AVALOS, Marta; BARRERA, John; MEZA, Cristian; EYHERAMENDY, Susana; VANDENBORGHT, Louise-Eva; ENAUD, Raphaël; DELHAES, Laurence

Métadonnées

Licence d’utilisation du document

AVALOS, Marta
Statistics In System biology and Translational Medicine [SISTM]
Bordeaux population health [BPH]
Centre Inria de l'Université de Bordeaux
Université de Bordeaux [UB]

BARRERA, John
Universidad de Valparaiso = Valparaiso University

MEZA, Cristian
Universidad de Valparaiso = Valparaiso University

Langue

Communication dans un congrès

Ce document a été publié dans

ICSDS 2024 - International Conference on Statistics and Data Science, 2024-12-16, Nice.

Résumé en anglais

Studies on chronic respiratory diseases and their associations with the microbiome -whether from the respiratory tract, gut (gut-lung axis), or environmental sources (e.g., indoor dust)- present unique statistical challenges. These studies often involve small sample sizes and high-dimensional data, with hundreds or thousands of bacterial and fungal abundances measured. The data are compositional, frequently zero-inflated, and either geographically grouped (e.g., environmental microbiome and asthma in the COBRA-Env study) or longitudinal (e.g., cystic fibrosis patients in the French LumIvaBiota cohort).Meta-analysis is an essential tool for synthesizing knowledge by pooling data from multiple studies to increase statistical power. However, microbiome research often shows high variability across studies in terms of methodologies, sample collection, sequencing techniques, and population characteristics. Exploring the sources of heterogeneity is crucial for ensuring comparability. Penalized meta-regression methods, such as the Lasso, provide a way to address heterogeneity in cases with few studies (such as those on chronic respiratory diseases linked to the microbiome) and many potential explanatory factors.In this talk, we will illustrate the statistical challenges associated with microbiome studies in chronic respiratory diseases and discuss strategies to address these complexities.< Réduire

Mots clés en anglais

Biostatistics

Machine Learning

High-dimensional

Application

Computation

URI

https://oskar-bordeaux.fr/handle/20.500.12278/206933

Métadonnées

Licence d’utilisation du document

Challenges of Microbiome Data Analysis in Chronic Respiratory Disease Studies: Examples from Three French Studies

Langue

Ce document a été publié dans

Résumé en anglais

Mots clés en anglais

URI

Unités de recherche

Métadonnées

Partager cette publication !

Licence d’utilisation du document

Challenges of Microbiome Data Analysis in Chronic Respiratory Disease Studies: Examples from Three French Studies

Langue

Ce document a été publié dans

Résumé en anglais

Mots clés en anglais

URI

Unités de recherche