Dear Allen Brain Institute team
I am currently analyzing gene expression data from the Allen Brain Cell Atlas and have a question regarding the histograms provided for individual genes.
I noticed that the bin widths in some histograms appear to be slightly inconsistent (e.g., alternating between values such as 0.36 and 0.37).
This raised the possibility that the bin boundaries may have been rounded (for example, to two decimal places) for display purposes.
I would greatly appreciate it if you could clarify the following:
-
Are histogram bin assignments computed using the original high-precision values (prior to any rounding), with rounding applied only for visualization?
-
Or are the bin boundaries (and/or data values) rounded before binning is performed?
In other words, can the displayed irregular bin widths be interpreted as rounding artifacts, while the underlying binning remains uniform and based on higher-precision values?
Understanding this point is important for accurately interpreting the distributions and for any downstream quantitative analysis.
If the underlying binning is performed using higher-precision values, is it appropriate to assume that the original bins were uniformly spaced, and to reconstruct evenly spaced bin boundaries from the displayed values? Or could such reconstruction introduce inaccuracies due to uncertainty in the original bin edges?
Thank you very much for your time and for providing such a valuable resource to the community.
