http://www.nlm.nih.gov/factsheets.dir/ISM.html (Einblicke ins Internet, 10/1995)

UMLS® Information Sources Map (ISM)

Background

Information Sources Map Design Goals

The Information Sources Map (ISM) is a knowledge source which makes use of the UMLS Metathesaurus and UMLS Semantic Network to support the following functions:

) Determine which information sources are most likely to be relevant to a particular inquiry.
) Supply human-readable information to users about the scope, probable utility, and access conditions of particular sources.
) Automatically connect to relevant information sources.
) Automatically conduct a successful retrieval session on one or more sources.

The current version of the ISM file contains records for 13 information providers and 63 distinct information sources. The information sources include 44 provided by the NLM and 19 provided by non-NLM sources. The information sources are varied and include not only major bibliographic databases covering biomedical research, clinical practice, and bioethics, but also: diagnostic expert systems such as AI/RHEUM (NLM), DxPlain (Massachusetts General Hospital, Iliad (University of Utah), and Quick Medical Reference (QMR, CAMDAT Corporation); and factual databases concerned with genetics, protein and nucleic acid sequences, drugs, toxicology, and environmental health.

Elements of the Information Sources Map Record

Records in the current ISM address the first two of the functions listed above: determination of relevance (source selection) and descriptive data about the sources. Four elements are used to index the conceptual scope of information sources:

Medical Subject Headings (MeSH) terms from the UMLS Metathesaurus
MeSH Subheadings which denote the contexts in which the main MeSH headings are applicable
Semantic Types from the UMLS Semantic Network
Semantic Type Relations, which link together two semantic types with a Relation defined within the Semantic Network.

The application of indexing terms to information sources is similar to the indexing of the biomedical literature, however in contrast to the indexing of articles, where index terms such as MeSH terms are applied at the most specific level possible, the index terms applied to an entire database or other information resource are chosen at the level of highest generality which is applicable across the entire resource.

The most novel of the indexing elements in the Information Sources Map is the attempt to portray the conceptual scope of an information resource by the use of UMLS Semantic Type Relations. A Semantic Type Relation can be thought of as two nouns (Semantic Types) connected by a verb (Relation), where the Types and Relations are drawn from valid values defined in the Semantic Network. For example, the type relation:

Hazardous or Poisonous Substance|causes|Pathologic Function

is used to describe the content of a toxicology database. This approach should expedite the development of software tools which allow the use of natural language queries to discover relevant sources.

Those portions of the Information Sources Map concerned with automatically connecting to and retrieving information from a given source (the procedural components) remain under development. Standards and methods for network-distributed information selection and retrieval are evolving rapidly, and the NLM intends to base further developments of the ISM on carefully-selected communications and software standards which appear both technically sound and likely to become widely available. Some of the systems which have been evaluated in this respect include WAIS, Project Gopher (University of Minnesota), World-Wide Web (CERN, Geneva), Hytelnet, Knowbot technology, and various Internet-based library retrieval systems. Certain of the emerging communications standards underlying a number of these systems, in particular ANSI Z39.50 and ISO/OSI ASN.1, are likely to appear as part of the procedural aspects of the ISM; it is also likely that the ISM will be able to interact with several of the previously mentioned network information retrieval systems.

For maximum utility, the ISM must be able to be used by all information collectors and vendors to describe their biomedical information resources. The ISM aims to create a distributed maintenance environment in which each information provider builds and maintains the ISM components which depict and support connection to its resources; software tools with graphical user interfaces are under development to aid ISM participants in preparing entries for the ISM database. After initial development and testing are complete, wide participation in the creation of a shared biomedical Information Sources Map will be sought.

Distribution Format

The current version of the Information Sources Map is available in IBM PC-compatible file format as part of the 1992 edition of the UMLS Knowledge Sources CD-ROM, which is an ISO 9660 disc. ISM-related files include the Information Sources Map, a file describing ISM field definitions and formats, and a file containing valid values for ISM fields which require a controlled vocabulary.

The ISM data files are represented as ASCII relational tables, in a format analogous to the UMLS Metathesaurus and Semantic Network files.

Application Procedures

Those who wish to obtain copies of UMLS products are required to sign a one-year experimental agreement with the NLM. To obtain a packet with sample records, documentation, and copies of the experimental agreement, send a written request to:

Betsy L. Humphreys
UMLS Project Officer
National Library of Medicine
8600 Rockville Pike
Bethesda, MD 20894

Other personnel associated with the ISM Project:

Other fact sheets in the UMLS series:

Printed copies of this document are available from:

Office of Public Information
National Library of Medicine
8600 Rockville Pike
Bethesda, Maryland 20894

NLM HyperDOC / Information Sources Map FactSheet / April 1994