I am trying to better understand the target distribution and target specificity scores and I came across a few cases that confused me. The documentation states that the tissue distribution for a gene should be 0 when expression is seen in more than 1/3 of the tissues, but not all the tissues. However in the example below, all of the genes seem to exhibit the same level of broad in tissues detection, but SPAST has a score of -1 while the others have 0. Can you explain more precisely how this is calculated and why there is a difference between these genes? Is the code for this available somewhere?
Nice to have you around here. It’s not surprising you are confused because there is a data inconsistency in this widget. Whereas the traffic-light color is derived from the assessment performed by HPA using their transcriptomics data, the data visualised in the widgets is derived from a meta-analysis of the ExpressionAtlas transcriptomics data.
Whereas is generally consistent there are some inconsistencies like the one observed in SPAST. This is of course not ideal and we are working towards a revised version of the underlying data that will make the assesment more consistent.
