Closed jordandekraker closed 2 years ago
The latest commit now uses brainspace to calculate affinity, gradients and align them to the reference gradient in one step. Currently correlation matrices and gradients are calculated in a single snakemake rule, would you think it would be worthwhile to split them into separate rules or just keep them as is?
I think its fine, so you can avoid having to save and then reload the exact same data. Especially if you make good error logging :)
https://github.com/myousif9/hippograd/blob/cd29d0777ac0d563b3ac54cd249c0ba6130dcfd0/hippograd/workflow/rules/post_hippunfold.smk#L21
calcualte correlation matrix here, and use that as the input to Brainspace (it will know what to do!)