search

rmarkello / abagen

A toolbox for working with Allen Human Brain Atlas microarray expression data
https://abagen.readthedocs.io
BSD 3-Clause "New" or "Revised" License
92 stars 41 forks source link

Integrate improved MNI-space registration #148

Open gdevenyi opened 4 years ago

gdevenyi commented 4 years ago

I "published" an improved MNI-space registration based on Chris' original work: https://github.com/gdevenyi/AllenHumanGeneMNI

Mostly the registrations tried harder, and I managed to fix the subject with the terrible cerebellum.

rmarkello commented 4 years ago

This looks awesome, thank you!! I'll try to integrate it into the codebase soon as an additional option for the parameter that controls which coordinates are used.

Just to confirm: I assume your new coordinates are in the same order as the original coordinates? You've dropped the well IDs from the final CSV files but it looks like they're identical in the TAG files. Just want to ~double~ triple confirm because some of them are quite dissimilar (in Euclidean distance) to the ones from Chris and co.

gdevenyi commented 4 years ago

Sorry, I should improve the docs a bit, the csvs here: https://github.com/gdevenyi/AllenHumanGeneMNI/tree/master/transformed-points/recombine the final files, which do have the well ids.

rmarkello commented 4 years ago

Oh, beautiful! I thought those were the original because of their formatting; clearly I didn't look closely enough.

This is great, thanks so much @gdevenyi !

gdevenyi commented 4 years ago

I saw from your presentation that there are surface-registered data available, my blue-sky would be to integrate my registrations for the subcortical/cerebellar with suface-based registration for the cortex, since I know I can't get good registrations there.