The concept of a gene was established in the era of classical genetics and is now essential for life science for elucidating the molecular basis of the coding of genetic information necessary to realize the body of an organism and its biological functions. However, an ontology fully representing multiple aspects of a gene is still not available. In this study, we dissected the biological and ontological definitions of bearers of genetic information, including genes and alleles. Based on this analysis, we then proposed a basic way of modeling an ontology that represents the common definitions in classical and molecular genetics. This ontology is available at http://www.brc.riken.jp/lab/bpmp/ontology/ontology_gene.html.
1 INTRODUCTION “Gene” is one of the most fundamental concepts composing the basis of modern biomedical science, established in the 1900s. Today, genetic information is known to be coded molecularly using a sequence of four types of polynucleotide bases, providing blueprints for the development of an organism’s body and its biological functions.
Recently, gathering information using a computer has become more and more important as a component of research studies in life science. Ontology provides us with one of the most important means of processing varieties of data and representing knowledge models. Currently, several biomedical ontologies have been constructed with the aim of integrating a variety of information produced by different fields of biology. Therefore, development of a common model/ontology for genes is one of the key issues in bioinformatics studies.
However, the simple modeling of a gene and its allele into an ontology can raise some problems. For example, because a gene can be formed of two kinds of molecular entities, DNA or RNA, we cannot define a single class of gene under a hierarchical tree of molecular classifications.
Another problem with the classification arises with the hierarchy of “gene => allele => molecular instance”. This hierarchy corresponds to a hierarchy of “Gdf5 of mouse (gene) => Gdf5Rgsc451 (allele) => molecular instance of Gdf5Rgsc451”. As a result, we cannot define the gene or allele class in the ontology. In this case, each class of gene and allele seems to have another super-class, termed “gene” and “allele”, respectively.
In this study, we dissected the role of a gene and built an ontology that represents a consistent data model of the basic concepts of genetics, including genes, alleles, and nucleic acid molecules.
In classical genetics, a gene is defined as a “particle-like thing to define a genetic trait”. Genetic information conveyed by a gene was thought to have effects on specific body features or biological functions and to result in biological variation. To explain the levels of biological variation, a gene, or a factor that controls a specific trait in a species, and an allele, or a variant of a gene that produces phenotypic variations of a trait, must be clearly discriminated in genetics.
On the other hand, in molecular biology, genomic segments are classified according to their roles or features as genes, non-genics, promoters, exons, introns, and so on. A gene is defined as a segment that codes one or multiple products (functional polypeptide or RNA).
The role of “bearer of genetic information” is essential for a gene. It is notable that the definition of a gene was made prior to the discovery of its material basis. To put it extremely, any molecular entity that bears genetic information in a “unit of” functional role could be termed a gene.
The role of a gene can be classified into two categories. One is the transmission of genetic information from parents to their offspring. This is driven by a series of biological processes resulting in the self-replication of nucleic acid chains. Additionally, through the effects of mutation (the alteration of genetic information), the self-replication processes contribute to genetic variation. In this sense, the hierarchy of gene to allele corresponds to the levels of genetic variation: namely, inter- and intra-species variations.
Another role of genes is to encode the design plan of individual organisms, namely, the coding of gene products. This function is accomplished by a sequence of biological processes including gene transcription and the translation of mRNA to a polypeptide. This role is equivalent to the regional classification of genomic segments in molecular biology.
In summary, the classifications of genomic segments can be categorized into two dimensions. One is the direction to increase genetic diversification based on self-replication; the other is a region-based classification of genomic segments. These classifications are distinguished by their fundamental roles in providing different contents of genetic information through different biological processes. For instance, a gene, or a species-specific variation in a genome segment that codes a gene product, can be represented as the point of intersection of these two systems (Fig. 1). the content is a sequence of polypeptides or functional RNA as a design plan.
The above-mentioned problem, i.e., that the concepts of gene and allele seem to contain “another super-class” can be solved if they are regarded as top-level concepts for roles in a specific context. 3
Next, we further dissected the detailed roles of genomic segments. 3.1
The two kinds of genetic information carried by a gene are commonly represented by the specific form, or the functional structure of the polynucleotide molecule. This structure can be compared to an artificial information bearer in which symbol sequences (e.g., text) are conveyed on an information media (e.g., paper).
In the description of information using text, a “symbol” is the most fundamental unit. In the English language, the letter “G” in the alphabet represents an abstract meaning (informational object) of the symbol “G” with the form of a line image in the shape of “G”. The coding of genetic information by a molecule has the same structure. The genetic information of “G” is representation of the abstract symbol of “G of the genetic code,” the molecular entity of which corresponds to a guanine nucleotide. (Mizoguchi 2004)
Using a sequence of such molecular symbols, representing the polynucleotide groups, the genetic information for an organism can be conveyed. This is analogous to a paper document conveying a representation of a story or a specification of something using a symbol sequence.
On the other hand, the content represented by the two kinds of genetic information is clearly different. With selfreplication, the content represented is the design plan for the nucleotide chain as itself. In the coding of a gene product, 3.2
The roles of self-replication and diversification can be classified as follows: (1) Level general to all organisms:
A molecular entity, such as polynucleotide group, is termed as a genomic segment and plays a role in bearing information for self-replication under the general context of the organism. (2) Level reflecting the identity of a biological species: The genomic segment plays a specified role in bearing species-specific genetic information under the context of the population of the species (or the gene pool). At this level, the genomic segment may be often termed as a genetic marker. In this paper, we call it an “s-segment”. (3) Level reflecting inter-species variation:
Under the context of species populations, another role exists: encoding inter-species variations. In this context, the s-segment is referred to as an allele. 3.3
On the other hand, under the context of an organism, a genomic segment, which codes a polypeptide or a functional RNA, is generally termed as a “gene”. This is not a speciesspecific gene, as described above. In this paper, we refer to this context as a “gene type” for discriminatory purposes. 4
Based on these analyses, we worked out an ontological
model to describe the concept of gene. We referred to Yet
Another More-Advanced Top-level Ontology (YAMATO),
which represents a basic framework of roles and
representation
With the Hozo ontology editor, concepts are
fundamentally categorized into basic concepts, which can be defined
context-independently, and role concepts to be played by an
entity, which are dependent on a specific context. An entity
playing a role concept is termed as a “role-holder”. For
example, a teacher, a person who plays the role of a teacher
role in a school is a role-holder. The player is selected from
the basic concepts or role-holders
Definitions of key concepts in natural language are shown below and their schematic representation in Hozo is depicted in Figure 2 Fig. 2. Modeling of gene, allele and genetic information according to the Hozo ontology editor. A: Class tree of basic concepts.
B: Composition of organism and population of a species. (i)
Genomic segment =def A role holder played by a molecular entity or polynucleotide group, under the context of organism. (ii) Gene type =def A role holder played by a genomic segment, which has both of the two kinds of genetic information, i.e., information for selfreplication and coding of gene product, and inherited genomic segment under the organism context. (iii) S-segment =def A role holder played by a genomic segment under the context of a population of a species, a Mendelian population. (iv) Gene =def A role holder played by a gene type under the context of a population of a species, which is defined by the specialization of the s-segment. context
Legend Role and role –holder ! (gray text indicates a role-holder defined in another context) cardinality Kind of a slot! p/o: part_of! a/o: attribute_of
Player! “(RH)”: player is a
role holder.! “#”: equivalent of punning in OWL2.! (v) Mutation =def A role holder played by a genomic segment, which is the terminal state in the process under the context of the mutation process. (vi) Variant of s-segment =def A role holder played by a s-segment under the context of a mutation process in gene pool. (vii) Allele =def A role holder played by a variant of an ssegment under the context of a gene pool of a population of a species. (viii)Genetic information entity =def A YAMATO: representation, composed of a molecular sequence of a representation form and YAMATO:specification as a content. (ix) Molecular sequence =def A YAMATO:symbol sequence composed of a molecular symbol. (x) Molecular symbol =def A YAMATO:symbol, which is a YAMATO:content. (xi) Representation of molecular symbol =def A YAMATO:representation, which is composed of a
radical group as a form and a molecular symbol as a content. (xii) Major allele =def A role holder played by an allele under the context of the population of a species. (xiii)Loss of function allele =def A role holder played by a gene allele under the context of an organism.
Hozo’s reasoner dynamically generates a classification of role-holders and their players by referring to interrelationships among slots and class restrictions (Fig. 3). Because the classification of genomic segment is dependent on multiple contexts, the inferred hierarchy represents multiple inheritances of role-holders. Here, the hierarchy demonstrates the practical orders of classification as “polynucleotide group => genomic segment => s-segment => allele => major allele”, “s-segment => gene => gene allele => loss of function allele” with the help of roles.
Fig. 3. An partial “is_a” hierarchy of classes (represented as “W”) and role-holders (RH) generated by the Hozo-reasoner. One roleholder appears multiple times in the hierarchy because of multiple inheritance. 5 5.1
The meaning of the term “gene” seems to have changed
historically. However, biologists have no difficulty thinking
in Mendelian terms when applying traditional genetic
techniques and are able to switch seamlessly to molecular
conceptions of the gene
First, we defined the body of information carried by a
gene using YAMATO, which is a top-level ontology based
on the traditional and commonplace Aristotelian ontology,
such as the Basic Formal Ontology (BFO) and the
Descriptive Ontology for Linguistic and Cognitive Engineering
(DOLCE). The ontology of representation in YAMATO
shows the definition of the general structure for variously
styled “content-bearing informational entities”, or
representations, composed of a representation form and a content.
A physical thing conveying a representation (i.e., books or
electric document files) is defined as a representing thing.
The ontology of representation provides a detailed theory
for the classification and instantiation of these abstract and
physical entities
A genetic information entity, which is transmitted from a parent to its progeny (or from a genomic segment to a replicated segment) was modeled as a specific combination of a symbol-sequence pattern of nucleotides and informational content to specify a one-dimensional structure of itself or a non-self molecule (gene product). This model revealed the separation of “information for self-replication”, which all genomic segments have, and “coding of gene products”, which only genes have. The modeling of a genomic segment analogous to a representing thing enables multiple copies of a specific gene to share the same genetic information.
Another achievement is the modeling of a multiplex
classification of genomic segments that play different roles in
different contexts. By dissecting the biological role of
genetic information entities and their contexts (i.e., organisms
or biological populations), we demonstrated the systematic
organization of concepts derived from genomic segment
without multiple inheritance. This achievement was
obtained thanks to the Hozo tool, in which the role theory of
YAMATO is well embodied as operations to edit ontologies
Various broad concepts of “gene” have been proposed
among existing biomedical ontologies. For example, the
Sequence Ontology (SO) classifies a gene as a “region”,
which is a “sequence feature” composed of bases and a
polypeptide region composed of amino acids
Hoehndorf et al. proposed a system of axioms for SO’s
top-level categories based on three primitive terms
(Molecular sequence, Syntactic sequence, and Abstract sequence) to
specify the meaning of sequence-related terms used in the
biological domain
Recently, the SO was revised to define the mereological,
spatial, and temporal aspects of a biological sequence
Thus, we think our effort is complementary to sequenceoriented approaches with SO, providing a biological roleoriented solution for the problem of the multiplex classification of genomic segments in genetics.
Transferring these merits into the OBO Foundry ontologies (Smith et al., 2009) is one of the major future issues for this study. The current version of our ontology is compatible with several domain ontologies. For example, the top level of molecular entity is fully compatible with the Chemical Entities of Biological Interest (ChEBI). However, open issues to build interoperability with the SO and Information Artifact Ontology (IAO) remain to be resolved.
In biological databases, genes and alleles are often instantiated at the level of an individual gene (e.g., Gdf5 and Gdf5Rgsc451 allele). This issue is not well addressed in this paper. From an ontological viewpoint, the following possibilities can be suggested: (a) They are instantiated from the view of a meta-model, as a rough equivalent of the “name” of class generated by the punning operation in OWL2. (b) They are instances of the specifications of genes and alleles. They are described in Fig. 2A as “specification of genomic segment”, which is a content of information for self-replication.
We prefer (b) to (a). The specification, as well as the
“plan”, which is a specification of a sequence of actions
mentioned elsewhere
In the SO, a gene is classified as a biological region defined as a “generically dependent continuant.” This means that the SO also represents individual genes as instances of abstract things. To establish inter-operability with the SO and the IAO, further consideration of specification and its realization of designed entities is needed.
This work was supported by the Management Expenses Grant for RIKEN BioResource Center, MEXT.