nemoarchive / analytics

Repository for the NeMO Analytics project.
MIT License
1 stars 0 forks source link

Create Patchseq profile for Flagship paper (Fig4) #138

Closed brianherb closed 3 years ago

brianherb commented 4 years ago

Patchseq -

@jorvis - the attached IT_test.zip file contains two files - a svg (IT_test.svg) and a data matrix (ITexpression.csv). The goal is to color the rectangles at the bottom of the image like a heatmap reflecting the expression from the data matrix.

IT_test.zip

image

jorvis commented 4 years ago

Can you upload the actual datafile to the devel version of NeMO? Then I can test the SVG on there.

brianherb commented 4 years ago

@jorvis - just checking, this site: https://gear-devel.nemoanalytics.org/ ?

brianherb commented 4 years ago

@jorvis - take a look at dataset 'Flagship Patchseq IT Test' - 403cf3ea-c5cb-7add-3142-a96945152f63

jorvis commented 4 years ago

Like this? (currently on production)

Screenshot from 2020-10-09 15-02-08

brianherb commented 4 years ago

OOOOOOOOOH! So does that mean it worked as is? That is fabulous!

jorvis commented 4 years ago

Yep, all good. I only had to modify our SVG processor to look for 'rect' elements to color in addition to the usual path and circle elements.

carlocolantuoni commented 4 years ago

gorgeous guys

On Fri, Oct 9, 2020 at 4:10 PM Joshua Orvis notifications@github.com wrote:

Yep, all good.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/nemoarchive/analytics/issues/138#issuecomment-706381069, or unsubscribe https://github.com/notifications/unsubscribe-auth/AH7KC7XFOEZIWNJKLYEA4M3SJ5U4TANCNFSM4SHRKKOA .

-- Carlo

brianherb commented 4 years ago

This is how I'm thinking of grouping neuron types - some have two rows, but each will be their own 'window'

Lamp5 image

Sncg / Vip image

Sst image

Pvalb image

IT image

ET image

carlocolantuoni commented 4 years ago

wow - very cool - after u and joshua implement all this, maybe you could pick a time to show us all how this works live along with the other flagship profile element you've built (like the MERfish etc). great to see it coming together

On Tue, Oct 13, 2020 at 11:43 AM Brian Herb notifications@github.com wrote:

This is how I'm thinking of grouping neuron types - some have two rows, but each will be their own 'window'

Lamp5 [image: image] https://user-images.githubusercontent.com/13540148/95883381-dbb7e300-0d48-11eb-8826-5043e9f2ea57.png

Sncg / Vip [image: image] https://user-images.githubusercontent.com/13540148/95883029-66e4a900-0d48-11eb-8902-e5486b5aacae.png

Sst [image: image] https://user-images.githubusercontent.com/13540148/95883150-9693b100-0d48-11eb-87e4-4c50b8506ef6.png

Pvalb [image: image] https://user-images.githubusercontent.com/13540148/95883275-c04cd800-0d48-11eb-870b-9cceda5c1a9f.png

IT [image: image] https://user-images.githubusercontent.com/13540148/95883474-f8541b00-0d48-11eb-9837-9028c4b7b4a8.png

ET [image: image] https://user-images.githubusercontent.com/13540148/95883543-11f56280-0d49-11eb-8485-79f481d74bce.png

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/nemoarchive/analytics/issues/138#issuecomment-707830729, or unsubscribe https://github.com/notifications/unsubscribe-auth/AH7KC7TT2BR3VAHD6R7XRS3SKRYQ7ANCNFSM4SHRKKOA .

-- Carlo

brianherb commented 4 years ago

New datasets on nemoanalytics:
(celltype - dataset name - unique id)

IT - Flagship Patchseq IT - 097fca87-8679-054b-d223-46ba4618f624

Lamp5 - Flagship Patchseq Lamp5 - 98db4be8-4a23-6a19-70ce-75444c79b6ed

Pvalb - Flagship Patchseq Pvalb - c0f77e66-ab22-6e4a-642a-c41f1ec6b9a3

ET - Flagship Patchseq ET - 5a8a7aae-248e-206b-d239-5b584afe8978

Sncg/Vip - Flagship Patchseq Sncg_Vip - 15b97730-aee5-d142-b6d3-ca1bd3c79197

Sst - Flagship Patchseq Sst - f973b4f0-ed0a-9ddb-c6c7-2fa5e0817796

brianherb commented 4 years ago

A new profile named 'MOp Patchseq' has been created here: https://www.nemoanalytics.org/p?l=1206f7ce

Unfortunately when all panels are together they do not render properly. If there is just one panel in the profile, it looks fine, but putting two or more causes problems.

brianherb commented 4 years ago

image

jorvis commented 4 years ago

Yes, mixing a lot of wide images with different aspect ratios is a problem.

brianherb commented 3 years ago

Fixed issue by reducing the size of the background image.