After a few seconds we get a summary and an overview of the data in a boxplot window. Another window is on the right, let’s ignore it by now.

We can see that there are several conditions, normalized although a bit skewed towards high expression outliers.

The boxplot is also a parallel coordinates, where you can select profiles based on expression levels, by dragging the small handles at the tips of each boxplot.

Let’s drag the bottom handle at the last column until expression reaches approximately 4.0. We get over 500 genes. They are summarized as a large colored area.

But... wait, what does column names mean?

Yes, that GSM77333 is not very informative...

Press Ctrl+l (or go to Analysis/Show Labels...) and select Cycle and IntervalGroup (select several items by holding shift). These are experimental factors (more info on the GEO entry for GSE3431)

Ok, so these are genes upregulated on late timepoints of the cell cycle. Let’s narrow the search by selecting those that are also lower on mid timepoints. Drag the top handle of any mid condition to approximately 2.0

Now we have around a dozen of genes, and a nice profile, periodical along the cycles.

On the right window, we can see their heatmap. Row names are uninformative as they are probe ids... Ctrl-l again and select GENENAME.

We can keep on exploring the data, just remember that you can use Ctrl+Z,Y,0 to undo, redo or reset selections.

However, we could require a more systematic and rigorous analysis. Lets select Analysis/Differential Expression...

Ahhh, p-values and stuff, now we are talking. Ok, here you can select two groups of experimental conditions and perform a differential analysis (e.g. early vs late). Or you can select the entire set of experimental factors and perform all comparisons at once.

Let’s select the IntervalGroup entry on both lists (default parameters should be fine).

We’ll get a differential analysis for every combination of conditions, and see a Venn-like diagram with each Differential Expression Group (DEG)

There are four DEGs, two of them larger and highly overlapped. Click on EARLY VS LATE (genes with high expression on early timepoints respect to late ones). Its genes will be visualized on the other views

You can continue analyzing results, performing other comparisons or trying other methods.

As a last feature, we can add up some info about gene functions. With the selected group above, click on View/Word Cloud. It opens a fourth window with the compilation of GO BP annotations found for these genes. The number in parenthesis and the size correspond to the number of selected genes that are annotated with the term.

Let’s take a tour through BicOverlapper 2.0 major features. You can also watch it here.