Linking flow cytometry data against CellMarker

Let us now consider a flow cytometry example dataset:

import lamindb as ln
import lnschema_bionty as bt

ln.track()

ℹ️ Instance: testuser1/mydata

ℹ️ User: testuser2

ℹ️ Added notebook: Transform(id='OWuTtS4SApon', v='0', name='12-flow', type=notebook, title='Linking flow cytometry data against `CellMarker`', created_by='bKeW4T6E', created_at=datetime.datetime(2023, 3, 30, 23, 17, 53))

ℹ️ Added run: Run(id='kXZegnhfZcQNtvYGY6C3', transform_id='OWuTtS4SApon', transform_v='0', created_by='bKeW4T6E', created_at=datetime.datetime(2023, 3, 30, 23, 17, 53))

Show code cell content Hide code cell content

filepath = ln.dev.datasets.file_fcs()

filepath

PosixPath('example.fcs')

Configure a CellMarker reference for parsing features

Because the file is a standard .fcs file, ln.File - under the hood - can parse it using readfcs.

Alternatively, we can load it into memory: AnnData = readfcs.read_fcs(filepath).

We’ll use the CellMarker-ontology based reference to link features:

reference = bt.CellMarker(species="human")

Parse features

features = ln.Features(filepath, reference)

✅ 14 terms (87.5%) are mapped.

🔶 2 terms (12.5%) are not mapped.

features

Features(id='IV2WkLDyi2WNBMvLrtiG', type='cell_marker', created_by='bKeW4T6E')

features.cell_markers[:3]

[CellMarker(id='CM_CD19', name='CD19', ncbi_gene_id='930', gene_symbol='CD19', gene_name='CD19 molecule', uniprotkb_id='P15391', species_id='NCBI_9606'),
 CellMarker(id='CM_CD14', name='CD14', ncbi_gene_id='4695', gene_symbol='CD14', gene_name='NADH:ubiquinone oxidoreductase subunit A2', uniprotkb_id='O43678', species_id='NCBI_9606'),
 CellMarker(id='CM_CCR7', name='CCR7', ncbi_gene_id='1236', gene_symbol='CCR7', gene_name='C-C motif chemokine receptor 7', uniprotkb_id='P32248', species_id='NCBI_9606')]

Track data with features (cell markers)

See also

Basic queries:

• Query data

• Query book

file = ln.File(filepath, features=features);

ln.add(file)

File(id='tsjjl9Tl837wKMUmxkXC', name='example', suffix='.fcs', size=6785467, hash='KCEXRahJ-Ui9Y6nksQ8z1A', source_id='kXZegnhfZcQNtvYGY6C3', storage_id='8Pj12JLb', created_at=datetime.datetime(2023, 3, 30, 23, 17, 57))

Querying data by features

We can now query datasets by cell markers:

files = (
    ln.select(ln.File)
    .join(ln.File.features)
    .join(ln.Features.cell_markers)
    .where(bt.CellMarker.gene_symbol == "CD14")
)
files.df().head()

	name	suffix	size	hash	source_id	storage_id	created_at	updated_at
id
tsjjl9Tl837wKMUmxkXC	example	.fcs	6785467	KCEXRahJ-Ui9Y6nksQ8z1A	kXZegnhfZcQNtvYGY6C3	8Pj12JLb	2023-03-30 23:17:57	None