Sharing advances in compositional data analysis
In 2025, Irene Creus-Martí, member of the IMaLeVICS research group, participated as an invited speaker at the 11th Course and Open Seminar on Compositional Data Analysis (CoDaCourse 2025), organised by the University of Girona.
The event is one of the main international references for compositional data analysis, bringing together experts in statistics, geology, chemistry, environmental science, and biology to explore theoretical advances and new applications of CoDa methodology.
Compositional methodology for microbiome research
In her invited seminar, Irene presented methodological developments focused on the analysis of microbiome datasets through compositional data analysis. This approach allows researchers to properly interpret relative abundance data — a key challenge in microbiome studies, where the total sum constraint makes conventional statistical methods inappropriate.
Her presentation illustrated how CoDa-based tools can help understand bacterial dynamics and temporal variability, offering interpretable and robust models for biological data. You can find more details in the paper “Compositional dynamic models for bacterial time series” published in Computers in Biology and Medicine.
🔗 Learn more about the CoDaCourse community at University of Girona – CoDa Research Group
Strengthening collaboration and interpretability
Participation in CoDaCourse 2025 reinforces IMaLeVICS’s commitment to interpretable modelling and to collaboration with national and international experts in statistical innovation. These exchanges are key to developing reliable computational tools, such as the R package CoDaLoMic, presented earlier at ICSDS 2024.
IMaLeVICS continues to advance in the development of interpretable machine learning techniques applied to vision, imaging, and biological systems — bridging statistical rigour and real-world impact.