I am interested in hearing people opinions about what is the best/accepted way to display the fold change values in a graph, specifically the fold change axis scale?
I have seen it presented in one of 2 ways 99% of the time, but neither of these are ideal as far as I can tell: 1. On a scale from 0 to whatever, where values between 0 and 1 represent down-regulation, and values >1 represent up-regulation. But the problem with this method is that it is harder to appreciate the extent of down-regulation. 2. Using inverse values where negative values (<-1) represent down-regulation and positive values (>+1) represent up-regulation. However, the problem with this approach is that a fold change of +1 is essentially equal to -1, and there shouldn’t be anything between these values. So this is wasted space, but more to the point, the difference between, say -1.1 and +1.1, looks much bigger than it actually is.
Why do people not use graphs that show the inverse values, but cut out the section between -1 and +1? So the y-axis may, from the bottom up, go -2, +/-1, 2 etc? Is it purely because the axis would not me mathematically accurate?
I have attached a PDF with the same data plotted in all three ways discussed above.