Talking with Steve McCarroll led to the thought that it might be cool to be able to directly extract count data for "metacells"... that is, pseudobulk the data in various ways. For example, you might want to add together counts of all cells in certain conditions. Like all cells from the same (dataset, donor, cell_type)... sum their counts and return one "metacell" per (dataset, donor, cell_type).
This doesn't have to be a cellarium-cloud feature though. We could just use a normal extract and then do post-processing in cellarium-ml (some kind of onepass) to aggregate to the meta-cell level. We were just wondering if it could be done nicely in BigQuery.
Talking with Steve McCarroll led to the thought that it might be cool to be able to directly extract count data for "metacells"... that is, pseudobulk the data in various ways. For example, you might want to add together counts of all cells in certain conditions. Like all cells from the same (dataset, donor, cell_type)... sum their counts and return one "metacell" per (dataset, donor, cell_type).
This doesn't have to be a cellarium-cloud feature though. We could just use a normal extract and then do post-processing in cellarium-ml (some kind of onepass) to aggregate to the meta-cell level. We were just wondering if it could be done nicely in BigQuery.