How to get the cluster to which a cell belongs, using AllenSDK?

121 Views Asked by At

On the "A Cellular Taxonomy of the Mouse Visual Cortex" page, "Explore the Data" section, the top row shows the list of cluster names of mouse visual cortex cells:

Allen cell type clusters

Question: Given a mouse visual cortex cell's specimen ID (e.g. 505808144), how do I find the cluster name to which that cell belongs?

I am able to get cell metadata via AllenSDK get_cell(id) method, but the returned metadata does not seem to contain the name of the cluster.

1

There are 1 best solutions below

2
On BEST ANSWER

The Cell Types Database contains two groups of data sets -- large sets of transcriptomics-only cells (see http://celltypes.brain-map.org/download), and one large set of cells with electrophysiology, some of which also have morphology (http://celltypes.brain-map.org/data). Detailed information on the experimental overview and methods used in these studies is available as whitepapers in the Documentation section (http://help.brain-map.org/display/celltypes/Documentation).

The example cell in your question has electrophysiology and morphology, but no transcriptomics, so it does not inherently belong to a transcriptomic cluster.

The AllenSDK currently supports downloading metadata for electrophysiology-morphology cells, but not the transcriptomics-only cells. You can download metadata these cells as CSV files. If you would like to be able to access transcriptomics metadata from the AllenSDK instead of flat CSV files, please create an issue on the AllenSDK issue tracker (https://github.com/alleninstitute/allensdk/issues).

Good news -- in a couple days we will be releasing cluster assignments on much of the data you see here (http://celltypes.brain-map.org/data), including the mouse visual cortex data set. This will include many more cells for some data sets. You will be able to download this data, including cluster assignments and metadata, as CSV files.

Note that mouse visual cortex data available within the Cell Types Database is a different set of cells (and profiling methodology) from the data in the Cellular Taxonomy science case study.