I'm using very noisy biological data, and I need to assume, when doing linear regression, that there are errors in both X and Y. To my understanding normal regression only assumes errors in Y. Is there a way to specify error values for X and Y in Mathematica, and take that into account when finding a linear relationship between the two?