Query & integrate data¶
import lamindb as ln
import bionty as bt
→ connected lamindb: testuser1/test-facs
ln.context.uid = "wukchS8V976U0000"
ln.context.track()
→ notebook imports: bionty==0.49.1 lamindb==0.76.4
→ created Transform('wukchS8V976U0000') & created Run('2024-09-06 08:30:51.815520+00:00')
Inspect the CellMarker registry 
¶
Inspect your aggregated cell marker registry as a DataFrame:
bt.CellMarker.df().head()
| uid | name | synonyms | description | gene_symbol | ncbi_gene_id | uniprotkb_id | source_id | organism_id | run_id | created_by_id | updated_at | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| id | ||||||||||||
| 41 | 3ZFziy5ims8J | CD14/19 | None | None | None | None | None | NaN | 1 | 2 | 1 | 2024-09-06 08:30:45.153391+00:00 |
| 40 | 31nZfqQo8yZg | CD103 | None | ITGAE | 3682 | P38570 | 28.0 | 1 | 2 | 1 | 2024-09-06 08:30:45.143810+00:00 | |
| 39 | 1iLDs6cZIpxj | CD69 | None | CD69 | 969 | Q07108 | 28.0 | 1 | 2 | 1 | 2024-09-06 08:30:45.143767+00:00 | |
| 38 | 525YfNUB967z | CD49B | None | ITGA2 | 3673 | P17301 | 28.0 | 1 | 2 | 1 | 2024-09-06 08:30:45.143730+00:00 | |
| 37 | 3IPMBjs68Vy1 | CXCR4 | None | CXCR4 | 7852 | P61073 | 28.0 | 1 | 2 | 1 | 2024-09-06 08:30:45.143692+00:00 |
Search for a marker (synonyms aware):
bt.CellMarker.search("PD-1").df().head(2)
| uid | name | synonyms | description | gene_symbol | ncbi_gene_id | uniprotkb_id | source_id | organism_id | run_id | created_by_id | updated_at | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| id | ||||||||||||
| 29 | 33vFR1q26vnM | PD1 | PID1|PD-1|PD 1 | None | PDCD1 | 5133 | A0A0M3M0G7 | 28 | 1 | 1 | 1 | 2024-09-06 08:30:24.538763+00:00 |
Look up markers with auto-complete:
markers = bt.CellMarker.lookup()
markers.cd8
CellMarker(uid='1xRpnOHIkdyE', name='CD8', synonyms='', gene_symbol='CD8A', ncbi_gene_id='925', uniprotkb_id='P01732', created_by_id=1, run_id=1, source_id=28, organism_id=1, updated_at='2024-09-06 08:30:24 UTC')
Query artifacts by markers ¶
Query panels and collections based on markers, e.g., which collections have 'CD8' in the flow panel:
panels_with_cd8 = ln.FeatureSet.filter(cell_markers=markers.cd8).all()
ln.Artifact.filter(feature_sets__in=panels_with_cd8).df()
| uid | version | is_latest | description | key | suffix | type | size | hash | n_objects | n_observations | _hash_type | _accessor | visibility | _key_is_virtual | storage_id | transform_id | run_id | created_by_id | updated_at | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| id | ||||||||||||||||||||
| 1 | qGXQ3oXRQrrfYSAf0000 | None | True | Alpert19 | None | .h5ad | dataset | 33374864 | QNP1c3p6scaAwPo9AW8fLw | None | 166537 | md5 | AnnData | 1 | True | 1 | 1 | 1 | 1 | 2024-09-06 08:30:34.881573+00:00 |
| 2 | cWmcWgk8zKjnt0Sj0000 | None | True | Oetjen18_t1 | None | .h5ad | dataset | 46506448 | WbPHGIMM_5GT68rC8ZydHA | None | 241552 | md5 | AnnData | 1 | True | 1 | 2 | 2 | 1 | 2024-09-06 08:30:45.686627+00:00 |
Access registries:
features = ln.Feature.lookup()
Find shared cell markers between two files:
artifacts = ln.Artifact.filter(feature_sets__in=panels_with_cd8).list()
shared_markers = artifacts[0].features["var"] & artifacts[1].features["var"]
shared_markers.list("name")
['Cd4', 'CD8', 'CD3', 'CD27', 'Ccr7', 'CD45RA']