Programa del congreso
Resúmenes y datos de las sesiones para este congreso. Seleccione una fecha o ubicación para mostrar solo las sesiones en ese día o ubicación. Seleccione una sola sesión para obtener una vista detallada (con resúmenes y descargas, si están disponibles).
|
Resumen diario |
| Sesión | |
|
Mi-S2.1-GIB I: Premios SEIB-Fenin (I) Lugar: Salón de Actos Presidente de la sesión: Elisabete Aramendi Ecenarro Presidente de la sesión: Juan Carlos Antony García Presidente de la sesión: Pablo Laguna Presentación de trabajos seleccionados para el Premio SEIB-FENIN para estudiantes del Grado en Ingeniería Biomédica. | |
    | |
15:11 - 15:22
Development of a Robust Data Processing Pipeline for GC-IMS: Correction of Technical Variability 1: Signal and Information Processing for Sensing Systems, Institute for Bioengineering of Catalonia (IBEC); 2: Department of Electronics and Biomedical Engineering, Universitat de Barcelona Abstract—Gas Chromatography–Ion Mobility Spectrometry (GC-IMS) is increasingly used for untargeted metabolomic profiling in biomedical research [1]. However, the stability of GC-IMS signals can be compromised by technical variability introduced during acquisition [2], [3]. These effects reduce data reliability and compromise the performance of downstream analyses, including classification and biomarker detection. This work presents a robust data correction pipeline designed to mitigate the impact of acquisition-related variability. Using 135 GC-IMS measurements from a single pooled urine sample acquired over nine consecutive sessions, two major sources of systematic noise were identified: the batch index (measurement session) and the elapsed time at room temperature before injection (acquisition order within a batch). Independent linear models fitted to each compound’s intensity revealed that approximately 35% of the total variance could be attributed to the batch index, and a 30% to elapsed time, suggesting a substantial linear association between acquisition order and signal variability. An orthogonal projection approach was applied in the sample space to eliminate the signal components linearly associated with the external variables, by projecting each feature intensity onto the subspace orthogonal to these effects. After correction, the proportion of technically stable features (Relative Standard Deviation, RSD < 20%) increased from 22.6% to 71%. The proposed pipeline offers a lightweight and generalizable solution to enhance GC-IMS data quality, improving its stability and reliability for classification and interpretation in metabolomic studies.
| |