Fig. 2

Raw read counts for the gene TP53 from the Australian Genome Research Facility site arranged by sample types (a, c, d, and b). The Y axis shows the read counts. The blank space in the middle represents where a 50–50 mixture of (a and b) would be located if one had been created and measured. By leaving this blank space, a visual interpretation can be made for the linearity between (a and b) by whether (c and d) mixture models fall on this linear line. If C or D do not fall on the linear relationship of A and B then the normalization method is imposing unwanted structure on the data. If all four samples (a, b, c and d) form a clear linear relationship then that normalization method is representing the true biological structure of the data