=Paper=
{{Paper
|id=None
|storemode=property
|title=Profiling of Semantically Annotated Proteins
|pdfUrl=https://ceur-ws.org/Vol-952/paper_47.pdf
|volume=Vol-952
|dblpUrl=https://dblp.org/rec/conf/swat4ls/HollunderMAHK12
}}
==Profiling of Semantically Annotated Proteins==
https://ceur-ws.org/Vol-952/paper_47.pdf
Profiling of Semantically Annotated Proteins
Hollunder J1, Mironov V2*, Antezana E2, Hoehndorf R3, Kuiper M2
1
Department of Plant Systems Biology, Flanders Institute for Biotechnology and Department of
Plant Biotechnology and Bioinformatics,
Ghent University, Technologiepark 927, B-9052 Ghent Belgium
2
Department of Biology, Norwegian University of Science and Technology (NTNU), Høg-
skoleringen 5, N-7491 Trondheim, Norway
3
Department of Physiology, Development and Neuroscience, University of Cambridge, Down-
ing Street, Cambridge CB2 3EG, UK
*
Corresponding author {vladimir.n.mironov@gmail.com}
The first two authors contributed equally
Abstract. We have exploited semantic annotations of biological entities to de-
velop a novel approach to infer new knowledge. We demonstrate this in four
use cases based on the Gene Expression Ontology, an applied ontology that we
developed to serve the needs of researchers involved in the analysis of genes
and proteins implicated in transcriptional control of pathways/diseases. We
have found that semantic annotations associated with biological entities in vari-
ous commonly used data sources support the identification of related entities,
thereby emulating associations that can be inferred from sequence or other
structural similarities between these entities. We demonstrate how those seman-
tic annotations can be used to make inferences about the respective biological
entities.
Keywords: ontology, annotation, semantic similarity, gene expression, pattern
identification, hypothesis generation
1 Gene Expression Ontology
The Gene Expression Ontology (GeXO) [1] is an application ontology that integrates
fragments of GO and the Molecular Interaction ontology (MI) with data from GOA,
IntAct, KEGG, SwissProt, and NCBI Gene. It also includes information on predicted
orthology relations among the proteins. The knowledge in GeXO covers three biolog-
ical species: human, mouse, and rat. GeXO comprises 168,417 terms of which 39,680
correspond to proteins. In the present study we attempted to assess the global implicit
informational value contained in GeXO.
2 Semantic profiles of protein terms
Protein features were extracted from GeXO in the form 'predicate-object.' The fea-
tures form a matrix with 39680 rows corresponding to proteins and 132360 columns
to features. The types of features we used are summarized in Table 1.
Table 1. Ten sets of features of proteins in GeXO, with their subject name space, predicate and
object name space.
namespace predicate namespace count
NCBIGene codes_for UniProtKB 21184
GO contains UniProtKB 1025
IntAct has_agent UniProtKB 30538
UniProtKB has_function GO 2619
GO has_participant UniProtKB 339
UniProtKB has_source NCBITaxon 3
UniProtKB is_a SSB 5
UniProtKB member_of KEGG 2045
UniProtKB orthologous_to UniProtKB 14413
GeXO transformation_of UniProtKB 60189
This feature matrix was used to compute semantic similarity among all the proteins
in the data set on the basis of the Jaccard index weighted by the information content
as described in [2]. We evaluated the quality of the computed semantic similarities
using ROC analysis.
For classifying false and true positives we used KEGG clusters of orthologous pro-
teins as positive sets. The KEGG cluster annotations were removed from the data set
prior to the analysis. The results in Figure 1 demonstrate the very high predictive val-
ue of semantic annotations (the results with the full data are given just as a reference).
To exclude an impact of sequence information on the analysis, we removed orthology
information from the data set, which affected the results only slightly. We concluded
that semantic annotations are able to reveal protein similarity even in the absence of
sequence information.
3 Patterns in semantic profiles
To identify recurrent patterns in the data, we used the DASS tool [3], which finds
closed sets in the data. Closed sets have the property that there is no superset that oc-
curs more frequently in the data set. A data set consists of a set of sets of elements
(referred to as host sets).
Fig. 1. ROC analysis of semantically annotated proteins: KEGG clusters of proteins were
used for classifying true vs. false positives. Analysis was performed on (1) the full set of anno-
tations, AUC 0.92; (2) the same as (1) but without KEGG cluster annotations, AUC 0.89; (3)
the same as (2) but without orthology annotations, AUC 0.89.
Fig. 2. Distribution of closed sets: The distribution of closed sets by (1) the number of fea-
tures in the set, (2) the number of times the set occurs in the data, (3) the number of KEGG
protein clusters associated with the set. All the distributions had a long tail excluded from the
plot. Very similar trends were observed for closed sets with p-values below 0.05.
Figure 2 provides the distribution of closed sets according to the size (number of
elements), or frequency (number of occurrence in the data set). The vast majority of
closed sets fall within a narrow range of the size and frequency. The sets of low fre-
quency are likely to be highly predictive due to their specificity. To have a more pre-
cise view on the predictive value of the set we focused on KEGG clusters, which de-
fine functionally distinct protein types, associated with closed sets. Figure 2 gives the
distribution of closed sets according to the number of associated clusters. The highest
number of sets was found to be associated with a single KEGG cluster, thus confirm-
ing the high predictive value of the closed sets. It is worth noting the very high num-
ber of closed sets without any associated KEGG cluster. In combination with the re-
sults of the ROC analysis this suggests that the closed sets could be used to classify
the proteins associated to those closed sets.
4 Use cases
To demonstrate the high predictive value of the closed sets, we extracted a subset
containing 106 transcription factors (TFs) of 40 distinct types known or suspected to
be involved in the response to the hormon gastrin.
The 106 TFs were subjected to clustering with a number of approaches on the basis
of associated closed sets. The resulting clusters were used as templates for screening
the total GeXO data set to identify hypothetical TFs and target genes (TGs). For an
initial validation of the identified candidates, we downloaded additional information
from UniProtKB (lookup for the term: gastric in http://uniprot.org) and mapped it on
the identified candidates. We identified more than 1700 potential candidates including
more than 460 genes with transcriptional activity (non-deep analysis, automated
screening by extracting information from http://www.uniprot.org/uniprot/ with a cus-
tomized Python script). Furthermore, we identified 53 known TGs and 28 genes
linked to the term gastric, whereas 11 TGs and 7 gastric genes occur in more than two
clustering solutions and represent (partially) supported hypotheses. Thus, these results
show that we can use the closed sets concept for predicting TGs and regulators in-
volved in response to gastrin. Additionally, we identified more than 400 novel candi-
dates occurring in more than two clustering solutions. Evidently, not all of these can-
didates are directly involved in this response, but they represent a good basis for fur-
ther (more detailed) analyses as well as possible wet-lab experiments.
References
1. Venkatesan, A., Mironov, V., Kuiper, M.: Towards an integrated knowledge sys-
tem for capturing gene expression events. ICBO, 3rd International Conference on Bio-
medical Ontology (2012)
2. Hoehndorf, R., Schonfield, P. N., Gkoutos, G. V.: PhenomeNET: a whole-phenome
approach to disease gene discovery. Nucl. Acids Research, 39, e119 (2011)
3. Hollunder, J., Friedel, M., Beyer, A., Workman, C. T., Wilhelm, T., DASS: efficient
discovery and p-value calculation of substructures in unordered data. Bioinformatics,
23, 77 (2007)