Towards an AI-based Genomic Medicine of Precision that Integrates Predictive and Explainable Knowledge Dimensions

• Óscar Pastor Universitat Politècnica de Valencia
• Salvador Navarro Universitat Politècnica de Valencia
• Alberto García Universitat Politècnica de Valencia
• Mireia Costa Universitat Politècnica de Valencia
• Ana León Universitat Politècnica de Valencia

Resumo

Understanding the human genome and deciphering the Language of Life is a grand challenge that modern sequencing technologies are making feasible by generating huge amounts of data whose correct interpretation has yet to be accomplished. To do it, two knowledge dimensions must be integrated: the predictive one, Machine Learning-oriented, that obtain accurate information from data, and the explainable one, Conceptual Modeling-based, that uses a symbolic representation to provide meaning to the data in order to understand and explain the semantics behind predictions. This position report discusses the problem, contextualizes it under a Life Engineering perspective, and it proposes how to face the design of AI-based data management platforms that follows the introduced ideas.

Referências

Spreeuwenberg, S. ([S.d.]). AIX: Artificial intelligence needs explanation. – Transparency increases the success of AI based decision support systems.

Costa, M., García, A. S. and Pastor, O. (1 dec 2023). The consequences of data dispersion in genomics: a comparative analysis of data sources for precision medicine. BMC Medical Informatics and Decision Making, v. 23, n. 3, p. 1–17.

Costa, M., García S, A. and Pastor, O. (2022). Conceptual Modeling-Based Cardiopathies Data Management. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), v. 13650 LNCS, p. 15–24.

Guizzardi, G., Pastor, O., Storey, V. C. and Kruchten, P. (1 nov 2023). Thinking Fast and Slow in Software Engineering. IEEE Software, v. 40, n. 6, p. 139–142.

León Palacio, A., García Simón, A., Reyes Román, J. F. and Costa, M. (mar 2024). The Delfos Platform: A Conceptual Model-Based Solution for the Enhancement of Precision Medicine [acepted for publication].

Liñares-Blanco, J., Pazos, A. and Fernandez-Lozano, C. (12 jul 2021). Machine learning analysis of TCGA cancer data. PeerJ Computer Science, v. 7, p. 1–47.

Palacio, A. L. and López, Ó. P. (27 dec 2018). Smart Data for Genomic Information Systems: the SILE Method. Complex Systems Informatics and Modeling Quarterly, v. 0, n. 17, p. 1–23.

Pastor, O. (2008). Conceptual modeling meets the human genome. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), v. 5231 LNCS, p. 1–11.

Pastor, Ó., León, A. P., Reyes, J. F. R., Garciá, A. S. and Casamayor, J. C. R. (18 jan 2021). Using conceptual modeling to improve genome data management. Briefings in Bioinformatics, v. 22, n. 1, p. 45–54.

Pastor, O. and Molina, J. C. (2007). Model-driven architecture in practice: A software production environment based on conceptual modeling. Model-Driven Architecture in Practice: A Software Production Environment Based on Conceptual Modeling, p. 1–302.

Reyes Román, J. F., Pastor, Ó., Casamayor, J. C. and Valverde, F. (2016). Applying conceptual modeling to better understand the human genome. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), v. 9974 LNCS, p. 404–412.

Swanson, K., Wu, E., Zhang, A., Alizadeh, A. A. and Zou, J. (13 apr 2023). From patterns to patients: Advances in clinical machine learning for cancer diagnosis, prognosis, and treatment. Cell, v. 186, n. 8, p. 1772–1791.