In the aggregation notebook in spatialdata-notebooks we color shapes at some point by aggregated points of gene b. However, it seems b is interpreted as color blue and the values in the table are ignored for the fill color:

Also one can see that the alpha is ignored. This is also the case currently in the notebook version in the scverse docs.