The practical application of the vast amount of genomic data publicly available for clinical practice is a bottleneck for advancing Precision Medicine. Unresolved issues in the genomics domain include dispersion, heterogeneity, discrepancies, a lack of standardization, and data quality issues. In this project, we create the Delfos platform, a conceptual model-based solution developed with a rigorous methodological and ontological foundation, with the main goal of minimizing the impact of these issues when transferring research findings to clinical practice. To validate this platform, we collaborate with geneticists and doctors to help them improve patient diagnosis and treatment in three health areas: cardiology, oncology, and inherited retinal diseases.
Recent advances in high-throughput sequencing technologies [
One of the most prominent problems to manage genomics data and interpreting the clinical
significance of DNA variants is associated with the vast amount of existing data repositories.
These repositories are classified into three types:
1. Raw Data: Constituted by raw omics data generated by sequencing technologies. These
data can be described in terms of DNA sequences, scientific articles, identification of DNA
variants, proteins, etc.
2. Data Collection: Databases and repositories that collect data on a particular aspect of
genomics. These repositories process and distribute “raw data”, making it available to the
community. Examples of such a repositories include PubMed [
Although these third-layer platforms represent a significant advancement in interpreting the
clinical significance of DNA variants, they are only a partial solution and have their own
limitations. Conflicting results are expected due to diferences in how clinical laboratories interpret
empirical evidence. Furthermore, the complexity and subjectivity of the interpretation process
make automation dificult [
All of these issues together form a bottleneck that prevents precision medicine from becoming a common and standardized practice. This includes i) eficiently translating research findings into clinical practice and ii) assessing the risk-benefit of any genetic-based drug or treatment for patients. In many cases, simple questions like “Is this variant the cause of the symptoms my patient is experiencing?” cannot be answered with confidence.
There is a need for additional tools and platforms to assist genome analysts in managing genomics information and help clinical experts rely more on their decisions.
The main goal of this project is to contribute to make precision medicine a common and standardized practice. This is subdivided as follows.
The first step is developing a Genome Information System (GeIS), called Delfos, that allows for: 1. Increase the automation of the interpretation of the clinical significance of DNA variants. 2. Increase the level of confidence of clinical experts when interpreting the clinical significance of DNA variants.
Delfos is a conceptual model-based platform that can be classified in a novel type, which we called clinical actionability [10], that is placed on top of the “Integration and Interpretation” type. (see Fig. 1). Delfos implements the four stages of the SILE method [11], specially defined for managing genomics data.
The outputs of this step are the following:
The second step is to validate Delfos in a real-world clinical context. The validation considers three use cases. All three use cases aim to improve the interpretation DNA variants to deliver a better diagnostic and treatment to patients in the following areas: cardiology, oncology, and inherited retinal diseases (IDR).
In first use case, we collaborated with two hospitals, namely, the Health Research Institute Hospital La Fe (IIS La Fe) and the Alicante Institute of Health and Biomedical Research (ISABIAL). The DNA sequences of 84 patients with some form of cardiomyopathy were analyzed, and the ifndings were discussed with domain experts.
In second use case, we collaborated with two hospitals, namely, the IIS La Fe and the Valencian Biomedical Research Institute (INCLIVA). The DNA sequences of 84 cancer patients were analyzed, and the findings were discussed with domain experts.
In the third use case, we are collaborating with a research group of the IIS La Fe specialized in rare diseases. First, the DNA sequences of 20 patients with some form IDR were analyzed. Like with the previous use cases, the findings were be discussed with domain experts. We are preparing an additional round of analysis at this moment with additional patients.
After finishing these three use cases, we will be able to assess whether the tool can improve patient diagnosis and treatment as well as if it is ready to be used in clinical practice on a regular basis.
In addition to these, we also validated the platform in a non-clinical context. We collaborated with the Valencian Institute of Agricultural Research (IVIA) in the agri-food field. The tool was used in a food improvement program to make citrus varieties more palatable and resistant to extreme weather conditions. Nearly 90 citrus varieties’ DNA sequences were analyzed, and the results were discussed with domain experts.
The outputs of this step are the following: 1. Validation of Delfos for cardiology: Doctors agreed that the results were correct and reported very positive feedback. A portion of this work has been published in: 4. Validation of Delfos for a non-clinical domain: We assessed that Delfos can also be used in non-clinical contexts that are relevant. Genomics has the potential to revolutionize food production and improve food quality and safety, making it an important tool for food improvement programs. Domain experts reported very positive feedback towards the use of Delfos to achieve their goals. A portion of this work has been published in:
The third step is to connect with non-clinical companies that ofer genetic counseling and could benefit from using the platform. The following actions have been carried out to improve the adoption of the tool: 1. We are in the process of creating a spin-of. We have already started to receive funding in order to define a business plan. We are collaborating with TRL+, an Spanish initiative committed to highlighting research projects in order to turn them into commercially and economically viable triple-impact spin-ofs. 2. We are also working with Viromii, a consulting firm focused on innovation and techtransfer, that supports the dissemination of the project results among the relevant institutions for the transfer of results. 3. The IP of the project has given talks in relevant national forums to disseminate the results of our project.
This project is highly relevant to the topic of Cyber-Human Systems. This topic refers to the interface between humans and technology, and our project explicitly addresses this. CyberHuman Systems aim to advance our ability to efectively work in complex environments, which is also a key objective of our tool. This tool (i.e., Delfos) deals with the practical application of genomic data in clinical practice, which involves the interaction between humans and computer systems. It aims to reduce the impact of various genomics challenges when translating research ifndings into clinical practice. These challenges include dispersion, heterogeneity, discrepancies, lack of standardization, and data quality issues, all of which require a are relevant when considering the interaction between humans and technology.
Furthermore, we contribute to another a relevant topic to the CAiSE community, namely, conceptual modeling. Conceptual modeling is an essential aspect of Information Systems Engineering, as it provides a way to represent and reason about complex systems in a structured way. Using conceptual models allows for a shared understanding among stakeholders, which is essential for successfully designing and implementing information systems. Model-based solutions have gained significant attention in the field of Information Systems Engineering due to their ability to capture complex relationships and interactions between various elements of a system, which can be particularly useful in the healthcare domain. This submission contributes to the field of conceptual modeling by proposing a model-based platform developed with a rigorous methodological and ontological foundation for the practical application of genomic data in clinical practice.
In addition, this project highlights the application of information systems engineering in the healthcare domain, a significant area of research and development; this aligns particularly well with the conference’s emphasis on innovative and rigorous research. As mentioned above, the main goal of the work herein presented is to minimize the impact of unresolved genomics issues when transferring research findings to clinical practice. This is a significant and timely challenge, as the correct management of the vast amount of genomic data generated by current sequencing technologies is a bottleneck for advancing Precision Medicine.
Finally, this project has the potential to stimulate further discussion in the CAiSE community about the use of Advanced Information Systems Engineering in healthcare and Precision Medicine, particularly in the context of model-based solutions.
Based on the three steps defined in Section 2, we indicate the current progress of the project (see Table 1). The design and implementation phases of the model-based tool have been completed, and it operates as expected. Two of the three use cases have concluded with promising results; the third is in the planning phase and will begin in a few months. To date, we have contacted four potential customers for our tool: the three hospitals with which we collaborated on the use cases, as well as BIONOS, the company that performed the patient sequencing process.
We wish to thank the IIS La Fe, INCLIVA, ISABIAL, and BIONOS for their huge contribution to this project and their willing to help. This work was supported by the Spanish State Research Agency through the PDC2021-121243-I00 grant (MICIN/AEI/10.13039/501 100011033) and coifnanced with ERDF and the European Union Next Generation EU/PRTR.
In progress (66%) In progress (50%)
The cardiology and oncology use cases have been validated. The IDR use case is on going.
We have started the creation of the spin-of and are currently contacting potential partners. variants, Genetics in Medicine 22 (2020) 336–344. doi:10.1038/s41436-019-0655-2. [10] A. León, A. García S., M. Costa, A. Vañó Ribelles, O. Pastor, Evolution of an adaptive information system for precision medicine, in: S. Nurcan, A. Korthaus (Eds.), Intelligent Information Systems, Lecture Notes in Business Information Processing, Springer International Publishing, 2021, pp. 3–10. doi:10.1007/978-3-030-79108-7_1. [11] A. León, J. Reyes, V. Burriel, F. Valverde, Data Quality Problems When Integrating Genomic Information, in: S. Link, J. C. Trujillo (Eds.), ER 2016 Workshops, LNCS, Springer International Publishing, Gifu, Japan, 2016, pp. 173–182. doi:10.1007/978-3-319-47717-6_15.