When we use a regression line to compare two analytical methods, e.g. a new, cheap method and an expensive reference method, we plot the results obtained with the new method against the results (on the same samples!) obtained with the other method. If the two methods give the same result, the best line will be:

y = x

Which method should be plotted on the x axis?
It makes no difference: in the results of both methods errors will be present
The new method, because its errors may be smaller than the reference method
The reference method, because the errors of this method are likely to be smaller than the errors of the other method
You should try both possibilities and pick the line that is closest to y = x