import HERBS probe data in alignment gui

[Andrianarivelo]

Hi,
I exported NPX 2.0 probes data (as pkl files, one for each shank), how can I format them to a format compatible with the ephys alignment-gui ?
Thanks

[jsiegle]

This fork of the IBL GUI uses the same input format as the original, which is documented here: https://github.com/int-brain-lab/iblapps/wiki/4.-Preparing-data-for-ephys-GUI

We have created a script for exporting ephys data into the compatible format when starting from a SpikeInterface sorting/recording objects, which you may find helpful: https://github.com/AllenNeuralDynamics/aind-ephys-ibl-alignment-gui-converter/blob/main/code/run_capsule.py

[Aexolowski]

I have a similar question, but want to know how to format the histological data from HERBS (the pickle files, one for each shank) to be able to use them with the IBL ephys alignment tool.

[jsiegle]

This is probably a better question to ask in the HERBS repo. Hopefully it's not too difficult to figure out how to .pkl files to the IBL-GUI-specific JSONs.

[Ic3-Fury]

Following up, we ran the code on our xyz_picks file to convert from CCF (mlapdv) --> Bregma (mlapdv) and it output a very small number - in the thousandths or the ten thousandths decimal place (xyz_picks_bregma.json). When comparing it to the sample xyz_picks file provided by IBL (xyz_picks.json) we see that the numbers in this json file are in the thousands, not in decimals.

Furthermore, we ran the sample code as is listed on the page:

`from iblatlas import atlas
import numpy as np

res = 25
brain_atlas = atlas.AllenAtlas(res)

ccf_mlapdv = np.array([[3000, 4000, 3000], [6000, 6000, 500] ], dtype=float)
bregma_mlapdv = brain_atlas.ccf2xyz(ccf_mlapdv, ccf_order='mlapdv')`

And we received the output coordinates: [-0.002739, 0.0014 , -0.002668], [ 0.000261, -0.0006 , -0.000168]. So we are unsure how the xyz_picks sample file relates to these decimal values.

[jsiegle]

I believe the AssertionError is due to a problem with the file naming, as it's not finding exactly one xyz_picks.json file in the data directory. Based on the screenshot, it may be because the .json extension is missing.

The xyz_picks file should be in units of microns, so the original scaling looks correct.

Looping in @mayofaulkner, who may be better suited to address some of these questions.

[Ic3-Fury]

Hello! Really appreciate you looping in Mayo, I think that would be great help here. We already tried renaming file "xyz_picks.json" and checked that it was a .json in it's properties...same error.

[mayofaulkner]

Hello,

The output from the ccf2xyz conversion is in metres but you will need to convert to um ( x 1e6) and then write to the json file. This should explain the difference in scales.

Is this single shank or multi shank data? If it is single shank then the file match is '*xyz_picks.json', if multi shank then you will need to provide one xyz_picks per shank and name them e.g *xyz_picks_shank1.json *xyz_picks_shank2.json etc

[Ic3-Fury]

Hi Mayo, the numbers now make a lot more sense knowing that they're in meters. We are using NPX 2.0 probes, so I'll work on making the files for each shank and naming them as you have suggested. I'll get back to you here as soon as I finish that and seeing if it works! Hopefully this should resolve the issues we have been having. Again, we really appreciate both yours and Josh's time and help!