POMA: Statistical Analysis Tool for Mass Spectrometry Data - 2020 Shiny Contest Submission

POMA: Statistical Analysis Tool for Mass Spectrometry Data

Authors: Pol Castellano-Escuder
Working with Shiny more than 1 year

Abstract: POMA ShinyApp is an easy-use interactive tool for statistical analysis of mass spectrometry data (e.g. proteomics or metabolomics). This app is based on POMA R package (https://pcastellanoescuder.github.io/POMA/) and its hosted in http://polcastellano.shinyapps.io/POMA/.

Full Description: MOTIVATION:

Like other high-throughput technologies, mass spectrometry usually faces a data mining challenge to provide an understandable and useful output to advance in biomarker discovery and precision medicine. Usually, mass spectrometry data is treated by univariate statistical analyses or by multivariate methods (supervised and unsupervised).

At the same way, biological interpretation of the results is one of the hard points and high knowledge of statistical analysis and computational programming is usually required. For this reason, several bioinformatics tools have emerged to simplify and improve the interpretation and understanding of the results. However, sometimes these tools don't accept complex databases, for example, those that have several covariates.

RESULTS:

Here we propose a free, friendly and fast online Shiny interface for the analyses and visualization of mass spectrometry data.

POMA allows users to go from the raw data to statistical analyses. The four main blocks of the analysis are the “Load Data” panel (where users can upload their data and an optional file of covariates), “Pre-processing” panel (that includes missing value imputation and normalization), "Summary Plots" panel (where users can visualize their data with interactive volcano plots, boxplots and heatmaps) and “Statistical analysis” (that include univariate and multivariate methods, correlation analysis, feature selection methods such as Lasso and Ridge regression, random forest, rank products, etc.). All these steps also include multiple types of interactive data visualization integrated in an intuitive user interface that no require programming skills to be used.

POMA slides at Toulouse useR!2019 (POMA app has been updated several times since this presentation): https://github.com/pcastellanoescuder/POMA_slides_useR2019


Category: Research
Keywords: metabolomics, proteomics, statistical-analysis, bioinformatics, visualization, mass-spectrometry, shinydashboard
Shiny app: http://polcastellano.shinyapps.io/POMA/
Repo: https://github.com/pcastellanoescuder/POMA_Shiny
RStudio Cloud: https://rstudio.cloud/project/1043415

Thumbnail:

Full image: