Multi-modal¶
Here, we’ll showcase how to curate and register ECCITE-seq data from Papalexi21 in the form of MuData objects.
ECCITE-seq is designed to enable interrogation of single-cell transcriptomes together with surface protein markers in the context of CRISPR screens.
MuData objects build on top of AnnData objects to store multimodal data.
# !pip install 'lamindb[jupyter,bionty]'
!lamin init --storage ./test-multimodal --modules bionty
Show code cell output
→ initialized lamindb: testuser1/test-multimodal
import lamindb as ln
import bionty as bt
ln.track()
Show code cell output
→ connected lamindb: testuser1/test-multimodal
→ created Transform('5MvP6WeoMC430000'), started new Run('QKewR8L6...') at 2025-04-15 16:37:36 UTC
→ notebook imports: bionty==1.3.0 lamindb==1.4.0
Creating MuData Artifacts¶
lamindb provides a from_mudata() method to create Artifact from MuData objects.
mdata = ln.core.datasets.mudata_papalexi21_subset()
mdata
Show code cell output
MuData object with n_obs × n_vars = 200 × 300
obs: 'perturbation', 'replicate'
var: 'name'
4 modalities
rna: 200 x 173
obs: 'nCount_RNA', 'nFeature_RNA', 'percent.mito'
var: 'name'
adt: 200 x 4
obs: 'nCount_ADT', 'nFeature_ADT'
var: 'name'
hto: 200 x 12
obs: 'nCount_HTO', 'nFeature_HTO', 'technique'
var: 'name'
gdo: 200 x 111
obs: 'nCount_GDO'
var: 'name'mdata_af = ln.Artifact.from_mudata(mdata, key="papalexi.h5mu")
mdata_af
Show code cell output
Artifact(uid='XQhZiiz1zBns4oK90000', is_latest=True, key='papalexi.h5mu', suffix='.h5mu', kind='dataset', otype='MuData', size=549984, hash='aFIJ7G9AIcxoEib8kecChw', n_observations=200, space_id=1, storage_id=1, run_id=1, created_by_id=1, created_at=<django.db.models.expressions.DatabaseDefault object at 0x7f94c8e00770>)
# MuData Artifacts have the corresponding otype
mdata_af.otype
Show code cell output
'MuData'
# MuData Artifacts can easily be loaded back into memory
papalexi_in_memory = mdata_af.load()
papalexi_in_memory
Show code cell output
MuData object with n_obs × n_vars = 200 × 300
obs: 'perturbation', 'replicate'
var: 'name'
4 modalities
rna: 200 x 173
obs: 'nCount_RNA', 'nFeature_RNA', 'percent.mito'
var: 'name'
adt: 200 x 4
obs: 'nCount_ADT', 'nFeature_ADT'
var: 'name'
hto: 200 x 12
obs: 'nCount_HTO', 'nFeature_HTO', 'technique'
var: 'name'
gdo: 200 x 111
obs: 'nCount_GDO'
var: 'name'Schema¶
# define labels
perturbation = ln.ULabel(name="Perturbation", is_type=True).save()
ln.ULabel(name="Perturbed", type=perturbation).save()
ln.ULabel(name="NT", type=perturbation).save()
replicate = ln.ULabel(name="Replicate", is_type=True).save()
ln.ULabel(name="rep1", type=replicate).save()
ln.ULabel(name="rep2", type=replicate).save()
ln.ULabel(name="rep3", type=replicate).save()
# define obs schema
obs_schema = ln.Schema(
name="mudata_papalexi21_subset_obs_schema",
features=[
ln.Feature(name="perturbation", dtype="cat[ULabel[Perturbation]]").save(),
ln.Feature(name="replicate", dtype="cat[ULabel[Replicate]]").save(),
],
).save()
obs_schema_rna = ln.Schema(
name="mudata_papalexi21_subset_rna_obs_schema",
features=[
ln.Feature(name="nCount_RNA", dtype=int).save(),
ln.Feature(name="nFeature_RNA", dtype=int).save(),
ln.Feature(name="percent.mito", dtype=float).save(),
],
coerce_dtype=True,
).save()
obs_schema_hto = ln.Schema(
name="mudata_papalexi21_subset_hto_obs_schema",
features=[
ln.Feature(name="nCount_HTO", dtype=float).save(),
ln.Feature(name="nFeature_HTO", dtype=int).save(),
ln.Feature(name="technique", dtype=bt.ExperimentalFactor).save(),
],
coerce_dtype=True,
).save()
var_schema_rna = ln.Schema(
name="mudata_papalexi21_subset_rna_var_schema",
itype=bt.Gene.symbol,
dtype=float,
).save()
# define composite schema
mudata_schema = ln.Schema(
name="mudata_papalexi21_subset_mudata_schema",
otype="MuData",
components={
"obs": obs_schema,
"rna:obs": obs_schema_rna,
"hto:obs": obs_schema_hto,
"rna:var": var_schema_rna,
},
).save()
Show code cell output
! record with similar name exists! did you mean to load it?
| uid | name | is_type | description | reference | reference_type | space_id | type_id | run_id | created_at | created_by_id | _aux | _branch_code | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| id | |||||||||||||
| 1 | s0GowB3a | Perturbation | True | None | None | None | 1 | None | 1 | 2025-04-15 16:37:38.041000+00:00 | 1 | None | 1 |
! record with similar name exists! did you mean to load it?
| uid | name | dtype | is_type | unit | description | array_rank | array_size | array_shape | proxy_dtype | synonyms | _expect_many | _curation | space_id | type_id | run_id | created_at | created_by_id | _aux | _branch_code | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| id | ||||||||||||||||||||
| 4 | RJIRkiuFrFkv | nFeature_RNA | int | None | None | None | 0 | 0 | None | None | None | True | None | 1 | None | 1 | 2025-04-15 16:37:38.103000+00:00 | 1 | {'af': {'0': None, '1': True, '2': False}} | 1 |
mudata_schema
Schema(uid='uBmIjPJrzVhKoOZ9II4p', name='mudata_papalexi21_subset_mudata_schema', n=-1, itype='Composite', is_type=False, otype='MuData', dtype='num', hash='muhs8kiT-fTL0SufVA-pOQ', minimal_set=True, ordered_set=False, maximal_set=False, space_id=1, created_by_id=1, run_id=1, created_at=2025-04-15 16:37:38 UTC)
Validate MuData annotations¶
curator = ln.curators.MuDataCurator(mdata, mudata_schema)
try:
curator.validate()
except ln.errors.ValidationError:
pass
! using default organism = human
! using default organism = human
! using default organism = human
! 96 terms are not validated: 'RP5-827C21.6', 'XX-CR54.1', 'RP11-379B18.5', 'RP11-778D9.12', 'RP11-703G6.1', 'AC005150.1', 'RP11-717H13.1', 'CTC-498J12.1', 'CTC-467M3.1', 'HIST1H4K', 'RP11-524H19.2', 'AC006042.7', 'AC002066.1', 'AC073934.6', 'RP11-268G12.1', 'U52111.14', 'RP11-235C23.5', 'RP11-12J10.3', 'CASC1', 'RP11-324E6.9', ...
12 synonyms found: "CTC-467M3.1" → "MEF2C-AS2", "HIST1H4K" → "H4C12", "CASC1" → "DNAI7", "LARGE" → "LARGE1", "NBPF16" → "NBPF15", "C1orf65" → "CCDC185", "IBA57-AS1" → "IBA57-DT", "KIAA1239" → "NWD2", "TMEM75" → "LINC02912", "AP003419.16" → "RPS6KB2-AS1", "FAM65C" → "RIPOR3", "C14orf177" → "LINC02914"
→ curate synonyms via .standardize("columns")
for remaining terms:
→ fix typos, remove non-existent values, or save terms via .add_new_from("columns")
curator.slots["rna:var"].cat.standardize("columns")
curator.slots["rna:var"].cat.add_new_from("columns")
curator.validate()
Show code cell output
! using default organism = human
! using default organism = human
! using default organism = human
Register curated Artifact¶
artifact = curator.save_artifact(key="mudata_papalexi21_subset.h5mu")
Show code cell output
! using default organism = human
! using default organism = human
! using default organism = human
→ returning existing schema with same hash: Schema(uid='tQt1swwzScOyxegHtaPa', name='mudata_papalexi21_subset_obs_schema', n=2, itype='Feature', is_type=False, hash='s_pyvTup4kDNqfE0Uda-RQ', minimal_set=True, ordered_set=False, maximal_set=False, space_id=1, created_by_id=1, run_id=1, created_at=2025-04-15 16:37:38 UTC)
! using default organism = human
! 12 unique terms (6.90%) are not validated for symbol: 'CTC-467M3.1', 'HIST1H4K', 'CASC1', 'LARGE', 'NBPF16', 'C1orf65', 'IBA57-AS1', 'KIAA1239', 'TMEM75', 'AP003419.16', ...
! using default organism = human
→ returning existing schema with same hash: Schema(uid='H4ZIxaJLpBYrlabb2j1I', name='mudata_papalexi21_subset_rna_obs_schema', n=3, itype='Feature', is_type=False, hash='P7Ov2yyw84ZWPjhvoarMvA', minimal_set=True, ordered_set=False, maximal_set=False, space_id=1, created_by_id=1, run_id=1, created_at=2025-04-15 16:37:38 UTC)
→ returning existing schema with same hash: Schema(uid='AIKI1xc8kvgEz3AYTGHH', name='mudata_papalexi21_subset_hto_obs_schema', n=3, itype='Feature', is_type=False, hash='O1BhyIsVj6eIX62YDnrhtA', minimal_set=True, ordered_set=False, maximal_set=False, space_id=1, created_by_id=1, run_id=1, created_at=2025-04-15 16:37:38 UTC)
artifact.describe()
Show code cell output
Artifact .h5mu/MuData ├── General │ ├── .uid = 'HOk9sjYwgYsKU9JZ0000' │ ├── .key = 'mudata_papalexi21_subset.h5mu' │ ├── .size = 549984 │ ├── .hash = 'aFIJ7G9AIcxoEib8kecChw' │ ├── .n_observations = 200 │ ├── .path = │ │ /home/runner/work/lamin-usecases/lamin-usecases/docs/test-multimodal/.lamindb/HOk9sjYwgYsKU9JZ0000.h5mu │ ├── .created_by = testuser1 (Test User1) │ ├── .created_at = 2025-04-15 16:37:42 │ └── .transform = 'Multi-modal' ├── Dataset features │ ├── obs • 2 [Feature] │ │ perturbation cat[ULabel[Perturbation]] NT, Perturbed │ │ replicate cat[ULabel[Replicate]] rep1, rep2, rep3 │ ├── ['rna'].var • 172 [bionty.Gene] │ │ SH2D6 float │ │ ARHGAP26-AS1 float │ │ GABRA1 float │ │ HLA-DQB1-AS1 float │ │ HLA-DQB1-AS1 float │ │ HLA-DQB1-AS1 float │ │ HLA-DQB1-AS1 float │ │ HLA-DQB1-AS1 float │ │ HLA-DQB1-AS1 float │ │ HLA-DQB1-AS1 float │ │ SPACA1 float │ │ VNN1 float │ │ CTAGE15 float │ │ CTAGE15 float │ │ PFKFB1 float │ │ TRPC5 float │ │ RBPMS-AS1 float │ │ CA8 float │ │ CSMD3 float │ │ ZNF483 float │ ├── ['rna'].obs • 3 [Feature] │ │ nCount_RNA int │ │ nFeature_RNA int │ │ percent.mito float │ └── ['hto'].obs • 3 [Feature] │ technique cat[bionty.ExperimentalF… cell hashing │ nCount_HTO float │ nFeature_HTO int └── Labels └── .experimental_factors bionty.ExperimentalFactor cell hashing .ulabels ULabel Perturbed, NT, rep1, rep2, rep3
ln.finish()
Show code cell output
! cells [(15, 17)] were not run consecutively
→ finished Run('QKewR8L6') after 8s at 2025-04-15 16:37:44 UTC
# clean up test instance
!rm -r test-multimodal
!lamin delete --force test-multimodal
Show code cell output
• deleting instance testuser1/test-multimodal