Importantly, as defined here, the measure of expression consistency is expected to be biased toward generating higher estimates for highly expressed and long genes.
It uses CP values (instead of relative quantities) as input and employs a different measure of expression stability from geNorm and NormFinder.
Since BPND values are often used in quantification of PDE10A as a measure of expression, it is important to also take affinity effects into account when quantifying PDE10A.
Some studies used the Euclidean distance of expression profiles as a measure of expression divergence [5], [28], [29], [30].
The read counts in these two categories were considered a measure of expression pertaining to the specific locus.
We did not use the correlation coefficient as a measure of expression divergence since there are problems with this measure for genes that are uniformly expressed (unpublished results).
Some studies used 1-EC as a measure of expression divergence [20], [21], and in this case the agreement between the assessed computational methods should be the same as the above analysis.
Second, Urrutia et al. [20] have reported that ED between human and mouse is correlated to the numbers of new Alu elements in humans, although the direction of the correlation depended upon the measure of expression divergence employed.
These studies suggest that the direct use of EST data might be problematic and that it should be normalized for length if transcript abundance is to be used as a measure of expression levels.
We also calculated the angle between the expression values and the log expression values as an alternative measure of expression divergence and obtained very similar results (results not presented).
Employing PCC as a measure of expression divergence yielded no significant correlation (Kendall's tau = 0.03 p value = 0.34) between expression similarity and DNA binding site motif similarity (Figure 1 in supplementary Data file S1).
