=Paper=
{{Paper
|id=Vol-2469/ERForum3
|storemode=property
|title=Integration and Analysis of Clinical and Genomic Data of Neuroblastoma applying Conceptual Modeling
|pdfUrl=https://ceur-ws.org/Vol-2469/ERForum3.pdf
|volume=Vol-2469
|authors=Sipan Arevshatyan,José Fabián Reyes Román,Verónica Burriel,Adela Cañete,Victoria Castel,Óscar Pastor
|dblpUrl=https://dblp.org/rec/conf/er/ArevshatyanRBCC19
}}
==Integration and Analysis of Clinical and Genomic Data of Neuroblastoma applying Conceptual Modeling==
https://ceur-ws.org/Vol-2469/ERForum3.pdf
Integration and Analysis of Clinical and Genomic Data
of Neuroblastoma applying Conceptual Modeling
Sipan Arevshatyan1[0000-0001-8718-2211], José Fabián Reyes Román1[0000-0002-9598-1301],
Verónica Burriel2, Adela Cañete3, Victoria Castel3, and Óscar Pastor1[0000-0002-1320-
8471]
1
PROS Research Center, Universitat Politècnica de València,
Camino Vera s/n. 46022, Valencia, Spain
2
Department of Information and Computing Sciences, Utrecht University,
Domplein 29, 3512 JE, Utrecht, The Netherlands
3
Pediatric Oncology Unit of Hospital Universitari i Politècnic La Fe, Valencia, Spain
siar5@doctor.upv.es, {jreyes|opastor}@pros.upv.es, v.burriel@uu.nl,
{canyete_ade|castel_vic}@gva.es
Abstract. Data management and analysis for risk assessment of rare and
complex diseases such as Neuroblastoma require efficient management of
multidisciplinary data. Recent advances in genomic testing are revealing
new publicly available data whose storage and analysis with clinical and
genomic data is becoming a big challenge. The use of Conceptual Modeling
(CM) techniques helps to define and structure the Neuroblastoma domain,
which serves as a basis to determine the information required for
diagnosing the disease. It is important to highlight that a Genomic
Information System (GeIS) based on a conceptual model allows improving
the adaptation of new requirements of the domain, and greatly simplifies
the integration and management of heterogeneous and homogeneous data.
The main objectives of this work are: i) to present a Conceptual Model of
Neuroblastoma (CMN), which defines all elements involved in the clinical
and genomic domain. ii) to apply the SILE method, in order to obtain all
(clinically) relevant variations associated with Neuroblastoma from
genomic data sources. The developed GeIS is intended to make the correct
exploitation of the validated data set to provide an early and efficient risk
assessment for patients with Neuroblastoma.
Keywords: Neuroblastoma, CM, GeIS, CMN, SILE Method, PM
1 Introduction
Since the first complete human genome was published in 2003, there has been an
astonishing progress regarding speed and cost. This task took 13 years and an
approximate expense of $3 billion [1]. With the barrier of $1000 per genome
Copyright © 2019 for this paper by its authors. Use permitted under Creative Commons
License Attribution 4.0 International (CC BY 4.0).
29
already broken, data acquirement is no longer a challenge for Next-Generation
Sequencing (NGS) technologies. The storing, managing and making sense of the
data has become the main issue since it requires a broad spectrum of specialists
including IT, computational biologists, genetic counsellors or pathologists [2].
Once the raw sequence data is obtained, it is aligned to the reference sequence.
According to a recent study performed on over 2,500 individuals, everyone differs
at 4.1 to 5 million sites from that reference genome [3]. The combination of these
genetic variations (or variants), known as “genotype”, together with environmental
factors, determines its host physical traits (known as “phenotype”). The main goal
of genomic medicine is to understand the genotype-phenotype relationships.
Considering that the reference sequence does not codify for any serious condition
(which is not clear is completely true [4]), the genetic driver of a disease can be
found among the group of differences between a patient’s genome and the
reference one. Therefore, Whole Genome Sequencing also known as WGS, may
assist in differential diagnosis [5]. Intensive research has aimed to reveal
associations between genetic variations and diseases through Genome-Wide
Association Studies (GWAS) [6]. These associations are especially interesting in
rare diseases in which hereditary studies are hindered by the high lethality in early
childhood. Representing more than 7% of all pediatric cancers and causing around
15% oncology deaths, rarity and early lethality unfortunately fit with
Neuroblastoma’s definition.
Neuroblastoma is the most common extra-cranial solid tumor in childhood. It
usually appears within the abdomen, neck, chest or pelvis, and its symptoms
depend on the tumor’s location [7]. The variability of clinical presentation and
likelihood of cure are examples of one of the most remarkable hallmarks of
Neuroblastoma: its clinical heterogeneity. Although some tumors undergo
spontaneous regression, others progress despite aggressive therapy [8]. This
disparate clinical behavior has been shown to be related to biological factors such
as age at diagnosis, tumor histology or genetic aberrations. For many years,
research groups have used different factors in order to stratify patients for risk-
based clinical trials. These differences prevented results from being compared. To
ease that analysis, several international efforts have resulted in standards for
patient classification and staging, such as the International Neuroblastoma
Staging System (INSS) [9], International Neuroblastoma Risk Group Staging
System (INRGSS) [10] and International Neuroblastoma Risk Group (INRG)
classification System [11].
The INSS consists of six stages (1, 2a, 2b, 3, 4 and 4S) according to the degree
of surgical resection. Patients are assigned a stage post-surgery. Since that fact
makes INSS not suitable for pre-treatment risk-based classification, a new
INRGSS was defined. In contrast to INSS, INRGSS classifies patients in four
stages (L1, L2, M and MS) [9] depending on clinical criteria and image-defined
risk factors. The INRGSS together with other genetic features as tumor ploidy,
chromosome 11q aberration, and “v-myc avian myelocytomatosis viral oncogene
neuroblastoma derived homolog” (MYCN) amplification are considered when
assigning patients to its corresponding INRG Classification System risk group.
30
The variations in the number and size of chromosomes as well as in the number
of MYCN repetitions are not the only genetic aberrations associated with
Neuroblastoma development and outcome. Also, smaller variations like
substitutions in genes such as “anaplastic lymphoma kinase” (ALK) [12],
“pairedlike homeobox 2b” (PHOX2B) [13] and “kinesin family member 1B”
(KIF1B) [14] are shown to be related to the hereditary disease. Similarly,
variations in regions between genes have been found to affect Neuroblastoma
progression. Moreover, they have a cumulative effect allowing to assign patients
to different risk-based groups, depending on the type of variations they carry.
These variations are normally presented in scientific publications. Although many
of them remain locked in unstructured sources, some researchers have focused on
targeting genetic variations in order to transfer them into curated clinical databases
[15]. Some of these text mining efforts have successfully located genetic
variations within papers and loaded into databases (e.g., ClinVar 1) with clinically
relevant information. Scientists also have the option to directly load the variations
they find in these databases or many others with different backgrounds.
ClinVar and Human Gene Mutation Database (HGMD, www.hgmd.cf.ac.uk/)
are examples of clinical repositories gathering information about the effect that
genetic variations have in humans, independently of the population or possible
treatments for patients who carry them. Drugs targeting specific variations can be
obtained by browsing the Comparative Toxic genomics Database (CTD,
ctdbase.org/). On the other hand, the variation frequencies in different populations
can be found in dbSNP 2, together with a small portion of the sequence containing
the variation. The complete sequence of genes is stored in databases such as
Nucleotide 3 or Ensembl (https://www.ensembl.org/). These archives sometimes
contain many sequences for a single gene. To facilitate study comparisons a
Reference Sequence (RefSeq, https://www.ncbi.nlm.nih.gov/refseq/) database was
created which stores a unique sequence per gene. This is only a small selection of
databases showing the existing heterogeneity among the so-called “genomic
chaos”. As seen above, valuable information in the decision-making process
following a genetic test is dispersed over several sources. The same variation can
be stored in different databases, which sometimes offers the same information.
That fact leads in the best of the cases to redundancy. The data does not always
match though; incongruities arise in these cases which can compromise the
patient’s response to the prescribed treatment [16].
The main objective of the research is to create a conceptual model of
Neuroblastoma which will integrate clinical and genomic data. Applying CM on
the genomic domain is highly helpful since it allows to define it accurately, hence
creating a robust structure from which data can be efficiently managed. According
to Olivé [17] the activity which ultimately defines the general knowledge on
which an Information System (IS) works is what we know as CM. Since an IS
1
https://www.ncbi.nlm.nih.gov/clinvar/
2
https://www.ncbi.nlm.nih.gov/snp/
3
https://www.ncbi.nlm.nih.gov/nucleotide/
31
built upon a non-described knowledge is considered to be unpredictable, the
obtaining of such a description is the main aim of CM. Once a genomic
conceptual model is created, it is possible to design Genomic Information Systems
(GeIS) [18], which play a key role in the biomedical domain. Taking into
consideration the heterogeneity, dispersion and redundancy which characterize the
genomic domain, the use of conceptual models structuring its basic features is of
great use in the way towards Precision Medicine (PM) [19]. Personalizing the
treatment, depending on the patient’s genome and environment, will only be
possible if there is a clear definition of the domain and a robust GeIS able to
integrate and manage multidisciplinary data.
To achieve our goal following this research line, in Section 2 we firstly give a
brief view of the current state of CM in the biological domain, as well as the
contemporary condition of Neuroblastoma research. A Conceptual Model of
Neuroblastoma (CMN) and an online tool built on it are introduced in Section 3.
Section 4 shows the process of searching and identifying relevant genetic
variations affecting Neuroblastoma development, and how to load them into our
database, which will allow data exploitation. Finally, the conclusions and future
works are presented in Section 5.
2 State of the Art
We study the state of the art in two different fields. Firstly, we describe how CM
is applied to the genomic domain in order to provide deep knowledge to generate
the CMN, and design and develop a system adapted to the needs of Pediatric
Oncology department of the Hospital Universitari i Politècnic La Fe (HUP/IIS La
Fe). Secondly, we study the Neuroblastoma’s domain and the available databases
in use by the hospital. This was an important step of the research in order to figure
it out how to create a conceptual model adapted to the needs of the department and
additionally, implement advanced data management techniques in order to
improve decision making processes.
2.1 Conceptual Modeling in Biology
Despite the large amount of publicly available genomic data repositories, it is not
usual to find stable conceptual models underlying them. Since most of the
accessible archives require storage improvements, CM in the biological domain
has been used over the years. Initial proposals in the field were introduced by
Paton [20], who presented several models defining cellular genetic components,
products and their interactions. The work developed by Ram [21] was focused on
the protein context and proved conceptual modeling usefulness in searching and
comparing large volumes of 3D structural data. CM is not only an approach to
32
describe and represent a specific domain but also aids in software production [34].
The models have been already used in bioinformatics. By applying this method
Garwood et al. [22] created user interfaces for querying biological data
repositories. Considering data archives designed so that variations associated to
diseases are stored, a proper data management may assist in generating a diagnosis
and recommending personalized treatments. Keeping in mind the principles of
PM, this research group 4 developed GeIS in order to manage some diseases like
usher syndrome [23], breast cancer [24], alcohol sensitivity [18], among others;
based on its respective domain conceptualization. More generally, with the aim of
creating a standardized vocabulary and unequivocally defining the genomic
domain, the Conceptual Model of the Human Genome (CMHG) [25-26] was
created representing all relevant information in the genomic domain as a whole.
This model led to the development of a Human Genome Database (HGDB) [26],
which can be exploited by using “GenesLove.Me” (see more information in [27]),
an internally matured tool believed to be of great use in genetic diagnostics.
The use of CM is gaining momentum as a software development approach in
the medical domain since it greatly improves geneticists’, lab scientists’ and
physicians’ work. Through the creation of GeIS, diagnostic tools can be developed
to allow the creation of accurate diagnosis once the right information is identified
and collected.
2.2 Neuroblastoma
Given the relative rarity of Neuroblastoma, small volumes of data are readily
available. Its correct management is then paramount in order to properly
understand the molecular basis of the disease. To make Neuroblastoma data
widely available and thus facilitate international, multi-disciplinary research, an
INRG database [28] was created with the specific aims of enabling complex
queries and linking their results to biological databanks and publicly available
datasets. In 2013, after two years of work and when the most recent progress
report was published, this database contained a total of thirty-four biological
metrics on approximately 11,000 patients, including tumor stage, MYCN status,
tissue availability, race, ethnicity or site of relapse [28]. Since several common
SNP (Single Nucleotide Polymorphism) alleles have been demonstrated to be
involved in Neuroblastoma tumorigenicity [29-30] genomic data stored in the
database of Genotypes and Phenotypes (dbGaP,
https://www.ncbi.nlm.nih.gov/gap/) the NGS and TARGET 5 data were planned to
be added to INRG database.
The INRG international collaboration intends to compensate the relative lack of
research focused on Neuroblastoma. This shortage was noticed after exhaustive
4
Research Center on Software Production Methods (PROS), http://www.pros.webs.upv.es
5
Therapeutically Applicable Research to Generate Effective Treatments,
https://ocg.cancer.gov/programs/target
33
searches for scientific information prior to the project in collaboration with the
Pediatric Oncology Unit of HUP/IIS La Fe (http://www.hospital-lafe.com/) in
Valencia (Spain). Nowadays, the information related to Neuroblastoma available
in databases is mainly clinical. The genetic tests performed on tumor cells in order
to diagnose the disease analyses only a few variations for which there is a known
treatment. This catalogue of druggable genetic variations is growing as new
therapeutic targets are discovered and innovative genetic tests are developed. The
information storage and analysis are becoming a challenge since it needs to meet
the new requirements. This work is an effort to unify, in a single database, clinical
and genetic information with the aim of assessing Neuroblastoma patient’s risk.
Building a GeIS on a database gathering interdisciplinary information will allow
unveiling genetic patterns within Neuroblastoma cohorts. Together with recent
technological advances in molecular testing, and also in the computer sciences
(e.g., applying Data Sciences, Artificial Intelligence, and Big Data techniques),
could help and improve our understanding about genetic basis of the disease, and
would ultimately lead to an efficient diagnosis. A precise diagnosis will in turn
assist in the process of developing more effective targeted therapies.
3 Domain Definition
The CM and ontological characterization of basic features making up a specific
environment assure consistency, correction and an efficient exploitation of its
specialized datasets. This is especially true for those repositories designed to store
complex data coming from heterogeneous sources.
Neuroblastoma represents an excellent example of a disease whose treatment
choice, application and traceability require advanced technologies able to gather
and manage data coming from many different fields, ranging from histological
tests to nuclear medicine. The users of these technologies fit quite different
profiles encompassing pathologists, geneticists, lab assistants or
bioinformaticians. A precise definition of the domain is highly advantageous in
order to build applications able to store valuable information for each of them
separately and, most important, to obtain worthwhile conclusions from its
aggregated analysis. For that reason, the CMN (Fig. 1) was created in
collaboration with experts of the Clinical and Translational Cancer Research
Group (GICT-Cancer) from HUP/IIS La Fe. This CMN is the result of iterative
meetings and discussions with the experts of the Research Group. Through the
discussions and analyzes carried out, the CMN was generated, which facilitates
the understanding of the domain of Pediatric Oncology.
34
Fig. 1. The Conceptual Model of Neuroblastoma (CMN)
35
Next, the description of the CMN is presented, which contains all the elements
of the previously defined domain. In the model the “Patient” class is characterized
by attributes such as “name”, “NRG”, “date of birth” and “date of diagnosis”,
“age” or “weight” and could be considered the center of the model. This class
represents information about individuals attending a specific healthcare institution,
represented in the model by the “Hospital” class. To start with, the patient is
assigned a “Status” class, which is divided into “Deceased”, “Relapse” and
“Disease Free” classes respectively. Given the Neuroblastoma heterogeneity,
different diagnostic tests are performed depending on biological and genetic
factors. Several standard procedures regulating tests were developed, which are
represented by “Protocol” class. The disease development stages are also
dependent on the protocol the patient is assigned to. They are represented by
“Phase” class, whose intersection with patients in a specific phase results in the
“Episode” class. The former class refers to tests performed, diagnosis arising from
the tests and the treatment applied to the patient. One or several tests can be
carried out during a single episode of the disease so its progression can be tracked.
Since several tests coexist, with both common and unique attributes, “Test” class
is divided into “Histological”, “Radiological”, “Nuclear Medicine”, “Biological”
and “Minimal Residual Disease”. The “Histological” class has two classes which
inherit from it. These classes are “Histological Marrow” class and “Histological
Tumor” class, depending on the analyzed tissue.
The tumor tissue can be used for genetic analysis, represented in the model by
“Genetic Test” class. The results from these tests are referred by “Genetic Result”.
All the previous analysis is performed in order to characterize the “Tumor” class,
whose size, location and possible dissemination (i.e., “Metastasis”) are recorded.
The tumor features together with other significant test results assist in generating
the “Diagnosis”, which ultimately determines the prescribed “Treatment”. Since
there are several treatment options, this entity is specialized in the classes
“Chemotherapy”, “Surgery”, “Transplant Aut”, “Radiotherapy”, “MIBG” and
“Maintenance”. Some of these treatments involve toxic substances whose effects
need to be tracked. For that reason, the “Toxicity” class was created which refers
to the toxicity level. The “Genetic Results” class has a relationship with the
“Variation” which represents every genomic variation in that patient. If those
variations have been reported in scientific literature and are stored in a public
genomic database, a relation will be defined with “External DB” class.
A “Variation” is related with the "Chromosome" in which it is located, and also
it could be located inside a “Gene”. The relationship between the “Gene” and
“Chromosome” classes shows the chromosome where the gene is located. The
“Variation” class is divided into “Precise” and “Imprecise” classes to specify if
the position of the mutation is known. In particular “Deletion”, “Insertion” and
“Inversion” classes show the position of the deletion, inversion and insertion of
nucleotide sequence in the DNA sequence of the chromosome while the “Indel”
class represents consistent variations in insertions and deletions at the same time
in the DNA sequence of the chromosome. In contrast to “Precise” class the
“Imprecise” does not contain the position as it is unknown. This class could be
36
specialized in 2 subclasses to define two kinds of imprecise variations, such as
changes in the number of copies of certain DNA fragments (also called Copy
Number Variation) - “CNV” class – and big changes affecting the structure of the
chromosome – “Structural” class. Besides that, the "Amplification" class
represents an increase in the expression of a specific "Gene", which is represented
by a relationship between these two classes.
This model allows non-experts to understand the environment and easily use
tools built on it. In particular, it was used as a basis to design an online tool meant
for the management of clinical and genomic data collected from patients with
Neuroblastoma. Furthermore, its structure does not limit its extension but leaves it
open to evolution by adding possible diagnostic tests, drug targets or innovative
treatments.
Several technologies were used in the development process of the web-based
tool (prototype) aimed for Neuroblastoma clinical and genomic data management.
Among these technologies we can highlight, Java (https://www.java.com/en/)
using Java Persistence API 6 for the database loading process. JavaScript
(https://www.javascript.com/) was used in the development of the tool itself.
Specifically, the backend was designed with Node.js (https://nodejs.org/en/), Jade
(https://jade.tilab.com/) while Bootstrap (https://getbootstrap.com/), jQuery
(https://jquery.com/) were used for frontend development.
An analysis of the available data storage technologies was carried out in order
to determine the most appropriate one based on the type, quantity and subsequent
use of the data to be stored. In this stage, the relational SQL technology was
selected. However, in the future in case of the expansion of the data and according
to the needs of the doctors the data could be migrated to NoSQL technologies (e.g.
MongoDB or Neo4j among others).
This prototype is based on the conceptual model mentioned above and currently
is being tested by the experts of the GICT-Cancer from HUP/IIS La Fe. In this
stage, the prototype is being checked for errors, bugs with the aim of improving
the prototype according to the needs of the doctors.
4 Applying the SILE Method to Neuroblastoma
In order to assess Neuroblastoma risk from NGS data, the SILE (Search-
Identification-Load-Exploitation) method was followed [31]. Its main goal is to
systematize the search and identification of genomic information to be loaded,
analyzed and exploited by a GeIS based on the CMHG [26]. A summary of the
activities taking place at each level of the method is defined in Table 1.
As previously stated, Neuroblastoma has been associated to/with a wide variety
of genetic variations by means of different study methods. Since risk assessment
in Neuroblastoma does not only rely on biological issues but also on clinical and
6
https://www.oracle.com/technetwork/java/javaee/tech/persistence-jsp-140049.html
37
genetic features, the SILE method [35] represents a good choice to unify and
efficiently use all the available information.
Table 1. Description of each level of the SILE method [31]
Level Description
(S) Search Determination of the information context, required to solve a
concrete need, as well as the selection of data source from which to
extract information
(I) Identification Determination of a reliable and relevant dataset to be used to
populate a database which structure is delimited by the CSHG
(L) Load Population of the database with the data identified in the previous
level
(E) Exploitation Extraction of knowledge form the database by using tools to analyse
and interpret genomic data
4.1 Results obtained from the SILE method
• Search. An exhaustive research on existing integrative databases was
carried out in order to study other group’s strategies (e.g., where to obtain
variations from and how to properly annotate them). DisGeNET
(www.disgenet.org/) was the major finding of this intensive search. Its
creators define it as a discovery platform which integrates information on
gene-disease associations (GDAs) from public data archives and the
literature. Interestingly, DisGeNET classifies data as “Curated”,
“Predicted” and “All” depending on the original source it comes from.
We also have used ClinVar and dbGaP genomic repositories to find that
variations [32].
• Identification. In this stage, the data which was collected from different
data sources presented in the first stage of the SILE method, was
analyzed in order to remove possible redundancies and other quality
issues. The main aim of this step is to prepare the data for loading into
the database. At the beginning there were 996 variations collected from
different databases. After the Identification stage the complete validated
dataset consisted of 375 clinically relevant variations annotated in order
to allow a GeIS to efficiently locate and display the data [32-33].
Furthermore, search and identification processes allowed us to spot
several data quality issues such as redundancy or inconsistency which
must be solved before loading process.
• Load. A Database of Neuroblastoma (DBN) was developed in our
research group in order to efficiently store the clinical and genomic data.
It was based on the CMN mentioned above (Fig. 1). Both the DBN and
CMN were developed with the aim of defining all features related to
research, diagnose and treatment of Neuroblastoma and creating a strong
structure on which a GeIS has been developed. Based on the GeIS
previous studies, Neuroblastoma data was loaded into the DBN so that it
38
could be exploited using an online tool, the GeIS, which is presented in
the “Exploitation” section.
• Exploitation. The exploitation will be carried out through the prototype
(Fig. 2) developed for the GICT-Cancer from HUP/IIS La Fe. After the
testing of the prototype by the experts it will be available for clinicians to
use. The exploitation of the prototype will ultimately lead to a genetic
risk assessment based on validated evidences available on public
resources.
Fig. 2. The prototype developed to integrate and analyse clinical and genomic data
The design and the development of the prototype are based on emphasizing the
efficiency, usability, and security aspects (e.g., the prototype will take into account
all the current regulations of Data Protection Law 7). A Web-based approach is the
selected strategy to design the application as it has some well-reported advantages
for the considered working domain. The web-based software gives the user
flexible access to the all necessary tools or required data.
Although Model-Driven Engineering (MDE) tools can generate software code
automatically by using the model as an artifact we used model-based archetypes
[36] to specify user interaction strategies. Archetypes represent different views
offered to the users. In order to design the interface of the application, some
archetypes have been created based on the CM.
The prototype stores and manages patient’s demographic information, episode
description, complementary information, treatments, pathological and genomic
7
https://www.boe.es/buscar/doc.php?id=BOE-A-2018-16673
39
information. Additionally, the prototype has an analysis section where the
clinicians can define certain queries and get information from the GeIS.
5 Conclusions and Further Work
The main result of this work was the definition of the CMN, with the aim of
making the disease widely understandable for any personal profile involved in its
diagnosis, treatment and traceability process. The thorough analysis of the
domain allowed us to spot which information was valuable for the disease
diagnosis, its availability and current challenges regarding its management. In
order to overcome the contemporary difficulties, a GeIS based on the CMN was
built which was capable of efficiently managing Neuroblastoma-related clinical
and genetic data. The implementation of the SILE method allowed to define a
validated dataset of variations associated with Neuroblastoma, this group of
variations was selectively loaded into our DBN, and finally the exploitation was
carried out through the developed prototype, which is currently being tested.
The construction of the CMN has been proved to be highly convenient in
assisting the development of the GeIS since it provides a robust structure
supporting a correct data organization. Besides that, the model is also flexible in
the way it could easily adapt to new diagnostic tests or treatment strategies, hence
growing as Neuroblastoma related knowledge does. Although only curated
databases have been browsed in order to obtain the validated set of variations,
other databases might be useful for doctors in their decision-making process.
Following this research line, the CMN will be extended by characterizing new
"concepts" or "knowledge" involved in the domain and adapting the developed
prototype according to the new requirements. This study can serve as a basis or
prototype for the risk assessment of many other genetic conditions. In the future it
is foreseen to provide an IS that can be used in the Hospitals of the Community of
Valencia (and even in the country) where all the studies and analyzes on
Neuroblastoma can help researchers to continue advancing in the treatment and
monitoring of this disease. It is important to highlight that the use of CM is very
beneficial and efficient for the construction of GeISs, providing a strong structure
for the data they are meant to manage, and capable of adapting heterogeneous,
disperse, redundant and changing environment it ultimately defines.
Acknowledgment
The authors would like to thank the members of the PROS Research Center
Genome group for the fruitful discussions regarding the application of CM in the
medicine field. In addition, we would like to thank Dra. Desireé Ramal, Dra.
Vanessa Segura, Dra. Blanca Martínez, and Dra. Yania Yáñez as experts in
Neuroblastoma for their contribution to this research. This work was supported by
the Spanish Ministry of Science and Innovation through Project DataME (ref:
TIN2016-80811-P), the Preparatory Action - UPV / IIS La Fe (C07-ClinGenNBL,
40
2018) and the Generalitat Valenciana through project GISPRO
(PROMETEO/2018/176).
References
1. Van Dijk, E.L. et al.: “Ten years of next-generation sequencing technology”, Trends
Genet., vol. 30, pp. 418-426 (2014).
2. Mardis, E.R.: “The $1,000 genome, the $100,000 analysis?”, Genome Med., DOI:
10.1186/gm205 (2010).
3. Auton, A., Abecasis, G.: The 1000 Genomes Project Consortium, “A global reference
for human genetic variation”, Nature, vol. 526, pp. 68-74 (2015).
4. Chen, R., Butte, A.J.: “The reference human genome demonstrates high risk of type 1
diabetes and other disorders”, Pac Symp Biocomput., pp. 231-242, Jan. 2011.
5. Gonzaga-Jauregui C., Lupski J.R., Gibbs R.A.: “Human Genome Sequencing in Health
and Disease”, Annu Rev Med, vol. 63, pp. 35-61 (2012).
6. Li, X. et al.: “Genome-wide association study identifies four SNPs associated with
response to platinum-based neoadjuvant chemotherapy for cervical cancer”, Sci Rep,
vol. 7, DOI: 10.1038/srep41103 (2017).
7. Maris, J.M.: “Recent advances in neuroblastoma”, N Engl J Med., vol. 362(23), pp.
2202-2211 (2010).
8. Cao, Y. et al.: “Research progress of neuroblastoma related gene variations”,
Oncotarget, vol. 5, DOI: 10.18632/oncotarget.14408 (2016).
9. Monclair, T. et al.: “The International Neuroblastoma Risk Group (INRG) staging
system: an INRG Task Force report”, J Clin Oncol., vol. 27, pp. 289-297 (2009)
10. Cohn, S.L. et al.: “The International Neuroblastoma Risk Group (INRG) classification
system: An INRG Task Force Report”, J Clin Oncol., vol. 27, pp. 289-297(2009).
11. Castel, V. et al.: “Prospective evaluation of the International Neuroblastoma Staging
System (INSS) and the International Neuroblastoma Response Criteria (INRC) in a
multicentre setting”, Eur J Cancer., vol. 35, pp. 606-611 (1999).
12. Bourdeaut, F. et al.: “ALK germline mutations in patients with neuroblastoma: a rare
and weakly penetrant syndrome”, Eur J Hum Genet., vol. 20, pp. 291-297 (2012).
13. Van Limpt, V. et al.: “The Phox2B homeobox gene is mutated in sporadic
neuroblastomas”, Oncogene, vol. 23, pp. 9280-9288 (2004).
14. Yeh, I-T. et al.: “A germline mutation of the KIF1Bb gene on 1p36 in a family with
neural and nonneural tumors”, Hum Genet, vol. 124, pp. 279-285 (2008).
15. Jimeno Yepes, A., Verspoor, K.: “Literature mining of genetic variants for curation:
quantifying the importance of supplementary material”, Database (Oxford). DOI:
10.1093/database/bau003 (2014).
16. León, A. et al.: “Data Quality problems when integrating genomic information”.
Conceptual Modeling (ER2016): 35th International Conference, 3rd. Workshop
Quality of Models and Models of Quality (QMMQ 2016), pp. 173-182 (2016)
17. Olivé, A.: “Conceptual Modeling of Information Systems”, 1st ed, Springer-Verlag
(2007).
18. Reyes Román, J.F., Pastor, Ó.: Use of GeIS for early diagnosis of alcohol sensitivity.
In Proceedings of the International Joint Conference on Biomedical Engineering
Systems and Technologies (pp. 284-289). SCITEPRESS-Science and Technology
Publications, Lda (2016).
41
19. Roden, D.M., Tyndale, R.F., “Genomic Medicine, Precision Medicine, Personalized
Medicine: What’s in a Name?” Clin Pharmacol Ther., vol. 94, pp. 169-172 (2013).
20. Paton, N. W. et al.: “Conceptual Modelling of Genomic Information”, Bioinformatics.,
vol. 16, pp. 548-557 (2000).
21. Ram, S. Wei, W.: “Modeling the semantics of 3D protein structures”, In Conceptual
Modeling-ER 2004, Springer Berlin Heidelberg, pp. 696- 708 (2004).
22. Garwood, K. et al.: “Model-driven user interfaces for bioinformatics data resources:
regenerating the wheel as an alternative to reinventing it,” BMC bioinformatics., vol.
7, p. 532 (2006).
23. Burriel, V. et al.: “Design and Development of an Information System to Manage
Clinical Data about Usher Syndrome Based on Conceptual Modeling”, BIOTECHNO
(2013).
24. Burriel, V., Pastor, Ó.: “Conceptual Schema of Breast Cancer: the background to
design an efficient information system to manage data from diagnosis and treatment of
breast cancer patients”, IEEE-EMBS International Conference on Biomedical and
Health Informatics (BHI) (2014).
25. Reyes Román, J.F., Pastor, Ó., Casamayor, J.C., Valverde, F.: Applying conceptual
modeling to better understand the human genome. In International Conference on
Conceptual Modeling (pp. 404-412). Springer, Cham (2016).
26. Reyes Román, J.F.: Diseño y Desarrollo de un Sistema de Información Genómica
Basado en un Modelo Conceptual Holístico del Genoma Humano. PhD Thesis,
Universitat Politècnica de València. https://riunet.upv.es/handle/10251/99565 (2018).
27. Reyes Román, J.F., García, A., Rueda, U., Pastor, Ó.: GenesLove.Me 2.0: Improving
the Prioritization of Genetic Variations. In: Damiani E., Spanoudakis G., Maciaszek L.
(eds) Evaluation of Novel Approaches to Software Engineering. ENASE 2018.
Communications in Computer and Information Science, vol 1023. Springer (2019).
28. Cohn, S.L.: “The interactive International Neuroblastoma Risk Group (INRG)
database (db) 2nd year progress report” (2013).
29. Maris, J.M. et al.: “A genome-wide association study identifies a susceptibility locus to
clinically aggressive neuroblastoma at 6p22”, N Engl J Med. (358), pp. 2585-2593
(2008).
30. Capasso, M. et al.: “Replication of GWAS-identified neuroblastoma risk loci
strengthens the role of BARD1 and affirms the cumulative effect of genetic variations
on disease susceptibility”, Carcinogenesis., vol. 34, pp. 605-611 (2013).
31. León, A., Pastor, Ó.: Smart Data for Genomic Information Systems: the SILE
Method. Complex Systems Informatics and Modeling Quarterly, (17), pp.1-23 (2018).
32. Burriel, V. et al.: GeIS based on conceptual models for the risk assessment of
neuroblastoma. In 2017 11th RCIS (pp. 451-452). IEEE (2017).
33. Soler, C.: Diseño de un sistema de información genómica para el diagnóstico del
Neuroblastoma. https://riunet.upv.es/handle/10251/85716, Bachelor degree Project
(2017).
34. Pastor, O., Molina, J.C.: Model-driven architecture in practice: a software production
environment based on conceptual modeling. Springer Science & Business Media
(2007).
35. León, A., Pastor, Ó.: From big data to smart data: a genomic information systems
perspective. In 2018 12th RCIS, pp. 1-11. IEEE (2018).
36. Burriel, V.: Diseño y Desarrollo de un Sistema de Información para la Gestión de
Información sobre Cáncer de Mama. PhD Thesis, Universitat Politècnica de València.
https://riunet.upv.es/handle/10251/86158 (2017).