RFLOMICS: R package and Shiny interface for Integrative analysis of omics data - Archive ouverte HAL Accéder directement au contenu
Poster De Conférence Année : 2022

RFLOMICS: R package and Shiny interface for Integrative analysis of omics data

Nadia Bessoltane
Gwendal Cueff
Audrey Hulot
Delphine Charif

Résumé

The acquisition of multi-omics data in the context of complex experimental design is a widely used practice to identify entities and decipher the biological processes they are involved. The investigation of each omics layer is a good first step to explore and extract relevant biological variability. The statistical integration could then be restrained to pertinent omics levels and features. Such analysis of heterogeneous data remains a technical challenge with the needs of expertise methods and parameters to take into account data specificity. Furthermore, applying different statistical methods from several tools is also a technical challenge in term of data management. In this context, we developed RFLOMICS, an R package with a shiny interface, to ensure the reproducibility of analysis, with a guided and comprehensive analysis and visualization of data in a framework which can manage several omics-data and analysis results. RFLOMICS currently supports up to three types of omics (RNAseq, proteomics, and metabolomics), and can deal with multi-factorial experiments (up to 3 biological factors). It includes methods chosen based on expert feedback. This application is divided into three key steps. The first step allows the user to import the experimental design file and abundance matrix for each dataset (read counts for RNA-Seq, signal intensity for metabolomics and proteomics), and set up the statistical model and contrasts associated to the biological issues. The second step is to perform a full analysis for each dataset, which includes : i- quality control to check for batch effects or identify outlier samples that can be removed, ii- filtering and normalization of RNA-Seq data, or transformation of prot/meta data, iii- differential expression analysis using edgeR for RNA- Seq and limma for prot/meta data, iv- co-expression analysis using coseq, and finally, v- functional enrichment analysis. The third step is to integrate selected omics layers using the unsupervised methods proposed by MOFA. All the results as well as the raw data, and all information necessary for reproducibility of analysis are managed and stored thanks to the MultiAssayExperiment object. An HTML report can be generated, summarizing all analysis steps, using rmarkdown R package. RFLOMICS provides the same framework that allows the user to perform the analysis of multi-omics project from A to Z, taking into account the complexity of the design. It guarantees the relevance of the used methods, and ensures the reproducibility of the analysis. The interface offers an interesting flexibility between the visualization of the results and the data manipulation (filtering, parameter setting). Future development will include the implementation of supervised integration methods, and a docker image to facilitate deployment.
Fichier principal
Vignette du fichier
2022-07-05_JOBIM_poster_rflomics.pdf (7.38 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04546139 , version 1 (15-04-2024)

Identifiants

  • HAL Id : hal-04546139 , version 1

Citer

Nadia Bessoltane, Christine Paysant-Le-Roux, Gwendal Cueff, Audrey Hulot, Delphine Charif. RFLOMICS: R package and Shiny interface for Integrative analysis of omics data. Journées ouvertes en biologie, informatique et mathématiques (JOBIM), Jul 2022, Rennes, France. . ⟨hal-04546139⟩
0 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More