The Allen Brain Cell Atlas (ABC Atlas) has been updated with new data!
Data from Transcriptomic diversity of cell types across the adult human brain is now explorable in the ABC Atlas. This dataset includes more than three million cells sampled from the adult human brain. Samples were isolated from ~100 dissections from three donors and assayed using single-nucleus RNA sequencing. The resulting cells were clustered into hierarchical groups of 31 superclusters, 461 clusters, and 3313 subclusters. Additionally, categorical neurotransmitter type annotations were assigned to clusters based on expression (or lack thereof) of one or more neurotransmitter-associated marker genes. Included in this dataset includes two tSNE plots - one plot contains approximately 900k non-neuronal cells and the other contains approximately 2.5m neuronal cells.
I was able to download the 20241115 version of WHB-10Xv3 dataset’s cell_metadata file at AWS S3 Explorer
It looks that the metadata file listed all 3 million plus cells, but did have a column for cell type annotation. I guess that the “cluster_alias” column can be mapped to cell type annotation.
Is there another version of the cell_metadata file having cell type annotation, or another file having cluster_alias to cell type annotation mapping?
You can take a look at this tutorial jupyter notebook: 10x RNA-seq gene expression data (part 1) — Allen Brain Cell Atlas - Data Access This is walking through joining the cell metadata tables that you’ve found with the cell type annotations and walking through some basic extraction of gene expression data. You are correct that cluster_alias is the column that joins the cells and taxonomy/celltype annotations.
If you are comfortable with python, there is the classes in this repository that allow for more programmatic access to the data on S3 vs browsing the s3 bucket on the web. The repository also contains all of the tutorial notebooks we have which you can download and modify to your needs.
I’m trying to access the WHB-10Xv3 20241115 dataset. Is it structured the same as the WMB dataset, so the notebook tutorial will work if I simply change the dataset reference?
If you have a cluster-alias to cell type mapping file handy, can you simply share it?
Thank you for your detailed tutorial and support. I was able to retrieve my genes of interest from the WHB-10Xv3 20241115 dataset.
Since the dataset contains over 3 million cells, I’m looking to filter for subsets based on QC metrics and homogeneity in each cell types. Is there mapping data available for each cell, such as the number of reads mapped and the number of genes detected?
Hi Sean, could you post this question as a new thread? Your question and the answer may be helpful for other folks right now it may be lost under this original announcement post where people may not expect to look. Thanks!