Fig. 1

Regression analysis on gene saturation indexes for TraDIS dataset. The histogram represents the frequency of saturation indexes of genes. A gene saturation index is the ratio of insertion sites with insertions on the number of potential insertion sites. The regression uses the density of frequencies, as we are using density functions to fit the data. A parametric regression algorithm is used to divide the bimodal distribution into two distributions. An exponential distribution is fit to essential genes, and a Gamma distribution to non-essential genes. Although the Gamma distribution deviated from observation at higher saturations, the two distributions fit the data nicely in the saturation range where the division happens