search

DailyDreaming / load-project

1 stars 0 forks source link

Reconcile differences between methods to generate cell counts #113

Open theathorn opened 4 years ago

theathorn commented 4 years ago

Lon's old method vs current method.

hannes-ucsc commented 4 years ago

Use .attic/cell_counts.json and make report to determine if there actually are discrepancies.

GPelayo commented 4 years ago

This is a table of all the discrepancies between .attic/cell_counts.json and make report.

Accession ID .attic/cell_counts.json make report
GSE44183 48 47
GSE75659 1318 1316
GSE99795 147 0

@hannes-ucsc Also, GSE95435 has counts in .attic/cell_counts.json but has no spreadsheet associated with it in this repo. GSE95435 only appears in the geo_series_accessions field of spreadsheets/existing/GSE114374_project.json

hannes-ucsc commented 4 years ago

The off-by-ones are fixed in #142. I looked at the geo files in GSE99795 and hesitate to agree with @DailyDreaming's call to count transcripts as cells. We'll stick with our call not to consider that project's matrix. That leaves GSE95435 to resolve.