< MARCAR DataTools: Documentation University of Tübingen: Documentation Aktuell

Documentation

Contents



Introduction

The current gold standard for evaluation of the carcinogenic potential of drug candidates is the 2-year chronic exposure rodent assay. The IMI MARCAR project is a ongoing 5-year project aiming to identify early biological indicators for the prediction of carcinogenic effects, particularly for the detection of non-genotoxic carcinogens (NGCs).
The MARCAR DataTools are designed for visualization and analysis of the datasets generated in the course of the MARCAR project. Upon completion the DataTools will provide several approaches to simplify the mechanistic and automated analysis of the diverse datasets, which have been generated using mRNA and miRNA microarrays, DNA methylation analysis, and protein profiling in rats and mice.
The MARCAR DataTools are still under active development and will be updated as new functionality is implemented. If you encounter any errors while using the DataTools or have suggestions for improvements, please write an email to Michael Römer.

How to get started

Currently, the MARCAR DataTools are only available to members of the MARCAR Consortium. To access the tools please register on our Galaxy server with your email address used for MARCAR correspondence or contact Michael Römer for assistance.
The MARCAR DataTools are hosted on our Galaxy server. To use the tools, just select the tool you wish to use and follow the instructions found below the tool in the "How to use the tool" section. To help you select the appropriate tool for your analysis, please check the following short description of available tools. For a detailed description of each tool please refer to the Tools section.
  • MARCAR DataViz

    Generates heatmaps for all data types and box and profile plots of intensity ratios for all datasets. Used to generate an overview of specific datasets or assess the performance of a biomarker across several datasets.
  • MARCAR DataComp

    Generates intensity ratio vs. intensity ratio plots and Venn diagrams of differential genes for pairwise comparison of samples from all datasets. Used to compare two samples and generate hypotheses based on similarities.
  • MARCAR DataLoad

    Generates gene-sample matrices for all data types and datasets. Allows extraction of data from the database for use with your own favourite analysis tool.


  • Tools

    Currently, the MARCAR DataTools comprise three tools used for visualization and data analysis:
  • MARCAR DataViz

    MARCAR DataViz is a visualization tool that aims at simplifying the generation of heatmaps, box plots, and profile plots to generate hypotheses or visually validate biomarkers found in earlier studies. MARCAR DataViz incorporates the datasets obtained from rodent in vivo studies by the MARCAR consortium. Currently, MARCAR DataViz only supports microarray data for mRNA transcripts, support for miRNA transcripts, DNA methylation and protein abundance is planned for future releases. In addition to the MARCAR datasets, MARCAR DataViz offers easy access to the genes on the list of potential biomarkers that was generated in the course of the MARCAR project. Alternatively, the user can select any gene of interest from the genome. Due to different study designs, particularly different choices of the used microarray platform, not all genes are represented in all studies. For further details on specific studies, please refer to the Datasets section.
    MARCAR DataViz further allows the visualization of orthologouos genes in other organisms (currently, only rat and mouse are supported) by mapping the probes on microarrays for different organisms based on sequence homology.
    To facilitate the generation of hypotheses, MARCAR DataViz allows for user-selected cutoffs when working with intensity ratios or p-values. By choosing appropriate cutoffs, the user can filter genes based on their differential expression between treated and control animal tissue.
    Step-by-step instructions on how to use the tool can be found on the Galaxy site of the tool, below the entry mask.
  • MARCAR DataComp

    MARCAR DataComp is a visualization tool for the pairwise comparison of samples from diverse datasets to allow the generation of hypotheses based on similarity or diversity. MARCAR DataComp incorporates the datasets obtained from rodent in vivo studies by the MARCAR consortium. Currently, MARCAR DataComp only supports microarray data for mRNA transcripts, support for miRNA transcripts, DNA methylation and protein abundance is planned for future releases.
    MARCAR DataComp generates two types of plots. First, Intensity vs. intensity ratio plots that allow a comparsion of samples based on the observed expression profiles, e. g., to identify a correlation between gene deregulation. Second, Venn diagramms of deregulated genes as a simplified visualization of similiarities in terms of gene deregulation. The intensity ratio and p-value cutoffs used for the generation of the plots can also be changed.
    Step-by-step instructions on how to use the tool can be found on the Galaxy site of the tool, below the entry mask.
  • MARCAR DataLoad

    MARCAR DataLoad is a tool that provides access to the datasets generated during or previous to the MARCAR project and allows to download gene-sample-matrices for various data types for either a specific dataset or multiple datasets at once. Currently, MARCAR DataLoad only supports microarray data for mRNA transcripts, support for miRNA transcripts, DNA methylation and protein abundance is planned for future releases.
    In addition, MARCAR DataLoad can generate gene-sample matrices containing orthologouos probesets from other organisms (currently, mapping rat to mouse and vice versa is supported) based on probe sequence homology.
    MARCAR DataLoad also offers a filter to only include genes into the gene-sample matrix, if they are deregulated in at least one sample in any selected dataset.
    Step-by-step instructions on how to use the tool can be found on the Galaxy site of the tool, below the entry mask.


  • Data types

    The MARCAR DataTools offer functionality for visualizing and exporting several data types in different stages of processing. Currently, the following data types are offered:
  • Normalized intensity


    Normalized intensity generally refers to the RMA-normalized signal intensity recorded by the image scanner of the corresponding microarray platform. This data type may be useful for researchers who wish to compare the absolute expression values in the samples or use their own post-processing to detect expression differences between treated and control samples.
  • Intensity Ratio


    Intensity ratio refers to the ratio of the normalized signal intensities observed in the treated and the control samples. The intensity ratio per treatment group (i.e., group of biological replicates with equal treatment) is computed as the ratio of the mean of the signal intensity observed in the treated samples and the mean of the signal intensity observed in the control samples. To facilitate the interpretation of the intensity ratios, by default the log2-ratio of the intensity ratio is used. However, most MARCAR DataTools have options to use absolute intensity ratios.
  • p-value


    The p-value is a measure of significance observed in a T-test performed between the treated and control samples. The p-values reported by the MARCAR DataTools have been corrected for multiple testing using Benjamini-Hochberg correction. For the computation of p-values, at least three treamtment and control samples, respectively, are required per treatment group.

  • Please note that depending on the design of the different studies and the tool you are using, the available data types may vary. For further information refer to the Tools and Datasets sections or contact Michael Römer.

    Datasets

    Currently, the following datasets are included in the MARCAR DataTools:
    Dataset Organism DataViz DataComp DataLoad Metadata sheet
    BSP rat study Rat Metadata sheet
    MUW rat liver study Rat Metadata sheets
    MUW rat liver tumor study Rat Metadata sheet
    Rat bladder study Rat Metadata sheet
    Mouse liver tumor study Mouse Metadata sheet
    CD-1 mouse study Mouse Metadata sheet
    CAR KO mouse study Mouse Metadata sheet
    CTNNB1 KO/HU study Mouse Metadata sheet