	AbstractAngle	4238
2.31.1	Properties	4238
2.31.2	Associations	4238
2.31.3	Operations	4238
2.32	Formula	4238
2.32.1	Properties	4238
2.32.2	Associations	4238
2.32.3	Operations	4238
2.33	FormulaElement	4339
2.33.1	Properties	4339
2.33.2	Associations	4339
2.33.3	Operations	4339
2.34	Crystal	4339
2.34.1	Properties	4339
2.34.2	Associations	4440
2.34.3	Operations	4440
2.35	Query and Collection	4440
2.36	Class Diagrams	4440
2.37	Recommendation	5952
2.37.1	Conformance Points	6154
2.37.2	Details	6154
3	Glossary	6557
Appendix A: UML Use Cases	6961
Actor Catalogue	6961
Master Use Case	7062
Detailed Use Case Scenarios	7163
Appendix B: Java Code	7668
Appendix C: Use Cases for Chemistry	7870




Table of Figures

Figure 1 The representation problem. (CTL File Formats, MDL Information Systems)	13
Figure 2 Intended solution	16
Figure 3 Platform-Independent Model with Stereotypes	18
Figure 4 Summary	20
Figure 5 CSAR/CML Diagram	21
Figure 6 Search Module	39
Figure 7 Search Module (Properties Classes)	40
Figure 8 DB Module	41
Figure 9 Collection Module	42
Figure 10 Atom	43
Figure 11 Bond	44
Figure 12 Coordinates	45
Figure 13 Formula	46
Figure 14 Molecule	47
Figure 15 Other	48
Figure 16 MOLUTIL Module	49
Figure 17 Structure and Translation	50
Figure 18 Architecture	51
Figure 19 Processing diagram	52
Figure 20 Sea Search and DB Interface Modules interacting with each other to perform a 
chemical structure search against database.	54
Figure 21 General Use Case Scenario	61
Figure 1 The representation problem. (CTL File Formats, MDL Information Systems)	1713
Figure 2 Intended solution	2117
Figure 3 Platform-Independent Model with Stereotypes	2319
Figure 4 Mappings and transformations.	2319
Figure 5 Summary	2622
Figure 6 CSAR/CML Diagram	2723
Figure 7 Search Module	4841
Figure 8 Search Module (Properties Classes)	4842
Figure 9 DB Module	4943
Figure 10 Collection Module	5044
Figure 11 Atom	5245
Figure 12 Bond	5346
Figure 13 Coordinates	5447
Figure 14 Formula	5548
Figure 15 Molecule	5649
Figure 16 Other	5750
Figure 17 MOLUTIL Module	5851
Figure 18 Structure and Translation	5952
Figure 19 Architecture	6053
Figure 20 Processing diagram	6154
Figure 21 Search and DB Interface Modules interacting with each other to perform a 
chemical structure search against database.	6456
Figure 22 General Use Case Scenario	7062



Acknowledgements

The authors of this document wish to express their appreciation to those listed below (in 
alphabetic order) for their contributions of ideas and experience. Ultimately, the ideas 
expressed in this document are those of the authors and do not necessarily reflect the views 
or ideas of these individuals, nor does the inclusion of their names imply an endorsement of 
the final product. Moreover, the inclusion of the CML materials does not necessarily reflect 
that the author of the CML endorses the content of this proposal.

David Benton     
w.david.benton@gsk.com 
Robert DeWitte   
robert.dewitte@acdlabs.com 
James P. Fry
james.fry@pfizer.com 
Saikrishna R. Gangeddula 
saikrishna.gangeddula@mail.com 
Shogo Ito 
shogo@gandolf.acad.emich.edu 
Peter Murray Rust 
peter.murray-rust@nottingham.ac.uk 
Wendy L. Sharp
wsharp@gandolf.acad.emich.edu 
Valery Tkachenko 
valt@acdlabs.ru 
Scott Markel
scott.markel@lionbioscience.com 
Tony Parsons     
tony_parsons@sandwich.pfizer.com 
Henry Rzepa      
h.rzepa@ic.ac.uk 
Richard Scott    
richard.scott@denovopharma.com 







1	Administrative

© Copyright 2002 and 2003 by
Intelligent Solutions and
LION Bioscience
The organizations listed above hereby grant a royalty-free license to the 
Object Management Group, Inc. (OMG) for world-wide distribution of 
this document or any derivative works thereof, so long as the OMG 
reproduces the copyright notices and the below paragraphs on all 
distributed copies.
The material in this document is submitted to the OMG for evaluation.  
Submission of this document does not represent a commitment to 
implement any portion of this specification in the products of the 
submitters.
WHILE THE INFORMATION IN THIS PUBLICATION IS 
BELIEVED TO BE ACCURATE, THE COMPANIES LISTED 
ABOVE MAKE NO WARRANTY OF ANY KIND WITH REGARD 
TO THIS MATERIAL INCLUDING BUT NOT LIMITED TO THE 
IMPLIED WARRANTIES OF MERCHANTABILITY AND 
FITNESS FOR A PARTICULAR PURPOSE.  The companies listed 
above shall not be liable for errors contained herein or for incidental or 
consequential damages in connection with the furnishing, performance 
or use of this material.  The information contained in this document is 
subject to change without notice.
This document contains information, which is protected by copyright.  
All Rights Reserved.  Except as otherwise provided herein, no part of 
this work may be reproduced or used in any form or by any means 
graphic, electronic, or mechanical, including photocopying, recording, 
taping, or information storage and retrieval systems without the 
permission of one of the copyright owners.  All copies of this document 
must include the copyright and other information contained on this 
page.
The copyright owners grant member companies of the OMG 
permission to make a limited number of copies of this document (up to 
fifty copies) for their internal use as part of the OMG evaluation 
process.
RESTRICTED RIGHTS LEGEND.  Use, duplication, or disclosure by 
government is subject to restrictions as set forth in subdivision (c) (1) 
(ii) of the Right in Technical Data and Computer Software Clause at 
DFARS 252.227.7013.CORBA, OMG and Object Request Broker are 
trademarks of Object Management Group.



1.1	Submission Contact Point

Intelligent Solutions:
Juan Esteva
751 Malena Dr.
Ann Arbor, MI 48103
734.996-3628
juan.esteva@intellsolutions.com 
 
LION bioscience:
Mitchell Miller
955 Ridge Hill Lane
Suite 30
Midvale, UT 84047
801.569.1390
mitchell.miller@lionbioscience.com 
 


1.2	Overview to the Material in the Submission

A set of specifications describing the chemical structure access and retrieval processes is 
provided herein.  The specifications are given using class hierarchies, use case scenarios, and 
sequence statements.  Appendix A contains the XML representation for the core CML and 
Appendix B contains sample code. The XMI file for the entire project is provided upon 
request.



1.3	Statement of Proof of Concept

Several code segments have been developed to test the use case scenarios presented below. 
Please also refer to Appendix A and Appendix B.

Search Chemical Structure (against legacy database)
Create CML molecule from a legacy format.
Select the values for use as parameters in a query.
Construct a query using interfaces in Search and DB modules.
Transport resulting query to server.
Vendor Implementation translates the query into the legacy format.
Vendor Implementation performs the search from the database.
Vendor Implementation translates the result into the CML object model.
Transport the result back to client.
Transform the result into a legacy format.

Search Chemical Structure (within collection of CML molecules)
Create CML molecule from a legacy format.
Select the values for use as parameters in a query.
Construct a query using interfaces in Search and Collection modules.
Search -the collection using the query built above.
Transform the result into a legacy format.

Add Chemical Structure
Create CML molecule.
Transport the new molecule to server. 
Vendor Implementation translates the insert query into the legacy format.
Vendor Implementation performs the insert into the database.
Vendor Implementation translates the status of the insert into standard format.
Transport the result back to client.

Update Chemical Structure
Search the database for the chemical structure to be updated.
Replace the corresponding CML molecule with a newly constructed CML molecule.
Transport the update query along with the new molecule to server. 
Vendor Implementation translates the query into the legacy format.
Vendor Implementation performs the update to the database.
Vendor Implementation translates the status of the update into standard format.
Transport the result back to client.



Delete Chemical Structure
Search the database for the chemical structure to be deleted.
Delete the corresponding CML molecule. [Actual deletion is performed by Vendor 
Implementation]
Transport the delete query to server. 
Vendor Implementation translates the query into the legacy format.
Vendor Implementation performs the delete to the database.
Vendor Implementation translates the status of the delete into standard format.
Transport the result back to client.

Translate Chemical Structures between Representations
Forward transformation will create a CML representation from a legacy representation.
Reverse transformation will create a legacy representation from a CML representation.

Manipulate Associated Intrinsic Properties
Given a legacy representation of a molecule or result set, create CML molecule.
Then gain access to its corresponding intrinsic properties – including atoms, bond types, 
connectivity, molecular weight, and molecular formula.




2	Chemical Markup Language within the Context of CSAR
2.1	Introduction

Peter Murray-Rust and Henry Rzepa state that there are a number of problems when trying 
to capture the contextual meaning of chemistry that impede the use of this information. For 
instance, the following extract  from a typical molecule science journal illustrates not only 
how precisely data and information must be represented, but also how much human 
perception is required to translate this information as presented in this (linear) form into e.g. 
a reproducible experiment or a mechanistic interpretation; 
"Thiamin phosphate synthase catalyzes the formation of thiamin phosphate from 4-amino-5- (hydroxymethyl)-
2-methylpyrimidine pyrophosphate and 5-(hydroxyethyl)-4-methylthiazole phosphate. The reaction involves... 
dissociative mechanism...carbenium ion intermediate...and pyrimidine iminemethide observed in the crystal..."
Note the profusion of chemical structure information, concepts and terms, which only a 
trained human chemist could easily process. Quantitative concepts and units are also 
ubiquitous;
"A 500 ul aliquot of 0.8 uM TP synthase in 50 mM Tris-HCl (pH 7.5) and 6 mM MgCl2 incubated at room 
temperature with 50uM CF3HMP-PP."
An even greater degree of human perception is required when handling graphical chemical 
representations which may contain many, often fuzzy and dangerous, human-only semantics 
(e.g. 2D representations of 3D properties, relative stereochemistry, aromaticity, hydrogen 
and other "weak" bonding, use of generic and "R" groups, reaction arrows and mechanisms, 
etc). The challenge therefore is to develop an infrastructure which can be routinely used to 
capture, store and appropriately filter and display such information (see Figure 1). 
Consequently, the ontological chemically-specific information types to support the capturing 
and utilization of this information should include;
?	Molecules and substances 
?	Reactions 
?	Analytical information, especially spectra 
?	Computation and simulation (QM, mechanics, dynamics, etc.) 
?	"Data-centric" concepts (numbers, units, arrays, matrices, etc.) 
?	Specialist software for display, editing, searching etc 
?	Support "adjoining" disciplines such as bio areas, materials science etc.) 

 
Figure 1 The representation problem. (CTL File Formats, MDL Information Systems) 
Figure 1 shows some of the common file formats that are used to represent chemical 
information. Therefore, the tasks that can be identified in implementing an XML solution 
include:
?	The creation of documents from both legacy sources (see Figure 1) of data and de 
novo by humans 
?	The creation and capture of metadata (dictionaries of terms, tables of contents, 
codes, etc.) 
?	Specification of namespaces (a reserved addressing scheme for information) 
?	Human validation of the system (conformance to agreed specifications) 
?	Machine validation of documents (according to a specified and agreed schema) 
?	Document transformation (XSLT) 
?	Rendering and display (XSL-FO, Domain-specific such as molecular representations) 
The design of an XML-based markup language should provide for;
?	a simple, extensible DTD or Schema (do not over-complicate and make it modular) 
?	Agreed semantics 
?	One (or more) agreed and published ontologies 
?	Agreed examples and conformance tests 
?	A community of critical mass 
Appropriate tools for accomplishing this should be identified. These might include; 
?	XML Writers 
?	XML Readers 
?	Legacy converters (difficult because of variation and ambiguity in the original data 
which may require some degree of perception for an accurate conversion) 
?	Validators 
?	Dictionaries 
?	Editors 
Custom written XSLT style sheets and generic editors will do some of these, but a DOM 
(Document object model, which represents a syntax-free abstraction of the data in memory) 
is probably essential for many subjects.
2.2	CML the Chemistry XML 
Markup Languages created using eXtensible Markup Language (XML) are becoming 
accepted as the medium for passing generic and domain-specific information over networks. 
In many cases this information represents persistent information objects which can be re-
used in other applications and other contexts.  Such messages and objects are likely to 
become the medium of chemical information flow, replacing the incompatible "legacy file 
formats" currently used in most applications.

XML makes a considerable contribution to software design because it can describe and 
document data structures. Thus, an XML document represents a tree in which the 
component nodes and attributes are precisely defined and to which any tree-based algorithm 
(enumeration, sorting, walking, diffing, merging, pruning, etc.) can be defined. Because trees 
are common and useful, the W3C XML activity has created an Application Programming 
Interface (API) for XML trees called the Document Object Model (DOM). This encourages 
standardization of terminology, documentation and algorithms. We believe this can form the 
core for component-based computing and describe its extension to chemistry.
2.2.1	The Semantic Web of Chemistry
Our dream is for a "Semantic Web" for chemistry. This concept epitomizes the ability for a 
machine to receive a document over a network without prior agreement on its contents and to carry 
out meaningful processing without human direction. The sender of the message has to represent:

1.	What the components of the document are
2.	Semantics (how the components are to be processed 
3.	The ontology used (what the components mean)

XML solves the first; the encoding, markup, structure and namespace is precisely defined. It 
deliberately does not address the others. 

There is no universally agreed-upon mechanism or sets of mechanisms for encoding 
semantics; the most common is to associate executable code (e.g. plug-ins or Java applets), 
which behaves in a predetermined way when installed on the recipient's machines. Agent 
technology and CORBA are alternative approaches but are normally only deployed for 
projects where there is much prior agreement between senders and recipients. In this paper, 
we propose a Java interface as a first step. 

2.2.2	Creating, reading and editing XML/CML documents
Creating systems for processing structured documents is more difficult than often 
appreciated.  It requires:

?	A precise, agreed specification that has been tested through prototype implementations
?	Tools to write compliant documents
?	Tools to read compliant documents
?	Tools to modify (edit) and validate compliant documents

If the specification is flexible, a read/write/edit tool must cater to all possible variants of 
documents, often a major challenge. We argue here that it must be implemented with a 
DOM, and a CMLDOM is critical to chemical interoperability. 
CML deliberately does not cover all chemistry but concentrates on "molecules" (discrete 
entities representable by a formula and usually a connection table). It supports a hierarchy 
for compound molecules (clathrates, macromolecules, etc.). It also supports reactions and 
macromolecular structures/sequences (though it can interoperate with other macromolecular 
XML languages as they are developed). It has no specific support for physicochemical 
concepts but can support labeled numeric datatypes of several sorts which can cover a wide 
range of requirements. It allows quantities and properties to be specifically attached to 
molecules, atoms or bonds. 
In addition, CML does not provide a robust and all-encompassing solution.  For instance: 
?	In CML, there are no classes to facilitate searches.
?	In CML, there are no classes to facilitate searches for properties. 
?	In CML, there is no separate class for Cartesian coordinates but has Polar class for 
Polar coordinates. 
The goal of the proposed model (see § 2.37) is to make use of the representational and 
translational capabilities offered by CML and complement them with classes that facilitate 
searches and allow the creation of collections (see Figure 2).

 
Figure 2 Intended solution
2.3	Mapping of Models
2.3.1	Platform-Independent and Platform-Specific UML Models
There are multiple ways to transform a platform-independent model (PIM) expressed using 
UML into a corresponding platform-specific model (PSM) expressed using UML:

1.	A human could study the PIM and manually construct a PSM, perhaps manually 
constructing the one-of refinement mapping between the two.
2.	A human could study the PIM and utilize models of known refinement patterns to 
reduce the burden in constructing the PSM and the refinement relationship between 
the two.
3.	An algorithm could be applied to the PIM to create a skeleton of the PSM to be 
manually enhanced by hand, perhaps using some of the same refinement patterns in 
item 2.
4.	An algorithm could create a complete PSM from a complete PIM, explicitly or 
implicitly recording the refinement relationship for use by other automated tools.

Note that the above list does not address the production of executable code from a 
platform-specific model. There also are variations in which the platform-specific model is 
not a semantically rich UML model but, rather, is expressed via a language such as IDL, Java 
interfaces, XML, etc.

Fully-automated transformations are feasible in certain constrained environments. The 
degree to which transformations can be automated is considerably enhanced when the 
following conditions are obtained:

?	There is no legacy to take into account.
?	The model that serves as input to the transformation is semantically rich.
?	The transformation algorithms are of high quality.

It is much easier to generate executable code for structural features (attributes, certain 
associations and similar properties) of a model rather than behavioral features (operations) 
because the behavior of property getters and setters are quite simple.

Automation of transformations is more tractable when the transformation is parameterized, 
(i.e. a human has a pre-defined set of options to select from) to determine how the 
transformation is performed. For example, a system that transforms an UML model to an 
XML DTD could allow some control over how a UML class's attributes are transformed, 
giving a human the choice to put them in an ATTLIST or to put each attribute in a separate 
ELEMENT [as in the UML model].

Some of these points are illustrated by a CORBA-specific model, corresponding to the 
fragment of a platform-independent model shown in Figure 3. Note that this model 
fragment has been annotated with a stereotype to denote that Account is an entity as 
opposed to a process. The model fragment has also been enhanced to indicate that the 
Account number constitutes a unique identifier for Account instances.


 
Figure 3 Platform-Independent Model with Stereotypes
The submitters made use of the capabilities offered by Rational Rose and the Unisys XML 
Metadata Interchange (XMI) to accomplish the transformation between the PIM to the 
PSM. Given two metamodels defined by the OMG, one describing UML and one describing 
Java, we can derive the APIs for these metamodels by applying mappings defined in the JMI 
specificati on. These APIs are the common metadata access APIs for UML tools and Java 
IDEs respectively. Since the respective metamodels describe the functions of the tools well, 
the JMI mapping results in a rich set of APIs for integration. The process is illustrated in 
Figure 4 below:

 

Figure 4 Mappings and transformations.
2.3.2	A formal approach to describing objects 

The Unified Modeling Language (UML) [Grady Booch and Jacobson 1999] is rapidly 
becoming the de facto method for the representation of Object-Oriented (OO) concepts.  
Most implementations have graphical interfaces and provide tools for supporting the 
creation of software. For example, a framework described in UML can be converted to an 
IDL or XML representation or to Java class libraries.  We have been able to transfer our 
CMLDOM Java bytecode to collaborators who have been easily able to use UML-based 
tools to represent and use the class functionality.

Although XML is not directly an OO approach, it is easy and common to map it into OO 
frameworks. XML can support most OO concepts, the most common being:

?	Abstraction. The formalization of the essential properties of a (real-world) entity. 
The abstraction is usually a subset of most human perceptions of the real world. For 
concrete subjects like chemistry, XML elements normally map well into these formalizations. Thus, 
elements in CML represent a core set of universal self-consistent properties of molecules, atoms, etc. 
from which we create corresponding OO abstractions
?	Encapsulation. Each object is a "black box" and clients of it only have access 
through the interface. The implementation is not specified and details cannot normally 
be obtained from the interface, which exposes components through methods. The 
interface is often tied closely to the XML elements and their attributes, though some of the methods 
are related to XML prose descriptions.
?	Aggregation. An object contains other objects (a "has A" relationship). This maps 
directly into the content model of XML DTDs, so that a DTD defines the object aggregation 
model. For CML, a <molecule> element can (directly or indirectly) contain <atom> 
elements, <bond> elements, etc. and a Molecule object can contain Atom objects, Bond 
objects, etc.
?	Polymorphism. Objects that share a common interface can be used in identical 
ways by other objects. Thus, all markup languages share common features (such as 
XMLAttributes) through XML itself. For example, MathMLDOM objects (the W3C's 
specification for mathematics) and CMLDOM objects both implement the W3CDOM interface, so 
either can be "plugged into" a generic W3CDOM-compliant tool. 
?	Inheritance. An object can inherit properties from another (its superclass), and can 
modify or add to them. Inheritance can be from a single superclass or many 
("multiple inheritance"); Java has eschewed multiple inheritance and so do we, using 
interfaces where appropriate. XML has no direct concept of inheritance, though the recent 
XMLSchema specification provides some of this functionality with multiaxiality through facets.  
CMLDOM is designed for inheritance as outlined below.

Other properties which XML does not normally support are:

?	State and Sequence. These relate to dynamic processes in the UML sense. They are 
not currently used in CML.
?	Behavior. This is normally provided through methods. XML is, at heart, a language 
for describing static documents, though many specific markup languages have found 
the need to specify behavior. This is true for chemistry but we defer the detailed behavior-
related methods to the following ontological paper [A Universal approach to Web-based 
Chemistry using XML and CML, P. Murray-Rust, H. S. Rzepa, M. Wright and S. 
Zara, ChemComm, 2000, 1471-1472.]. Since CML V1.0 does not formalize the behavior of 
CML objects, the CMLDOM interface will be the major way of achieving this.
  
 
 
 


2.4	Class Diagram: cml / Overview


 

Figure 554 Summary




 
Figure 665 CSAR/CML Diagram
2.5	Resolution of RFP Mandatory and Optional Requirements
The following is the list of requirements put forward in the Chemical Structure Access and 
Representation (OMG RFP LSR RFP 8).
2.5.1	Mandatory Requirements
Define interfaces that provide uniform access to 2-D chemical structures, associated intrinsic 
properties, and collections thereof, independent of underlying implementation. These 
interfaces shall include the functionality listed below. In all of these, the interfaces must 
handle stereochemistry, the associations between core structures and their salts, and the 
association between tautomeric structures.

Support the lifecycle of single as well as collections of 2-D chemical structures including 
associated intrinsic properties. Lifecycle support includes create, insert (into collections), 
update and delete. Intrinsic properties include, but are not limited to: unique identifier (if 
assigned), atoms, bond types, connectivity, 3-D spatial coordinates, molecular weight and 
molecular formula. Any calculations used must make available the method of calculation 
used, a description of the method, and literature references if available.

The proposal addresses the manipulation of these properties--please refer to the CML module 
definitions, which include the CML, NONCML, and MOLUTIL interfaces.

Search for and retrieve 2-D chemical structures from collections, independent of underlying 
storage implementation, using the following methods:

?	Exact match – search for all 2-D structures in a collection that exactly match a query 
structure.
?	Similarity – search for all 2-D structures in a collection that are loosely related to a 
query structure.
?	Substructure – search for all 2-D structures in a collection that contain a user-defined 
partial chemical structure.
?	Tautomer – search for all 2-D structures in a collection that are tautomers of a query 
structure.
?	Generic (i.e., Markush, R-group, etc.) – search for all 2-D structures in a collection in 
which the identified structures are contained in a query represented by a generic 
structure.
?	Search for intrinsic properties, as defined in 6.5.1.1 

Searches consisting of any combination of the above methods using Boolean and 
comparison predicates (<, >, =, ~, AND, OR, etc.). For example: find all 2-D structures 
that match a query structure and whose molecular weight is equal to some number.

Provide access to representations of 2-D chemical structures in each of the following 
formats regardless of underlying implementation: MOL, SD, and MOL2 files; SMILES 
strings and SLN. Provide the following operations on collections of 2-D chemical structures: 
union, intersection, subtraction, and disjunction.

The proposal addresses the requirement posted above. A series of use case scenarios have been 
documented that take into consideration not only the search criteria listed above but also include a 
composite search using both structure(s), associated intrinsic properties and extrinsic properties.See 
Appendix B and C.

Address the issues related to translating structures and structural queries between 
representations, for example the potential loss of information.

Central to the manipulation of chemical structures in this proposal is the fact that we are considering 
using the Chemical Markup Language (CML) [http://www.xml-cml.org/]. CML acts primarily 
as an abstraction of current chemical models (usually connection-table based). It is intended to be 
able to represent any of them faithfully and to allow syntactic, semantic and ontological conversion. 
We have made some progress with the more common files types (so far, MOL and SMILE 
representations have been translated into CML) but will also be dependent on any standardized 
ontology to emerge from the IUPAC efforts. 

Provide extensibility with respect to the scope defined in section 6.2 and functionality 
defined in 6.5.1 without requiring changes to the interfaces.

The proposal does this by providing parameterized interfaces.

Clearly define all objects in terms of the chemical concepts expressed by them.
Include a glossary defining terms and concepts used in the response.

The submission does this.



2.5.2	Optional Requirements
There are no optional requirements.

2.6	Relationship to existing or emerging OMG Specifications 
?	Model Driven Architecture. While the model provides only the data structures 
needed for the interchange and manipulation of Chemical Representation data and 
specifies no services, it does share the MDA approach of using the UML model as 
the basis of generating the DTD from the XMI.  Given this model it is now possible 
to generate data structures that meet different language standards.
?	XML Metadata Interchange. The model is submitted in XMI format, v1.1, 
generated from Rational Rose Enterprise Edition 2001 using the Unisys Rose XML 
Tools Version 1.3.2 add-in.  
?	Query Service. The Query service was not used,used; instead the submitters suggest 
that it is possible to implement a Query Service for Chemical Representation data 
using Xdb. Xdb is an XML document repository providing structured storage of 
XML data, at present using a Relational Database Management System (RDBMS) 
mapping over PostgreSQL.  XDB provides a fast and scalable XML database 
framework with support for both XML Path Language (XPath) and XML Query 
Language (XQL), and the ability to store XML documents and SAX APIs.
?	Bibliographic Query Service. Although the BQS specification is not directly 
incorporated, the attributes and annotation from associations of the Bibliographic 
Reference can be used in queries to data sources that support the interfaces in the 
BQS specification. 
?	Collection Service. The Collection Service is not used.  Instead, the W3C DOM 
and XML-DEV SAX parsers provide similar capabilities for XML data as the 
Collection Service for IDL data.  The DOM parser is ideal for smaller XML data and 
provides full navigation backwards and forwards.  The SAX parser provides a 
forward only traversal of the data, which is ideal for parsing large XML documents. 
  


2.7	[Chemical} Elements
2.7.1	Molecule
At the heart of our model is the Molecule entity which represents a chemically substance.   
One Molecule contains zero or more sub-Molecules (no limit on the depth with which 
Molecules can be nested.), zero or more Atoms and zero or more Bonds.  Molecules 
nested within a Molecule give our model the ability to accommodate sets of tautomers, 
conformers, residues, mixtures (not required for this submission, but definitely useful in 
chemistry) and other complex chemical entities.
Molecules have the following properties:
2.7.2	Properties
Fuzzy – boolean; when true, this molecule is only intended to query databases and 
probably does not reflect a real chemical substance that a researcher would make or 
isolate and then store in a database.
childRole – string; when this Molecule is associated with another Molecule, childRole 
contains a description of the relationship.
2.7.3	Associations
We use Associations to record the relationship between Molecule and other entities:

?	ParentMolecule – identifies the next Molecule up in the hierarchy.
?	subMolecules – the converse relationship, identifying the set of sub-Molecules under 
this Molecule.
?	molecularAtoms – a pointer to the set of Atoms that belong to this Molecule.
?	immediateParentMolecule – the converse relationship to molecularAtoms, identifying the 
Molecule immediately above the set of Atoms.
?	molecularBonds – a pointer to the set of Bonds that belong to this Molecule.
?	molecularBondAngles – a pointer to the set of BondAngles that belong to this 
Molecule.
?	molecularTorsions – a pointer to the set of TorsionAngles that belong to this 
Molecule.
?	molecularCrystal – a pointer to a Crystal unit, defining the crystal structure of this 
molecule.
?	molecularElectrons -- a pointer to the set of Electrons that belong to this Molecule.

2.7.4	Operations
The following operations define the behavior of Molecule.

?	addAtom – appends an Atom to the set indicated by molecularAtoms.
?	deleteAtom – removes an Atom from the set indicated by molecularAtoms.
?	addBond – appends a Bond to the set indicated by molecularBonds.
?	deleteBond – removes a Bond from the set indicated by molecularBonds.
?	addBondAngle – appends a BondAngle to the set indicated by molecularBondAngles.
?	deleteBondAngle – removes a BondAngle from the set indicated by 
molecularBondAngles.
?	addTorsionAngle – appends a TorsionAngle to the set indicated by 
molecularTorsionAngles.
?	deleteTorsionAngleAngle – removes a TorsionAngleAngle from the set indicated by 
molecularTorsionAngleAngles.
?	getAtomByID – a convenience method for locating an Atom given an (alphanumeric) 
ID.
?	addElectron – appends an Electron to the set indicated by molecularElectrons.
?	deleteElectron – removes an Electron from the set indicated by molecularElectrons.
?	addMolecule – appends a sub-Molecule to the set indicated by subMolecules.
?	addMoleculeWithRole – appends a sub-Molecule to the set indicated by subMolecules 
and specifies its chileRole value.
?	getAtom – given an integer, n, return the nth Atom in this Molecule's molecularAtoms.
?	deleteMolecule – removes a Molecule from the set indicated by subMolecules.
?	hasCoords – indicates where this Molecule has coordinates of a specified type (either 
2D or 3D)
?	getNonHydrogenAtoms – returns an array of Atoms associated with this Molecule 
(molecularAtoms), omitting hydrogens.
?	getCalculatedMolecularMass – returns the molecular weight of this Molecule by 
summing the weights of constituent Atoms.

2.7.5	MoleculeFactory
MoleculeFactory is an interface that defines the behavior of something that creates 
Molecules.  
2.7.6	Properties
none
2.7.7	Associations
none
2.7.8	Operations
?	createMolecule – initialize a Molecule with a set of Atoms and Bonds.
?	(Note: at some point we may want to add an additional input parm of type 
Molecule[])
?	getAtomFactory – return something that can create Atoms.
?	getBondFactory – return something that can create Bonds.

2.8	MoleculeUtil
The interface MoleculeUtil defines the behavior of something that calculates properties for 
a given Molecule.
2.8.1	Properties
none
2.8.2	Associations
none
2.8.3	Operations
?	getCalculatedFormula – automatically generate a molecular formula string.
?	getMolecularWeight – return a double-precision real value for the mass of the 
Molecule.

2.9	Atom
Atom represents a location within a molecule, generally a chemical atom (see glossary). 
A set of properties is defined.  Properties listed as arrays, are generally single-valued for 
registerable structures (see glossary) but may have zero to many values for query structures.
2.9.1	Properties
?	occupancy – whether the position occupied in coordinate space by this Atom actually 
has a chemical atom within it.
?	formalCharges – a value or set of values for this Atom indicating whether it has gained 
(<0) or lost (>0) electrons relative to the uncombined form of the [chemical] 
element.
?	hydrogenCounts – a value or set of values indicating the number of hydrogen atoms 
attached to this Atom.  These hydrogen atoms may be used for substructure 
querying or display.
?	nonHydrogenCounts – the number or allowed numbers of heavy Atoms attached to this 
Atom.
?	atomID – string identifier attached to this Atom.
?	substitutionCount – a query property of this Atom, used exclusively in database queries, 
indicating the number of heavy atoms attached. It can take non-negative integral 
values, plus '*' to indicate 'as drawn.'
?	hydrogenCount – this is a string property, distinct from the array of integers 
hydrogenCounts that corresponds to the MDL molfile field indicating the number of 
hydrogens that must be present.  It can take non-negative integral values, plus '*' to 
indicate 'as drawn.'
?	ringMembership – a query property indicating the number of rings in which this Atom 
participates.  It can take non-negative integral values, plus '*' to indicate 'as drawn.'
?	aromaticity – a Boolean that indicates whether  this Atom has 'aromaticity' as defined 
in the glossary.
2.9.2	Associations
?	xy – relates the Atom to a set of 2D screen coordinates for display
?	xyz – relates the Atom to a set of 3D coordinates specifying location in space.
?	xyzFract -- relates the Atom to a set of fractional crystal coordinates.
?	immediateParentMolecule—relates the Atom to the Molecule that encapsulates it.
?	[chemical] elementTypes – defines that Atom type by relating it to a ChemicalElement 
in a periodic table.
?	ligands – defines a set of Atoms that are bonded to this Atom.
?	atomParities – defines the chirality (if any) of this Atom.
2.9.3	Operations
?	addLigand – appends a new Atom to this Atom's list of attachments.
?	getLigands – returns an array of bonded Atoms.
?	getNonHydrogenLigands – convenience method to provide a list of bonded heavy 
Atoms.

2.10	AtomFactory
The AtomFactory interface specifies the behavior of things that create Atoms. 

2.10.1	Properties
none
2.10.2	Associations
none
2.10.3	Operations
	createAtom – instantiates a new Atom.

2.11	AtomParity
The AtomParity interface defines a generalized pattern of behavior for definitions of atom-
level chirality.  Atom-level chirality means that the atom has some 'handedness' as a 
tetrahedral atom with 4 unlike groups around it.  The setting for this chirality – returned by 
the getStereoCenter method – can be either numeric (as in MDL software) or a string (as in 
Daylight software).

Atom parity is optional in some systems; molecular chirality may be fully specified using 
bond markings.

The AtomParity interface is realized by two classes: NumericAtomParity and 
StringAtomParity, which store the parity as a double-precision real and a string, 
respectively.
2.11.1	Properties
none
2.11.2	Associations
none
2.11.3	Operations
?	equals – compares two different AtomParities which may be in different formats.
?	isChiral – returns true if the Atom has defined assymmetry, for example, as a 
tetrahedral atom with 4 different substituents .
?	isSpecified – returns true when isChiral returns true AND a specific parity is set.
?	getStereoCenter – returns the actual value for this stereocenter.


2.12	Bond
Bond represents a chemical linkage between two Atoms.  Properties listed as arrays are 
generally single-valued for registerable structures (see glossary) but may have zero to many 
values for query structures.

2.12.1	Properties
?	bondOrders – an indication of the strength of the Bond.  Common values include 
SINGLE, DOUBLE, TRIPLE, AROMATIC, RESONANT, ANY.
?	bondLength – the distance between the atoms on either side of the Bond.  This only 
has meaning for 3D structures.
?	Topology – a query property that indicates whether the Bond is part of a ring, 
excluded from ring membership or unspecified.
2.12.2	Associations
?	bondAtoms – relates the Bond to its constituent (pair of) Atoms.
?	bondStereos – specifies the stereochemistry of the Bond.
?	immediateParentMolecule – the container Molecule.

2.12.3	Operations
?	getAtomByID – returns the constituent Atom of this Bond having the specified ID.
?	getOtherAtom – given one Atom, return the second Atom constituent of this Bond.
?	contains—returns true if the specified Atom is part of this Bond.


2.13	Electron
Electron is reserved future use.

2.13.1	Properties
?	none
2.13.2	Associations
?	none
2.13.3	Operations
?	none
2.14	Isotope
Isotope represents one possible configuration of an atom of a given [chemical] element, 
defined by its mass. (Isotopes of a given [chemical] element differ from one another because 
of the number of neutrons.)
2.14.1	Properties
?	mass - inertial property of the atom
?	abundance - fraction of the [chemical]  containing this
2.14.2	Associations
?	none
2.14.3	Operations
?	none
2.15	IsotopeFactory
IsotopeFactory provides a uniform interface for things that create Istopes.
2.15.1	Properties
?	none
2.15.2	Associations
?	none
2.15.3	Operations
?	createIsotope – instantiates an Isotope object.
2.16	AtomFactory
AtomFactory provides a uniform interface for things that create Atoms.
2.16.1	Properties
?	none
2.16.2	Associations
?	none
2.16.3	Operations
?	createAtom – instantiates an Atom object.
2.17	BondFactory
BondFactory provides a uniform interface for things that create Bonds.
2.17.1	Properties
?	none
2.17.2	Associations
?	none
2.17.3	Operations
?	CreateBond – instantiates an empty Bond object.
?	createBondForAtoms – instantiates a Bond given a pair of Atoms.
2.18	NumericAtomParity
NumericAtomParity is a realization of the AtomParity interface that uses numbers to hold 
the parity information.
2.18.1	Properties
?	numericParity – the value of the current Atom's stereochemical parity.
?	STEREO_PARITY_UNKNOWN – an indication that no information is available about 
the current Atom's stereochemical parity, probably because of a lack of experimental 
data.
?	STEREO_PARITY_UNSPECIFIED – an indication that no information has been 
provided about the current Atom's stereochemical parity, probably because the property 
has not been given a value.
?	STEREO_PARITY_EVEN – an indication that the mathematical function describing 
the current Atom's stereochemical parity gives a value divisible by 2.
?	STEREO_PARITY_ODD – an indication that the mathematical function describing the 
current Atom's stereochemical parity gives a value not evenly divisible by 2.
2.18.2	Associations
?	none
2.18.3	Operations
?	none
2.19	StringAtomParity
StringAtomParity is a realization of AtomParity interface that uses text to hold the parity 
information.
2.19.1	Properties
?	SMILESString – the value of the current Atom's stereochemical parity.
?	STEREO_PARITY_L – an indication that the current Atom's stereochemical 
configuration resembles a standard 'L' Atom.
?	STEREO_PARITY_R – an indication that the current Atom's stereochemical 
configuration is classified as 'R' according to the Cahn-Ingold-Prelog rules.
?	STEREO_PARITY_@ – an indication that current Atom has substituents arranged in 
an 'anticlockwise' fashion.
?	STEREO_PARITY_@@ – an indication that current Atom has substituents arranged in 
a 'clockwise' fashion.
?	STEREO_PARITY_S – an indication that the current Atom's stereochemical parity is 
classified as 'S' according to the Cahn-Ingold-Prelog rules.
?	STEREO_PARITY_D – an indication that the current Atom's stereochemical parity 
resembles a standard 'D' Atom.
?	STEREO_PARITY_UNKNOWN – an indication that no information is available about 
the current Atom's stereochemical parity, probably because of experimental limitations.
?	STEREO_PARITY_UNSPECIFIED – an indication that no information has been 
provided about the current Atom's stereochemical parity, possibly because the property 
has not been given a value.
2.19.2	Associations
?	none
2.19.3	Operations
?	none
2.20	 ChemicalElement
ChemicalElement provides a way of describing the properties of a category of Atoms, 
related by having the same number of protons in the nucleus. 
2.20.1	Properties
?	atomicNumber – an integer that identifies this [chemical] element, equal to the number of 
protons in the nucleus.
?	atomicWeight – the gravitational mass of the [chemical]  element relative to carbon (a 
standard).  This number is generally a weighted average of the Isotopes that make up 
the [chemical] element.
?	covalentRadius – one half of the distance between  two singly-bonded atoms of the 
[chemical]  element.
?	symbol – 1-3 letters that are used to represent the [chemical]  element
?	VDWRadius – the closest a non-bonded atom can approach without incurring very 
strong repulsive forces.
?	valenceElectrons – number of electrons in the outermost shell of an atom of this [chemical]  
element.
2.20.2	Associations
None
2.20.3	Operations
None
2.21	BondStereo (new properties only)
2.21.1	Properties
?	BONDSTEREO_CIS – describes a double bond whose principal substituents 
('principal' is loosely defined) are on the same side.
?	BONDSTEREO_TRANS – describes a double bond whose principal substituents 
('principal' is loosely defined) are on the opposite sides.
?	BONDSTEREO_Z – describes a double bond whose principal substituents (defined 
according to Cahn-Ingold-Prelog rules) are on the same side.
?	BONDSTEREO_E – describes a double bond whose principal substituents (defined 
according to Cahn-Ingold-Prelog rules) are on the opposite sides.
?	BONDSTEREO_WEDGE – describes a bond that appears to be 'coming out of the 
page.'
?	BONDSTEREO_HATCH – describes a bond that appears to be 'going into the page.'
?	BONDSTEREO_UNKNOWN – describes a bond whose stereochemistry is not 
available because of experimental limitation.
2.22	Bond (new properties only…)
2.22.1	Properties
?	BOND_ORDER_SINGLE – a bond containing only a single shared pair of electrons.
?	BOND_ORDER_DOUBLE – a bond containing two shared pairs of electrons.
?	BOND_ORDER_TRIPLE – a bond containing three shared pairs of electrons.
?	BOND_ORDER_ANY – a bond that may contain any number of shared electron pairs.
?	BOND_ORDER_RESONANT – a bond whose electronic configuration cannot be 
described by an integral number of shared electron pairs.
?	BOND_ORDER_AROMATIC – a bond whose properties are in between those of 
single and double bonds and also reflects a strong interaction with adjacent aromatic 
bonds.
2.23	ChemicalElementFactory
ChemicalElementFactory provides a uniform interface for things that create 
ChemicalElements.
2.23.1	Properties
None
2.23.2	Associations
None
2.23.3	Operations
?	createChemicalElement – instantiates a ChemicalElement having a specified atomic 
number.
?	createIostopicallyLabeledChemicalElement – instantiates a ChemicalElement having a 
specified atomic number and a given set of Isotopes.
2.24	IsotopeSet
An IsotopeSet is a grouping of Isotopes that define the composition of a sample of the 
[chemical] element.
2.24.1	Properties
None
2.24.2	Associations
None
2.24.3	Operations
?	getMostCommonIsotope – returns the most prominent Isotope of the current [chemical] 
element.
?	getIsotopeAbundance – returns the fraction of a given Isotope within the set.
?	getAveragedWeight – returns the mean of the weights of the Isotopes making up the set.
2.25	IsotopeSetFactory
IsotopeSetFactory provides a uniform interface for things that create IstopeSets.
2.25.1	Properties
None
2.25.2	Associations
None
2.25.3	Operations
createIsotopeSet – instantiates an IsotopeSet.
2.26	PeriodicTable
A grouping of ChemicalElements that provides a complete representation of all the atom 
types used in some chemical system.  
2.26.1	Properties
None
2.26.2	Associations
None
2.26.3	Operations
?	getElement – given an atomic symbol, return the corresponding ChemicalElement.
?	getElementByAtomicNumber – given a number, return the ChemicalElement with that 
many protons.
?	getSymbol -- given a number, return the atomic symbol for the ChemicalElement with 
that many protons.
2.27	PeriodicTableFactory
PeriodicTableFactory provides a uniform interface for things that create PeriodicTables.
2.27.1	Properties
None
2.27.2	Associations
None
2.27.3	Operations
createPeriodicTable – instantiates an empty PeriodicTable.


2.28	BondStereo
Interface BondStereo defines the behavior of classes that define a Bond's stereochemistry.  
2.28.1	Properties
None
2.28.2	Associations
None
2.28.3	Operations
?	equals – provides a means of comparing the stereochemistry of two Bonds.
?	getValue – returns the value of the current Bond's stereochemistry.


2.29	Coordinate2
Coordinate2 represents a set of x, y coordinates which specify the placement of an atom on 
a 2D display grid. 
2.29.1	Properties
?	x – the abscissa of this coordinate set.
?	y – the ordinate of this coordinate set.
2.29.2	Associations
None
2.29.3	Operations
None
2.30	Coordinate3
Coordinate3 represents a set of x, y, z coordinates which specify the placement of an atom 
on a 2D display grid. 
2.30.1	Properties
?	x – the abscissa of this coordinate set.
?	y – the ordinate of this coordinate set.
?	z – the depth coordinate of this set.

2.30.2	Associations
None

2.30.3	Operations
none
2.31	AbstractAngle
A generalization of the behavior of bond (or 3-center) angles and torsional (or 4-center) 
angles.  Realizations: BondAngle and TorsionAngle.
2.31.1	Properties
?	angle – the measure of this item
2.31.2	Associations
?	units – specifies the measurement type of this angle (generally, radians or degrees)
?	angleAtoms – a pointer to the (3 or 4) Atoms that constitute this angle.
2.31.3	Operations
None

2.32	Formula
Formula represents a listing of the atoms and quantities within a Molecule.  It is built of 
formulaElements (q.v.) blocks.  Since there are multiple ways of calculating the molecular 
formula for a given molecule, (depending on, for example, counting salt fragments that are 
not explicitly included in the structure), there may be more than one Formula for a given 
Molecule and therefore, Formulas may contain sub-Formulas. 

2.32.1	Properties
?	overallCount – total number of atoms in this formula.
2.32.2	Associations
?	formulaFormulaElements – relates the Formula to the constituent formulaElements.
2.32.3	Operations
?	createFromString – initializes a Formula from text input.
?	addElement – append a [chemical] element (including a count) to the Formula. 
?	getElementTypes – returns an array representing the types of atoms present
?	getElementCounts – returns an array of numbers representing the number of times 
each type of atom (from the array returned by getElementTypes) occurs in the 
Formula. 
?	addFormula – appends a Formula representation to this Formula.
?	getFormulaAtoms – returns an array of Atoms represented herein.
?	deleteAllFormulas – remove all sub-Formulas from this Formula.
?	deleteFormula – remove one  sub-Formula from this Formula.
?	getFormalCharge – return the surfeit or deficit of electrons in this Formula.
?	setFormalCharge – change the surfeit or deficit of electrons in this Formula.
?	getCalculatedMolecularMass – returns the weight generated for this Formula.
?	getFormattedString – generate a printable text representation of Formula, given a 
display mode.

2.33	FormulaElement
It refers to a combination of atom type ([chemical] element) and count.  It is used in defining 
molecular formulas.
2.33.1	Properties
?	ElementType – pointer to an entry in a periodic table defining a kind of atom.
?	Count – the number of times an ElementType occurs in a formula unit.
2.33.2	Associations
None
2.33.3	Operations
None
2.34	Crystal
Bond represents a chemical linkage between two Atoms.  
2.34.1	Properties
?	ACELL – parameter of a crystal unit
?	BCELL – parameter of a crystal unit
?	CCELL– parameter of a crystal unit
?	ALPHA– parameter of a crystal unit
?	BETA– parameter of a crystal unit
?	GAMMA– parameter of a crystal unit
?	Z-FLOAT– parameter of a crystal unit
?	Z-INT– parameter of a crystal unit
?	SPACEGROUP– parameter of a crystal unit
2.34.2	Associations
None
2.34.3	Operations
None

2.35	Query and Collection

?	ChemSearchEngineManager needs to be very generic to provide lookups for various 
legacy databases.
?	To simplify the interface of ChemSearchEngin, the association between SearchCriteria 
and ChemSearchEngine was dropped.  SearchCriteriaGroup will have one or more 
SearchCriteria.
?	SearchableProperty will have an attribute called intrinsic_flag to indicate if a property is 
intrinsic or extrinsic.
?	SearchCriteria interface will have an optional attribute called factor to be used for, i.e., 
to indicate a similarity factor for SIMILAR ACCURACY QUALIFIER.
?	A row in ResultSet contains a column for Molecule (from CML) and as many columns as 
queried extrinsic properties.  Intrinsic properties will be included as attributes of the 
Molecule.
?	Split() in ChemSearchEngine splits the query into sub-queries: intrinsic and extrinsic 
(external properties) queries.  The latter may be stored in a conventional rational 
database.
?	ResultSet can be updateable for insert, delete, and update.

?	The following diagram shows Search and Collection Modules interacting with each other 
to perform a chemical structure search within a collection.
  
 
2.36	Class Diagrams

Search Module ? Holds interfaces that are used to search against legacy databases as well as 
collections.




 
 
Figure 776 Search Module
 
Figure 887 Search Module (Properties Classes)


DB Module ?Holds interfaces to interact with legacy databases
 
Figure 998 DB Module


Collection Module?represents a collection of chemical elements.
 
Figure 10109 Collection Module
The collection module makes use of the Java API public class Collections  which extends Object. This class consists exclusively of static 
methods that operate on or return collections. It contains polymorphic algorithms that operate on collections, "wrappers", which return a 
new collection backed by a specified collection, and a few other odds and ends. The elements of these collections are Molecules. 




Atom

 
Figure 111110 Atom
 
Figure 121211 Bond
 
Figure 131312 Coordinates
 
Figure 141413 Formula

 
Figure 151514 Molecule

 
Figure 161615 Other


 
 
Figure 171716 MOLUTIL Module


2.37	 Recommendation
The following figures provide a pictorial description of the proposed recommendation.

 
Figure 181817 Structure and Translation
 
 
Figure 191918 Architecture

The DB Interface module has been specified in a database independent manner. Consequently, any 
commercial database can be used to persist the objects mentioned throughout this standard.
 


Figure 202019 Processing diagram
2.37.1	Conformance Points

The UML model included here subsumes all the elements of the CSAR model including the 
CMLCore schema and the additional processing classes. Consequently, the UML model is 
normative. 
2.37.12.37.2	Details
The following paragraphs provide detailed descriptions of how the recommendation satisfies each of 
the requirements of the RFP.

Define interfaces that provide uniform access to 2-D chemical structures associated intrinsic 
properties, and collections thereof, independent of underlying implementation. These interfaces 
shall include the functionality listed below. In all of these, the interfaces must handle 
stereochemistry, the associations between core structures and their salts, and the association 
between tautomeric structures.

Support the lifecycle of single as well as collections of 2-D chemical structures including associated 
intrinsic properties. Lifecycle support includes create, insert (into collections), update and delete. 
Intrinsic properties include, but are not limited to: unique identifier (if assigned), atoms, bond types, 
connectivity, 3-D spatial coordinates, molecular weight and molecular formula. Any calculations 
used must make available the method of calculation used, a description of the method, and literature 
references if available.

Not applicable this time.

?	Search for and retrieve 2-D chemical structures from collections, independent of underlying 
storage implementation, using the following methods:
?	Exact match – search for all 2-D structures in a collection that exactly match a query structure.
?	Similarity – search for all 2-D structures in a collection that are loosely related to a query 
structure.
?	Substructure – search for all 2-D structures in a collection that contain a user-defined partial 
chemical structure.
?	Tautomer – search for all 2-D structures in a collection that are tautomers of a query structure.
?	Generic (i.e., Markush, R-group, etc.) – search for all 2-D structures in a collection in which the 
identified structures are contained in a query represented by a generic structure.
?	Search for intrinsic properties, as defined above
?	Searches consisting of any combination of the above methods using Boolean and comparison 
predicates (<, >, =, ~, AND, OR, etc.). For example: find all 2-D structures that match a query 
structure and whose molecular weight is equal to some number.

Processing Classes Involved
Search Module?Holds interfaces that are used to search against legacy databases as well as 
collections.
DB Interface Module?Holds interfaces to interact with legacy databases.
Collection Module?Represents a collection of chemical elements.


The following diagram shows Search and DB Interface Modules interacting with each other to perform a 
chemical structure search against database.

 
 
Figure 212120 Sea Search and DB Interface Modules interacting with each other to perform a 
chemical structure search against database.


3	Glossary 
Aromaticity
This is a quality possessed by many, many common compounds, from benzene to 
phenylalanine in which multiple double bonds, 'conjugate,' sharing electrons.  This sharing 
produces a structure of lower energy than one in which the double bonds re isolated.  (The 
above is very simple explanation of a common phenomenon.  For more information, see any 
basic text on organic chemistry.)
Different software systems define aromaticity differently.  The biggest differences are 
whether aromaticity is specified in atom types or in bonds or perceived from the 
arrangement of bonds.  Daylight's SMILES notation generally specifies aromatic atoms using 
lower-case letters.  For example, benzene (an aromatic ring of 6 carbon atoms) is defined as 
'c1ccccc1'.   Because the atoms are denoted with lower-case letters, they are distinguished 
from cyclohexane (a non-aromatic or 'aliphatic' ring of carbon atoms (C1CCCCC1).
MDL's molfile, on the other hand, defines aromatics from an arrangement of alternating 
single and double bonds within a ring of appropriate size.
Still other systems explicitly define aromaticity using explicitly designated aromatic bonds.
To make matters even more complicated, MDL supports an aromatic bond type that can be 
used for bonds within a molfile that can be used to query a database but not registered. 
Atom
The smallest particle of an [chemical] element that can exist either alone or in combination, 
retaining any properties of the [chemical] element.  We extend this definition to include points 
in space (that can be used to define the position of other points); 'superatoms' or atoms that 
represent a collection of other atoms.
Bit string
A contiguous set of characters consisting entirely of 1s and 0s.  A bit string can be used to 
encode a good deal of information in a compact way.
Bond 
A chemical link between two atoms.  Bonds are classified as ionic (transfer of electrons from 
one atom to another); covalent (sharing of electrons, generally an equal number from each 
atom); dative (sharing of two electrons from a single atom); or hydrogen (attraction of 
electron-starved hydrogen atom to electron rich heavy atom.)
Charge
A deficiency or excess of electrons on a particular object, giving rise to a positive or negative 
charge, respectively. (www.allwords.com)  Molecules can carry charges, which are often 
attributed to specific atoms.
Chiral
Adjective applied to a molecule that cannot be superimposed on its mirror image. ('Chirality' 
is the corresponding noun.)
An example of a chiral structure for 2-chloro-2-iodo-butane is shown below:


The compound has a mirror image that cannot be superimposed and is therefore termed its 
enantiomer.

Connection table
A means of representing the atoms contained within a molecule and the bonds that hold 
them together.

Counterion
A set of one or more bonded atoms, with opposite charge and generally smaller size, that 
accompanies another charged set of bonded atoms.
Cyclic, acyclic bonds
When chemical bonds occur within a ring, they are termed 'cyclic.'  'Acyclic bonds' by 
contrast, occur in open chain structures.
electrons (p or pi)
Molecules containing double or triple bonds typically have electrons that project outside the 
line between the atoms in the bond.  When more than one double or triple bond are in close 
proximity, the electrons in the pi bonds interact and spread out over all the atoms involved.
Fingerprints
Fingerprints are bit strings that are based on features of a chemical structure.  In this regard, 
they are similar to Structural keys (q.v.) Fingerprints are different from keys in that the bits 
they contain are typically 'folded over' or combined (using a logical OR) with one another to 
reduce the size of the string.
Heteroatoms
Atoms that are neither carbon nor hydrogen are considered 'heteroatoms' and are often 
handled differently by software systems.
Markush structure
It is common to represent chemical structures as a common core containing marked 
substitution sites, plus a set of possible structures for each substitution point.  These 
Markush structures can be used in several ways:
1)	to represent a set of compounds analyzed in order to determine 
the effect of varying substituents on compound activity (SAR, short of 'Structure Activity 
Relationships)
2)	to represent a set of compound produced using combinatorial 
techniques (synthesized by serially attaching different chemical groups to a common core)
3)	to produce a fine-tuned substructure query
In this document, we use one Assembly to represent the core, (atoms of type 'R' designate 
the substitution points), plus one additional Assembly for each R-Group
Molecule
The smallest particle into which an [chemical] element or a compound can be divided without 
changing its chemical and physical properties; a group of like or different atoms held 
together by chemical forces. (www.allwords.com) Generally, composed of atoms held 
together by bonds.
A molecule can represent:
1.	an entire chemical entity
2.	one portion of a complex chemical entity (such as a mixture, set of tautomers or 
conformers)
3.	a collection of other molecules as defined in 2) above
Orbital
A subdivision of a nuclear shell containing zero, one, or two electrons (m-w.com)
Query structure
Most chemical software systems require structures to meet certain requirements in order to 
be entered into a database or used for calculation.  A structure that meets these criteria is 
classified as 'registerable.'  (Generally, atom types must correspond to entities in the periodic 
table and bond types must be well defined.)
Additionally, the software will allow a chemist to draw structures that do not meet the 
criteria for registration but can be used to retrieve molecules out of a database using 
substructure searching (see below).  An example of a structure that is valid for query but not 
registration is one with a just one bond designated as aromatic.
Radical
An atom or a group of atoms with at least one unpaired electron. 
(www.allwords.com)
singlet – a radical with two unpaired electrons whose spins are opposite
doublet – a radical with a single unpaired electron
triplet – a radical with two unpaired electrons whose spins are aligned
Registerable structure
A chemical structure that meets the criteria for inclusion in a repository.  (Compare with 
'Query structure' above)

Similarity Search
A chemical data query in which a user seeks compounds that resemble a given structure 
without necessarily having a substructure (q.v.) match.  The resemblance is often intuitive 
to a chemist but has a mathematical basis in terms of common features or properties.  
Similarity searches typically include a cutoff value X so the user sees only structures having 
X% or more similarity to the query structure.
The mathematics of similarity searching require:

1.	A means of evaluating individual chemical structures.  Generally, this involves 
computing keys (q.v.) or fingerprints (q.v.).
2.	A metric for comparing keys or fingerprints from two structures.  The most 
common of these is the Tanimoto coefficient (q.v.)
At search time, a user-supplied query structure is compared with every other structure in the 
database and those with a similarity metric greater than the cutoff are considered 'hits.'  (Bit 
operations are typically fast enough to make large number of comparisons practicable.)
Spin state
A way of characterizing the angular momentum of electrons.  Individual electrons may spin 
'up' or 'down.'  Two or more electrons may have their spins parallel or antiparallel.
Stereochemistry
Studying the effect of configuration of atoms around assymetric atoms and bonds.
Structural keys
When compounds are registered into most chemical search software systems, the structures 
are scanned for the presence of predefined features, such as heteroatoms (q.v.) or 6-
membered rings.  Each key sets a bit within a string that may be hundreds of characters long.  
These keys are used for substructure and similarity searching. 
Substructure
One chemical structure is said to be a substructure of another if the first structure can be 
located within the second.  (The second is said to be the superstructure of the first.)  All 
structures are substructures of themselves. 

A substructure search scans a database for all substructural matches.

Tanimoto coefficient
Mathematical formula for evaluating the similarity of two structures 
SA,B = c/[a + b - c]

Where 
SA,B = 'similarity of structures A and B'
c = number of features in common between the given property in the two structures.  (In 
the case of structure keys or fingerprints, this means the number of ON bits when the two 
bit strings are ANDed.)
a = number of features ON in structure A
b = number of features ON in structure B
[From J. Chem. Inf. Comput. Sci. 1998, 38, 983-996]
Tautomer
'One of two or more structural isomers that exist in equilibrium and are readily converted 
from one isomeric form to another.' From http://www.bartleby.com/65/ta/tautomer.html .  
An illustration of the tautomers of ethyl acetoacetate is shown:


 


Valence
The number of bonds an atom has to other atoms.
 



Appendix A: UML Use Cases
Actor Catalogue


Actor
Definition
Client
This is a logical client that represents any given system 
that will interface with the provided IDLAPI. 



Master Use Case 




 

Figure 222221 General Use Case Scenario


Detailed Use Case Scenarios
Add Chemical Structure

Desired 
Outcome:
A chemical structure is inserted into database.
Entered When:
Actor invokes the insert function.
Finished When:
A chemical structure is inserted into database and 
a confirmation is returned to the actor.
Description:
The purpose of this function is to insert a chemical 
structure into database.  A chemical structure can 
be specified using CML.
 
	
Data Elements:
Name
Description
Chemical structure
A chemical structure to be inserted 
into database.


Add Chemical Structure – Add Unique Chemical Structure

Desired 
Outcome:
A chemical structure is inserted into database 
only if it will not be a duplicate.
 Entered When: 
Actor invokes the insert_if_not_duplicate 
function.
Finished When:
A chemical structure is inserted into database and 
a confirmation is returned to the actor.  A 
collection of duplicated chemical structures is 
returned if any duplicate is found.
Description:
The purpose of this function is to insert an unique 
chemical structure into database.  An exact 
match search will be performed to ensure a 
uniqueness of the input chemical structure.  A 
chemical structure can be specified using CML.

Data Elements:
Name
Description
Chemical structure
A chemical structure to be inserted 
into database.




Update Chemical Structure

Desired 
Outcome:
An existing chemical structure data in database 
is updated.
Entered When:
Actor invokes the update function.
Finished When:
An existing chemical structure data in database 
is updated and a confirmation is returned to the 
actor.
Description:
The purpose of this function is to update an 
existing chemical structure data in database.  A 
chemical structure can be specified using CML.


Data Elements:
Name
Description
Chemical structure
A chemical structure to replace an 
existing data.

Delete Chemical Structure

Desired Outcome:
An existing chemical structure in database is 
removed.
 Entered When:
Actor invokes the delete function.
Finished When:
An existing chemical structure in database is 
removed and a confirmation is returned to the 
actor.
Description:
The purpose of this function is to delete an 
existing chemical structure in database.  A 
chemical structure to be deleted can be 
specified using CML/ID.

Data Elements:
Name
Description
Chemical structure/ID
Identifier of chemical structure to be 
deleted.




Search Chemical Structure

Desired Outcome:
Collection of structures meeting the search 
criteria is returned.
Entered When:
Actor invokes the search function.
Finished When:
A collection of structures is returned.  An empty 
collection is returned if no structure meets the 
specified criteria.
Description:
The purpose of this function is to identify a 
collection of structures meeting the specified 
criteria.  Search criteria can be specified by a 
query string and/or structure (CML).  A 
returned collection is in CML, which can be 
transformed into another format such as 
SMILES, MOL, SLN.

Data Elements:
Name
Description
Query string
Specifies the search criteria.

Comments:

The following are sample query strings used in the MDL ISIS/Host:

molstructure = [{mjk38smasd903kqlads90rmlw9masksoaskdoq}]
molstructure tautomer 
[{mjk38smasd903kqlads90rmlw9masksoaskdoq}]
pdnum = '0123456-0000'
pdnum like '0123456-%'
mol.weight < 500

The following are sample query strings used in the Daylight DayCard 
(Daylight Chemistry Cartridge for Oracle).
http://www.daylight.com/meetings/mug2000/Delany/cartridge.html

select * from medium where exact(smiles, 'O=c1ccocc1') = 1
select count(smiles) from large where contains(smiles, 
'>>O=c1c(C)cocc1') = 1;

Exact Match

molstructure = [{mjk38smasd903kqlads90rmlw9masksoaskdoq}]
select * from medium where exact(smiles, 'O=c1ccocc1') = 1


Similarity

molstructure = [{mjk38smasd903kqlads90rmlw9masksoaskdoq}] and factor 
= 0.7


Substructure

molstructure sss [{mjk38smasd903kqlads90rmlw9masksoaskdoq}]
select * from medium where contains(smiles, 'O=c1ccocc1') = 1


Tautomer

molstructure tautomer [{mjk38smasd903kqlads90rmlw9masksoaskdoq}]
select * from medium where tautomer(smiles, 'O=c1ccocc1') = 1

Comments: (from Tripos' RFI) There are potentially different methods of 
working with tautomers, either recognizing the potential for rearrangement 
at registration into the chemical database with flags that highlight the areas 
where tautomerism can take place.  An alternative method would be to 
register the tautomers as different compounds and then flag their tautomeric 
relatives.  The latter is what PD does.

Generic

Comments: i.e. Markush, R-group, etc.

Intrinsic Properties

mol.weight < 500


Combination

molstructure sss [{mjk38smasd903kqlads90rmlw9masksoaskdoq}] and 
mol.weight < 500



Get Intrinsic Properties

Desired Outcome:
An intrinsic property of a chemical structure is 
returned.
 Entered When:
Actor invokes the access (get) function.
Finished When:
An intrinsic property of a chemical structure is 
returned to the actor.
Description:
The purpose of this function is to query an 
intrinsic property of a chemical structure.

Data Elements:
Name
Description
Intrinsic property
Name of an intrinsic property to be 
queried.


Translate Chemical Structure from One to Another

Desired Outcome:
A type of chemical structure representation is 
transformed into another type.
 Entered When:
Actor invokes the translate function.
Finished When:
A type of chemical structure representation is 
transformed into another type, which is then 
returned to the actor.
Description:
The purpose of this function is to transform the 
representation of one chemical structure into 
another representation.

Data Elements:
Name
Description
Structure representation type
Source representation.  
Structure representation
Chemical structure to be 
translated.
Structure representation type
Destination representation.  



Appendix B: Java Code
This appendix contains a sampling of the Java code generated for the project. 
The complete Java API and XMI files can be found at:

http://www.intellsolutions.com/serv01.htm



package pim_chem.cml;


public class BondStereo 
{
   public String BONDSTEREO_CIS = "CIS";
   public String BONDSTEREO_TRANS = "TRANS";
   public String BONDSTEREO_Z = "Z";
   public String BONDSTEREO_E = "E";
   public String BONDSTEREO_WEDGE = "WEDGE";
   public String BONDSTEREO_HATCH = "HATCH";
   public String BONDSTEREO_UNKNOWN = "UNKNOWN";
   public static final String STRING = "STRING";
   public static final String SYMBOL = "SYMBOL";
   public static final String CIS = "C";
   public static final String TRANS = "T";
   public static final String WEDGE = "W";
   public static final String HATCH = "H";
   public static final String UNKNOWN = "U";
   
   /**
    *   
    * @param obj
    * @return boolean
    * @roseuid 3B48CBF4033F
    */
   public boolean equals(Object obj) 
   {
    return true;
   }
   
   /**
    *   
    * @return java.lang.String
    * @roseuid 3B48CBF4034A
    */
   public String getValue() 
   {
    return null;
   }
}

//Source file: F:\\OMG\\pim_chem\\search\\SelectPropertyGroup.java

package pim_chem.search;


public interface SelectPropertyGroup 
{
   public SearchableProperty theSearchableProperty;
   
   /**
    * @param arg0
    * @return boolean
    * @roseuid 3A120EE20091
    */
   public boolean addSelectProperty(SelectProperty arg0);
   
   /**
    * @param arg0
    * @return boolean
    * @roseuid 3A120F2C0343
    */
   public boolean removeSelectProperty(SelectProperty arg0);
   
   /**
    * @return boolean
    * @roseuid 3A120F4A03C9
    */
   public boolean removeAll();
}


Appendix C: Use Cases for Chemistry
1)	Structure search
A)	using ISIS/Draw to sketch a molecule for a substructure search of a Daylight database
I simple structure (e.g., biphenyl)

II more complicated structure (disconnected fragments; atom lists; substitution counts; 
charges, etc.)
a)	two fragments (twofrag.mol)
b)	two fragments + atom list (tfraglist.mol)
c)	two fragments + atom list + substitution counts (tfsub.mol)



III R Group Query
IV S-Group data

A)	using the Ertl Java editor to generate a SMILES to search an MDL database
B)	browsing the hits
I Browsing regular structures
II Browsing structures with polymeric constructs
III Browsing structures to which user does not have rights

2)	Use ChemSymphony to search Web-based Available Chemical Directory (ACD) system for 
suppliers of a given set of structures
A)	superstructures of aromatic acid chlorides – very simple substructure to find a large class of 
compounds (aromacchlor.mol)
B)	p-nitrobenzoic acid – search for a specific compound (pnitrobenz.mol)

3)	Looking for similar compounds
A)	ISIS/Draw front-end to Unity (or RS3) database back end
B)	ChemSymphony front-end to MDL database

4)	Registration
A)	using ISIS/Draw to generate a molfile for registration into Daylight (convert to SMILES for 
direct chemical registration, as well as saving the molfile to an Oracle field)
I Simple structure; everything in molfile translates to SMILES
II Complex structure (charges, valence) but properties do translate
III Parts of the structure (brackets, S-Group data) do not translate to SMILES 
B)	using ChemSymphony to generate structures for registration to a Unity database



5)	Registration correction
A)	database is MDL; need to locate molecule by ID, replace structure.  Drawing tool is 
ChemSymphony
I simple structure – no loss of information
II more complicated structure – all information can be translated
III very complicated structure – some information does not map
  D. H. Peapus, H. J. Chiu and N. Campobasso, Biochemistry, 2001, 40, 10103-10114.
  	This section is taken verbatim from: Peter Murray-Rust and Henry S. Rzepa "Chemical Markup, XML 
and the World-Wide Web.  Part II: Information Objects and the CMLDOM"











15