Data structures for transcriptomics

The two main data structures used for transcriptomics data analysis in the Biomedical Genomics Workbench are tracks and experiments.

Tracks, also known as 'Genome Browser View', (see Genome Browser Tools) are the fundamental building blocks for data analysis in the Biomedical Genomics Workbench, where all information is tied to genomic positions. A central coordinate-system is provided by a reference genome, which allows different types of data or results for different samples to be seen and analyzed together.

Experiments (see Experiments), on the other hand, are used to represent complex relationships between expression samples, and to carry out statistical analysis (see statistical analysis) of differential expression.

Tracks and experiments are intimately related, and it is possible in most cases to convert from one type to the other.

From Tracks to Experiments

When carrying out RNA-seq analysis using the RNA-Seq Analysis tool, the starting point is a set of reads from a sequencing study. As part of the RNA-Seq Analysis tool, these reads are mapped onto a reference genome. The RNA-Seq tool produces expression tracks, which are compatible with the reference genome, and can be visualized together with the genome in the

Genome Browser View (see Genome Browser Tools).

You can find more information about the RNA-Seq Analysis tool in RNA-Seq Analysis).

Once expression tracks have been obtained from the RNA-Seq Analysis tool, they can be used as sequencing-based sets of expression values in setting up an experiment. This can be done using the Set up Experiment tool and is described in more detail in Experiments.

An experiment set up in this manner from expression tracks is intimately coupled to the tracks it originated from. To see this coupling in action, perform the following steps:

  1. Use the Set up Experiment tool on two or more expression tracks to set up an experiment, as described in Experiments.
  2. Save and open the resulting experiment, by double-clicking its name in the Navigation Area.
  3. Use the Create New Genome Browser View tool to create a new Genome Browser View from the expression tracks you used to set up the experiment, as described in Genome Browser Tools.
  4. Save and open the resulting Genome Browser View by double-clicking its name in the Navigation Area.
  5. Drag the experiment tab downwards, until you see the blue shadow indicating the resulting placement (figure 30.37), and drop it in place. You should now have a divided view, with the experiment in the bottom half (figure 30.38).
  6. Clicking on any line in the experiment will now automatically jump to the corresponding genomic location in the upper view. Use the Zoom to Selection (Image zoom_selection) button to zoom in to the desired genomic region.

Image drag_tab_to_drop_cancer
Figure 30.37: Dragging a tab to the lower half of the view area.

Image two_panel_experiment_cancer
Figure 30.38: After dropping a tab to the lower half othe view area.

From Experiments to Tracks

Experiments can be used to carry out statistical analysis on the expression values obtained from RNA-seq analysis as described in statistical analysis. The results of the statistical analysis are annotated on the experiment as additional columns.

It can be advantageous to visualize the results of the statistical analysis as tracks. The Extract Differentially Expressed Genes tool in the Biomedical Genomics Workbench enables the conversion of experiments to tracks. You can find the Extract Differentially Expressed Genes tool here:

        Toolbox | Microarray and Small RNA | Extract Differentially Expressed Genes (Image create_tracklist_16_h_p).