Hi there,
Is there a way of knowing whether a matrix was logarithmized just by looking at it (if someone else did the analysis).
Hi there,
Is there a way of knowing whether a matrix was logarithmized just by looking at it (if someone else did the analysis).
What do you know about the range of the initial data? If you assume that the logarithm has been taken, then the initial data cannot have zero or negative data. So you could check if you see negative values.
Still, it's dangerous to do something like that: you won't be able to tell the base of the logarithm (2, e, 10?) even if you can decide if it's likely that the data has been transformed.
Is this matrix a gene x expression microarray matrix?
If so, plot the distribution of the intensities for each column ... what is the range on the x axis? If each distribution looks "somehow" normal, it was most likely log transformed, and the range on the x-axis is in "the tens". If the distribution just looks like a spike, and the x-axis range is in the thousands, then it most likely has not been log transformed.
But this isn't the best way to approach this, as you don't know what values are used in your matrix, or what normalization was done to the data, etc. If this is data you are seriously considering using, it's probably best to ask "the source" what's going on with it.
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.