Open Biomedical Ontologies

Dear OBO collaborators,


We would like to submit our proteomics process ontology (ProPreO) for=20
inclusion in the OBO ontologies. A description of the ontology and how=20
to access it is given below.

Thank you for your consideration.

Will York - Complex Carbohydrate Research Center, University of Georgia
Amit Sheth - Large Scale Distributed Informations Systems Laboratory,=20
University of Georgia
Satya Sahoo - Large Scale Distributed Informations Systems Laboratory,=20
University of Georgia


  _ProPreO: comprehensive Proteomics data and process provenance Ontology=
_

Proteomics discipline and glycoproteomics in particular, are focused on=20
two core objectives:

a) Identification of a biomolecule =96 /what is it?/__

b) Quantification of the identified biomolecule =96 /How much of it is=20
there?/__

High-throughput experimental protocols for proteomics are rapidly=20
maturing and generating vast amounts of data. Similar to genomics, the=20
limiting step in this scenario will be the computational and related=20
analytical tools that are available to process this data and generate=20
useful information.

Many queries in proteomics involve comparison of the proteome in=20
different organisms, tissues, or cells in different developmental or=20
disease states. But, proteomics experimental protocols are characterized=20
by heterogeneity in sources of sample (biological organism), process to=20
generate data (separation techniques or mass spectrometric instruments),=20
parameters used in the process (instrumental parameters or separation=20
method parameters), and data formats used to store the data. Hence,=20
proteomics research requires not only finding specific data sets=20
obtained using the relevant biological sources, but also require one to=20
ensure that the data sets are comparable. For example, differences in=20
the sample preparation, data acquisition or data processing can=20
invalidate a comparison. Provenance, the information regarding the=20
ancestry of a dataset and the description of how the data is created,=20
transformed and processed, forms the foundation which allows for=20
multiple proteomics datasets to be compared in a relevant manner.=20
ProPreO, a proteomics process ontology, models not only /data=20
provenance/ but also /process provenance/ to enable consistent and=20
coherent comparison as well as analysis of proteomics datasets. Hence,=20
ProPreO is one of the first ontologies focused on capturing=20
comprehensive provenance information in proteomics and related fields.

As part of the bioinformatics research in the National Center for=20
Research Resources (NCRR) Integrated Technology Resource for Biomedical=20
Glycomics, ProPreO ontology is one of the two ontologies we have=20
developed. The other related ontology is GlycO, a domain ontology to=20
model the structure and function of glycans. These two ontologies enable=20
the creation of a computational framework for the annotation, retrieval=20
and analysis of high-throughput experimental proteomics and=20
glycoproteomics data, in order to facilitate the discovery of biological=20
knowledge that it embodies.

We adhered to four major criteria during the development of ProPreO:

a) *Logical rigor*: We are using ProPreO for annotation of experimental=20
proteomics data. Using this annotated experimental data; information=20
management application will not only be able to store, retrieve, and=20
integrate multiple datasets but also infer implicit knowledge that will=20
provide insight to proteomics researchers for hypothesis formulation and=20
validation. Hence, to allow computational tools to use ProPreO for=20
reasoning purposes, we ensured the absence of incorrectly determined=20
classes, incorrect or inappropriate naming schemes, and ill-defined=20
relationships between concepts in the ProPreO schema. ProPreO schema=20
includes 390 rigorously defined classes, 32 generic relations and 172=20
specific restrictions on the generic relations to correctly describe=20
each concept and its relation to other concepts.

b) *Amenability to existing bio-medical ontologies:* It is now well=20
understood and accepted that the life sciences domain requires multiple=20
ontologies to manage the inherent complexities of the domain. Hence, in=20
the scenario involving multiple but related ontologies, it is critical=20
that these ontologies can be used in an integrated manner by semantic=20
applications. We have followed the Basic Formal Ontology (BFO) (/Smith=20
B. et. al./ 2002) approach in class and relationship creation in=20
ProPreO. The three top-level classes of ProPreO are =93data=94 (datasets =
and=20
parameter data), (experimental) =93instrument=94, and (experimental) =93t=
ask=94.=20
Additionally, we created the relations in ProPreO by defining generic=20
and easily understandable relations at top-level classes. Using various=20
restrictions, we defined the application of the generic relations for=20
each class thereby effectively and efficiently modeling the=20
characteristics of each concept and its relation with other concepts=20
accurately. Currently, we are working on issues related to the=20
integration, mapping and alignment of ProPreO with ontologies listed in=20
the Open Biomedical Ontologies (OBO) repository.

c) *Use of OWL-DL language*: The Web Ontology Language (OWL) has three=20
flavors namely, OWL-Lite, OWL-DL and OWL-Full. As we planned ProPreO=20
ontology to be used by computational applications while being as=20
accurate as possible in expressing the inherent complexity of the=20
proteomics experimental domain, we chose OWL-DL as the language for=20
ProPreO. OWL-DL enables us to be expressive while ensuring acceptable=20
computational properties.

d) *Populated ontology*: We believe that an ontology schema is of=20
limited use without real world knowledge. We have populated ProPreO with=20
instances corresponding to concepts modeled as part of the ontology=20
schema. ProPreO has 3.1 million instances and 18.6 million triples. This=20
population of ProPreO with million of real world instances enables us=20
build computational tools that integrate the large volumes of=20
high-throughput experimental data within an overarching semantic=20
framework and reason over it for knowledge discovery.

These four criterions has enabled ProPreO to provide the formal semantic=20
foundation for modeling and incorporation of comprehensive provenance=20
information in wide ranging, high-throughput proteomics research.

ProPreO (version: 0.5) is available for download in the following=20
variants at http://lsdis.cs.uga.edu/projects/glycomics/propreo/:

a) *ProPreO schema*: The schema of the ontology featuring its 390=20
classes and attendant relations. This is an /*.owl/ file which is best=20
viewed using the Prot=E9g=E9 ontology development environment=20
(http://protege.stanford.edu/). The ProPreO schema file is relatively=20
small and hence may be used to gain an understanding of the structure of=20
the ontology and its applicability to various scenarios.

b) *ProPreO populated ontology*: This file includes the 3.1 million=20
instances and hence is a relatively large /*.owl/ file. ProPreO=20
currently is populated with instance related to human tryptic peptides,=20
their parent proteins and related enzyme entities. This populated=20
ProPreO ontology may be used as foundation for developing various=20
semantic applications that leverage its instances and its comprehensive=20
provenance framework.

For citation and further details on ProPreO: Satya S. Sahoo, Christopher=20
Thomas, Amit Sheth, William S. York, and Samir Tartir, =93Knowledge=20
Modeling and its application in Life Sciences: A Tale of two ontologies=20
<http://lsdis.cs.uga.edu/library/download/p1088-sahoo.pdf>,=94 the 15^th=20
World Wide Web (WWW, 2006) conference=20
<http://www2006.org/programme/item.php?id=3D1088http://www2006.org/progra=
mme/item.php?id=3D1088>,=20
Edinburgh, UK, May 2006.