Brainstem #128
Brainstem #128
Conversation
There was a problem hiding this comment.
The terms file is missing (please also rename - see comment) and a syntax error meaning I can't run it to check. Please check it is working for you.
I realise you're moving on so some of these task may not need to be done by you (may be done later), but it would be good to get it running properly and in the codebase.
| eIndexPM = {} | ||
| xiPM = {} | ||
| pointMarkers = {} | ||
| eIndexPM['caudal-dorsal'] = int(elemPerLayer/2) | ||
| eIndexPM['midRostCaud-dorsal'] = int(elemPerLayer / 2) + (elemPerLayer * int((elementsCountAlong/2)-1)) | ||
| eIndexPM['rostral-dorsal'] = (elemPerLayer*(elementsCountAlong-1)) + int(elemPerLayer/2) | ||
| eIndexPM['caudal-ventral'] = int(elemPerLayer/2) - (elementsCountAcrossMinor-1) | ||
| eIndexPM['midRostCaud-ventral'] = eIndexPM['midRostCaud-dorsal'] - (elementsCountAcrossMinor-1) | ||
| eIndexPM['rostral-ventral'] = int((elemPerLayer*(elementsCountAlong-1)) + (int(elemPerLayer/2) - elementsCountAcrossMinor + 1)) |
There was a problem hiding this comment.
We're now adding the marker points as annotation groups and putting the points in not just the marker group, but the brainstem group and their own annotation group.
We will add the point annotations with the name and term Id "None" (a string with the word None in it) and request the terms from SciCrunch using the google sheet the mappers maintain.
All the names need to be qualified with "brainstem" and point/junction so they are clear what anatomy they are for in SciCrunch. Since I can't run it yet, I can't comment any further on the locations. Avoid abbreviations like midRostCaud.
|
I have pushed these changes. The mesh file I have added includes updated markers for where cranial nerves 3-12 emerge from the brainstem, which was not considered at the time of this original pull request. |
|
Sorry, I have pushed the wrong version up. Please don't review this one yet, it won't work! |
|
This should work for you now. However, cranial nerve emergent markers are not included as they are computed in "brainstem_coordinates" which are not defined on this mesh. |
There was a problem hiding this comment.
Some feedback. We need to have a discussion/handover.
I'd like to get some idea of the sizes and shapes in different species.
Is the midbrain normally longer than that?
Can we make annotation groups for the caudal end where it attaches to the spinal cord, and the cranial end?
| eIndexPM['caudal-dorsal'] = int(elemPerLayer/2) | ||
| eIndexPM['midRostralCaudal-dorsal'] = int(elemPerLayer / 2) + (elemPerLayer * int((elementsCountAlong/2)-1)) | ||
| eIndexPM['rostral-dorsal'] = (elemPerLayer*(elementsCountAlong-1)) + int(elemPerLayer/2) | ||
| eIndexPM['caudal-ventral'] = int(elemPerLayer/2) - (elementsCountAcrossMinor-1) | ||
| eIndexPM['midRostralCaudal-ventral'] = eIndexPM['midRostralCaudal-dorsal'] - (elementsCountAcrossMinor-1) | ||
| eIndexPM['rostral-ventral'] = int((elemPerLayer*(elementsCountAlong-1)) + (int(elemPerLayer/2) - elementsCountAcrossMinor + 1)) | ||
| xiPM['caudal-ventral'] = [1.0, 0.0, 0.0] | ||
| xiPM['caudal-dorsal'] = [1.0, 0.0, 1.0] | ||
| xiPM['midRostralCaudal-ventral'] = [1.0, 1.0, 0.0] | ||
| xiPM['midRostralCaudal-dorsal'] = [1.0, 1.0, 1.0] | ||
| xiPM['rostral-ventral'] = [1.0, 1.0, 0.0] | ||
| xiPM['rostral-dorsal'] = [1.0, 1.0, 1.0] | ||
| for key in eIndexPM.keys(): |
There was a problem hiding this comment.
We're making the marker points as annotationGroups now. This makes the code cleaner as there isn't the risk of not matching by name, however it means we must request ILX terms from K-core; mainly requires putting the name and description on a Google spreadsheet.
These names are not sufficient however, as they must be qualified with 'Brainstem', 'midline' and 'point' to be fully prescriptive. Is 'rostral' or 'cranial' more appropriate here -- we need terms that make sense across species.
The names should be plain english, uncapitalised without camelCase.
I don't think we need triple qualified rostral+caudal+ventral/dorsal as only 2 should be sufficient (along with midline).
Wouldn't the midRostralCaudal* points be better on the medulla-pons boundary?
Alternatively to many of these points, we could define a midline group around the boundary...
There was a problem hiding this comment.
I am unsure what you want here. I think it's probably better to talk to the next mapper to make these changes.
No description provided.