The TRANSMET-CODA project gears around the application of statistical methodology for the analysis of compositional data (CoDa) in applied science and engineering. A second important goal concerns the development and improvement of compositional methods, in particular for applied scientific problems where this is needed. The project consists of three lines of investigation: a) Biomarkers and health science, b) Natural hazards and climate change, and c) Characterization, quantification and comparison of social and economical systems. The first line, TRANSMET-1, led by Dr. Graffelman, concerns the analysis of large databases of biomarkers, such as omics data (metabolomics, proteomics), genetic polymorphisms obtained by whole genome sequencing, and data from microbiome studies based on DNA sequencing. These types of data require techniques from the fields of multivariate analysis, statistical genetics and compositional data analysis. Methodological compositional contributions for the graphical representation of genetic data, for relatedness research, and for improved computational efficiency are important aspects of this research line. The statistical analysis of biomarker data is an active area of research, full of interesting challenges, right at the intersection of computer science, statistics, compositional methodology and biological knowledge, and relates directly to challenge 1 on human health of the current call. The second line, TRANSMET-2, led by Dra. Ortego, concerns the application of compositional methods for the study of multivariate extremal phenomena, considering their interdependence, with the use of Bayesian statistical techniques for quantifying the uncertainty in the estimations. This research line addresses challenge 5 on climate change and use of natural resources. The application of CoDa techniques for the detection of change points in compositional data with a spatio-temporal character of considerable interest. It is also important to consider the appropriate scale for the measure of dependence. For this purpose we employ new theoretical developments based on copulas in order to represent dependence. The study of dependence for discrete and mixed data with compositional techniques is also addressed. From this perspective, the use of statistical techniques for mixed data such as factor analysis or multidimensional scaling will be reconsidered. The third line, TRANSMET-3, led by Dr. Pérez and Dr. Ortego deals with challenges 6 (social sciences), 1 (health and demographic change), and 4 (transport) of the call. The main focus of this line is on the spatio-temporal evolution of social and economic indicators of the human population, which are used in many contexts. Indicators regarding WASH (Water Sanitation and Hygiene), basic services and international sustainable development, in particular those related to the Sustainable Development Goals, SDG, are, among others, considered. The use of compositional methods for monitoring theseaspects internationally requires a methodological development, in particular if the evaluation takes place at different administrative levels. Transmet-3 also introduces compositional techniques into transport engineering (e.g. in the evaluation of the international transport of goods) and the compositional characterisation of the food system by means of compositional differential equations.
Plan Estatal de Investigación Científica y Técnica y de Innovación 2017-2020
Programa Estatal de I+D+i Orientada a los Retos de la Sociedad
Galvan, I.; Graffelman, J.; Torrents, D.; Sumoy, L.; Moreno, V.; de Cid, R. International Workshop on Compositional Data Analysis p. 94 Presentation's date: 2019-06-07 Presentation of work at congresses