NIFSTD_Poster_Description_RevisedMM.doc - Download NIF ...

NIFSTD: Towards a Comprehensive Ontology for Neuroscience

Maryann E. Martone, Amarnath Gupta, Jeffry S. Grathe, Sarah M. Maynard,
Stephen D. Larson, Fahim T. Imam

University of California, San Diego, California

Abstract

As a core component of N
euroscience Information
Framework (NIF)

project
,

NIF

Standard (NIFSTD)
was
envisioned as a

set of modular ontologies that
provid
e
s

a
comprehensive collection of terminologies
to

describe neuroscience

data and resources.

We
presen
t

on the structure, d
esign principles and
current state of NIFSTD.

Introduction

The NIFSTD
is

a

critical
component

of
the
NIF
project

(
http://neuinfo.org
)
, designed

to enable an
effective concept
-
based search mechanism
against a
diverse collection of neuroscience resources.
The
core ontology
covers the major domains of
neuroscience
data (e.g.,
anatomy, cell

type,
techniques), as well as
digital resources

being created
throughout the neuroscience community. The
NIFSTD has

been assembled in a form that promotes
easy extension and modification
during

its evolution.

Basic Structure and Design Principles

The NIFSTD is constructed according to a set of best
practices that closely follow
those

es
tablished by the
Open Biological Ontology (OBO) community. The
ontologies in NIFSTD are built
in a modular fashion
,

each covering a distinct orthogonal domain of
relevance to neuroscience.

The modularity principle
facilitates flexibility in the use of NIFS
TD by
enabling users to accept only those domains they need

for annotating or for building additional ontologies
from NIFSTD core. The modules are standardized to
the same upper level ontology, the Basic Formal
Ontology (BFO), OBO Relations Ontology (OBO
-
R
O), and the Ontology of Phenotypic Qualities
(PATO). Within a module, we have established some
basic relationships among classes, e.g., a partonomy
of brain regions within NIF anatomy.
However, c
ross
-
module relationships are specified
in separate

bridge
files
through a set of object properties from standard
ontologies such as OBO
-
RO.
The use of separate

bridge files
maintains

the orthogonality of
each
module
.
The
N
IFSTD is expressed in OWL
–
DL

which
ensure
s

computational decidability and
automated reasoning
.

W
ithin the NIF,
NIFSTD is
served through an ontology management system
called OntoQuest which can read any OWL file and
uses a shredding scheme to populate an OW
L
-
compliant relational schema (currently in
PostgreSQL).
Ontoquest
supports operations for
navigating, path finding and hierarchy exploration as
well as term searching

in ontological graphs.

Current State

We have recently released version 1.0 of NIFSTD
bui
lding upon the 0.5 release described in Bug et al
1
.
NIF 0.5 was assembled in a relatively short period of
time from sources that contained overlapping domain
concepts. For example, the version 0.5 imported the
Subcellular Anatomy Ontology (SAO, Larson et.
al
2007) that mostly covered subcellular structures, but
also included neuronal cell types and molecules,
leading to duplicate and overlapping classes. It also
required a complicated redundant import tree leading
to frequent load error in Protégé
. For
NIFSTD 1.0,
the modularity principle has been vastly improved.
The classes from SAO have been placed under the
appropriate orthogonal modules and duplicate classes
have been removed. The dependencies between the
modules have been reduced to minimum; the mo
dules
can be loaded independently only under the imports
of upper level ontologies. The import hierarchy has
been simplified and re
-
engineered to eliminate the
redundancies.
http://purl.org/nif/ontology/n
if.owl
. A wiki
version of NIFSTD, the NeuroLex
(http://neurolex.org)

has been developed as the main entry point for
broader community to access, annotate and enhance
the NIFSTD contents in
an easier to use

and
collaborative
manner
.
The approv
ed contributions
from
NeuroLex
are implemented in the

NIFSTD
OWL files in a regular basis.

Conclusion

The standardized

approach of NIFSTD follows
the
best practices of ontology development that ensures
the core principles of ontologies being sharable,
reu
sable, and interoperable.

The NIFSTD continues
to evolve to incorporate new modules and contents as
well as implementing more detailed and useful cross
-
domain relations.

Acknowledgement

Supported by a contract from the NIH Neuroscience

Blueprint

HHSN27120
0800035C

via NIDA
.



Reference

1.

Bug W, Ascoli G, Grethe J, Gupta A, Fennema
-
Notestine C, Laird A, Larson S, Rubin D, Shepherd
G, Turner J, Martone M. The NIFSTD and BIRNLex
vocabularies: building comprehensive ontologies for
neuroscience. Neuroinform.2008;6:1
75
–
94