How to find MNI coordinates for gene expression zscore data

Technical
1

I am trying to pull data from the human gene microarray data and have successfully downloaded the the zscore data. The “Contents.txt” file that is downloaded with the data explains that the headers, “sample_mri_0, sample_mri_1, sample_mri_2 should be understood to be MRI coordinates of the sample in x,y,z order”. However the coordinates don’t seem to correspond to the MNI coordinate system. For example, a point identified in the structure “anterior orbital gyrus”, has coordinates 135, 102, 92, which is outside the MNI brain, assuming the 0,0,0 point is at the anterior commissure. Am I misunderstanding the coordinate system, or do I need to make some correction? When I visualize the data in BrainExplorer2, the coordinates look correct, so I feel like I am missig something.

2

Hi @axlander,

When you download the files from here here…
image
…you should have a zip file that includes multiple files. In the file called “SampleAnnot.csv” there are TWO sets of coordinates for each sample: the MRI coordinates that you mention, alongside the MNI coordinates. I’m not sure what the MRI coordinates represent, but this should allow immediate translation between both sets of coordinates. If you need more, please reply and I can ask around.

Best,
Jeremy

3

Hi Jeremy.

Thanks for your response. When I download the zip file, it has 4 files enclosed: Columns.csv, Contents.csv, Expression.csv and Probes.csv, but no SampleAnnot.csv. The Columns file has 17 columns including the MRI coordinates (sample_mri_0, sample_mri_1, sample_mri_2) as well as some columns of donor info, and others of brain regions. But no other coordinates.

Any ideas?

Thanks,

Alex

4

Hi Alex,

The files you need to download are the six files circled below. I’m not sure why the file you downloaded is missing the MNI coordinates, but these files have it. Unless you care about the microarray data itself (which you will also be downloading), the only thing you’ll need in the zip files is “SampleAnnot.csv”.

Best,
Jeremy

image
image827×788 33.6 KB

5

Hi Jeremy.

Thanks for the instructions for Dummies – I clearly need to work on my reading skills.

Alex

6

Hello @jeremyinseattle!
Could you specify which exact MNI space has been used for the AHBA?
I can’t seem to find any information on which MNI space the original MRI coordinates have been mapped to.

Thank you very much in advance!

Best,
Laura

7

Hello @ettchen,

A key author on this project confirmed that would be MNI152, referenced here: https://www.bic.mni.mcgill.ca/ServicesAtlases/ICBM152NLin2009

Best,
Jeremy

8

Thank you so much for the quick reply!!
Could he please specify which one exactly of the six templates from 2009?
There’s these six:

  • ICBM 2009a Nonlinear Symmetric
  • ICBM 2009a Nonlinear Asymmetric
  • ICBM 2009b Nonlinear Symmetric
  • ICBM 2009b Nonlinear Asymmetric
  • ICBM 2009c Nonlinear Symmetric
  • ICBM 2009c Nonlinear Asymmetric

Thank you again!

Best,
Laura

9

It is more than likely “ICBM 2009b Nonlinear Symmetric”, as that is what the underlying reference atlas is aligned to. More info is here: Atlases and Ontologies - Brain Cell Data Center (BCDC)

1 Like
10

Thanks again @jeremyinseattle !
I wasn’t sure if the gene expression is in the same space, as the atlas seemed to be more recent (Allen Human Reference Atlas - 3D, 2020 (new!)).
But I’m glad to have your confirmation now :slight_smile: