[mg-rast] MG-RAST: normalisation on base pairs

Hi, as far as I understood the underlying algorithms, a normalisation based on the number of assinged reads to a subsystem protein. However, what about a normalisation method based on the total number of assinged base pairs? Is it possible, to extract this data out of the output from the analysis step? Best regrads, Stefanie Muehlhausen