emquardokus
commented
1 week ago @andreasbueckle tissue blocks can be taken from the wall of the renal pelvis but you will not get cells from a hole/cavity.
So if the blocks are in the center of the renal pelvis (the center of the tube), they are not correctly registered and those should be confirmed to be actually from renal pelvis and not meant to be in the adjacent kidney tissue next to the renal pelvis. I've not see datasets nor have I heard of any sampling from the renal pelvis from the kidney team.
Urothelial cells line the renal pelvis.
Renal pelvis is contiguous with the ureter in a similar manner to large blood vessels narrow to smaller diameter blood vessels.
Please verify that the wall of renal pelvis were really sampled. I think this is a mistake.
andreasbueckle
For the kidney (4 sites), the WUSTL team registered all sites into the renal pelvis, which is a LANDMARK, and hence we get no collisions when landmarks are turned off (it is a cavity):
Same view with landmarks on (the site now lies neatly in the renal pelvis landmark):
View of the renal pelvis landmark in Babylon (with BB):
A solution for this issue could be to make the renal pelvis a crosswalked AS that gets picked up by the mesh and BB collisions. For this, it needs an UBERON or FMA ID. There exists one for it at https://www.ebi.ac.uk/ols4/ontologies/uberon/classes/http%3A%2F%2Fpurl.obolibrary.org%2Fobo%2FUBERON_0036295. Details are TBD. Unsure how to proceed in a timely manner since we have a way to go before the next release where the 3D Reference Object and asct-b-3d-models crosswalk (https://lod.humanatlas.io/ref-organ/asct-b-3d-models-crosswalk/latest/) can be adjusted if that is the route we want to go.
As of now, we need to drop 61 datasets: HBM455.PRVR.965: 2 datasets HBM486.QXHD.874: 4 datasets HBM556.CFNP.485: 52 datasets HBM752.DHVB.758: 3 datasets So total of 61