Hi,
This reference is very valuable and I would like to use it to map-reference my data. I have downloaded the h5 file "HumanFetalBrainPool.h5". Although it is very easy to get the CellClass, I think everyone would benefit from using also the 51 annotations you specify in AnnotationDescription. Still, the only way to link the cells with those 51 annotation types, is using the clusters, via the AnnotationPosterior matrix.
When I try to interpret this matrix, I found that the maximum annotation posterior for each cluster (which is 1) is usually associated with more than one annotation name. How I can link each cluster unambiguously to the corresponding annotation and use this as the "cell type" rather than the general cell class?
Hi,
This reference is very valuable and I would like to use it to map-reference my data. I have downloaded the h5 file "HumanFetalBrainPool.h5". Although it is very easy to get the CellClass, I think everyone would benefit from using also the 51 annotations you specify in AnnotationDescription. Still, the only way to link the cells with those 51 annotation types, is using the clusters, via the AnnotationPosterior matrix.
When I try to interpret this matrix, I found that the maximum annotation posterior for each cluster (which is 1) is usually associated with more than one annotation name. How I can link each cluster unambiguously to the corresponding annotation and use this as the "cell type" rather than the general cell class?