CellTypist 参考标签¶
公开免疫参考集给每个细胞一个标签，作为人工命名的线索。最终以 marker 为准。骨髓中性粒常被标成单核。
In [1]:
import time
import numpy as np
import scanpy as sc
from session_info2 import session_info
import celltypist
from celltypist import models
/home/han/software/miniconda3/envs/celltypist/lib/python3.12/site-packages/celltypist/classifier.py:11: FutureWarning: `__version__` is deprecated, use `importlib.metadata.version('scanpy')` instead
from scanpy import __version__ as scv
1. Download Reference¶
下载公开免疫参考模型。
In [2]:
# models.download_models(force_update=True)
In [3]:
model_high = models.Model.load(model="Immune_All_High.pkl")
model_low = models.Model.load(model="Immune_All_Low.pkl")
In [4]:
model_high.cell_types
Out[4]:
array(['B cells', 'B-cell lineage', 'Cycling cells', 'DC', 'DC precursor',
'Double-negative thymocytes', 'Double-positive thymocytes', 'ETP',
'Early MK', 'Endothelial cells', 'Epithelial cells',
'Erythrocytes', 'Erythroid', 'Fibroblasts', 'Granulocytes',
'HSC/MPP', 'ILC', 'ILC precursor', 'MNP', 'Macrophages',
'Mast cells', 'Megakaryocyte precursor',
'Megakaryocytes/platelets', 'Mono-mac', 'Monocyte precursor',
'Monocytes', 'Myelocytes', 'Plasma cells', 'Promyelocytes',
'T cells', 'pDC', 'pDC precursor'], dtype=object)
In [5]:
model_low.cell_types
Out[5]:
array(['Age-associated B cells', 'Alveolar macrophages', 'B cells',
'CD16+ NK cells', 'CD16- NK cells', 'CD8a/a', 'CD8a/b(entry)',
'CMP', 'CRTAM+ gamma-delta T cells', 'Classical monocytes',
'Cycling B cells', 'Cycling DCs', 'Cycling NK cells',
'Cycling T cells', 'Cycling gamma-delta T cells',
'Cycling monocytes', 'DC', 'DC precursor', 'DC1', 'DC2', 'DC3',
'Double-negative thymocytes', 'Double-positive thymocytes', 'ELP',
'ETP', 'Early MK', 'Early erythroid', 'Early lymphoid/T lymphoid',
'Endothelial cells', 'Epithelial cells', 'Erythrocytes',
'Erythrophagocytic macrophages', 'Fibroblasts',
'Follicular B cells', 'Follicular helper T cells', 'GMP',
'Germinal center B cells', 'Granulocytes', 'HSC/MPP',
'Hofbauer cells', 'ILC', 'ILC precursor', 'ILC1', 'ILC2', 'ILC3',
'Intermediate macrophages', 'Intestinal macrophages',
'Kidney-resident macrophages', 'Kupffer cells',
'Large pre-B cells', 'Late erythroid', 'MAIT cells', 'MEMP', 'MNP',
'Macrophages', 'Mast cells', 'Megakaryocyte precursor',
'Megakaryocyte-erythroid-mast cell progenitor',
'Megakaryocytes/platelets', 'Memory B cells',
'Memory CD4+ cytotoxic T cells', 'Mid erythroid', 'Migratory DCs',
'Mono-mac', 'Monocyte precursor', 'Monocytes', 'Myelocytes',
'NK cells', 'NKT cells', 'Naive B cells',
'Neutrophil-myeloid progenitor', 'Neutrophils',
'Non-classical monocytes', 'Plasma cells', 'Plasmablasts',
'Pre-pro-B cells', 'Pro-B cells',
'Proliferative germinal center B cells', 'Promyelocytes',
'Regulatory T cells', 'Small pre-B cells', 'T(agonist)',
'Tcm/Naive cytotoxic T cells', 'Tcm/Naive helper T cells',
'Tem/Effector helper T cells', 'Tem/Effector helper T cells PD1+',
'Tem/Temra cytotoxic T cells', 'Tem/Trm cytotoxic T cells',
'Transitional B cells', 'Transitional DC', 'Transitional NK',
'Treg(diff)', 'Trm cytotoxic T cells', 'Type 1 helper T cells',
'Type 17 helper T cells', 'gamma-delta T cells', 'pDC',
'pDC precursor'], dtype=object)
2. Load Data and Check Normalization¶
读入上一篇导出的 h5ad，确认已经是归一化表达。
In [6]:
adata = sc.read("../celltypist/seurat_object.h5ad")
adata.obsm["X_umap"] = np.asarray(adata.obsm["UMAP_HARMONY"])
In [7]:
adata.X[1000:1010, 500:510].toarray()
Out[7]:
array([[0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],
[2., 0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],
[1., 0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 3., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],
[0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],
[1., 0., 0., 0., 0., 0., 0., 0., 0., 0.]])
In [8]:
adata.X.sum(axis=1)
Out[8]:
matrix([[6292.],
[ 999.],
[2523.],
...,
[ 696.],
[6616.],
[6425.]], shape=(148682, 1))
In [9]:
adata.X.expm1().sum(axis=1)
Out[9]:
matrix([[1.27089863e+062],
[3.58504247e+009],
[6.07603023e+037],
...,
[2.90489192e+013],
[3.61033306e+161],
[1.47285655e+121]], shape=(148682, 1))
In [10]:
sc.pp.normalize_total(adata, target_sum=1e4)
In [11]:
adata.X.sum(axis=1)
Out[11]:
matrix([[10000.],
[10000.],
[10000.],
...,
[10000.],
[10000.],
[10000.]], shape=(148682, 1))
In [12]:
adata.X.expm1().sum(axis=1)
Out[12]:
matrix([[5.04999958e+098],
[4.36944669e+095],
[5.71107545e+149],
...,
[2.72747399e+193],
[1.55661597e+244],
[3.87772117e+188]], shape=(148682, 1))
In [13]:
sc.pp.log1p(adata)
3. Prediction¶
给每个细胞一个参考标签和置信度。置信度低的不要单独拿去写机制。
In [14]:
t_start = time.time()
predictions_low = celltypist.annotate(
adata, model="Immune_All_Low.pkl", majority_voting=True
)
t_end = time.time()
print(f"Time elapsed: {t_end - t_start} seconds")
🔬 Input data has 148682 cells and 38606 genes 🔗 Matching reference genes in the model 🧬 5824 features used for prediction ⚖️ Scaling input data 🖋️ Predicting labels ✅ Prediction done! 👀 Can not detect a neighborhood graph, will construct one before the over-clustering /home/han/software/miniconda3/envs/celltypist/lib/python3.12/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html from .autonotebook import tqdm as notebook_tqdm ⛓️ Over-clustering input data with resolution set to 25 🗳️ Majority voting the predictions ✅ Majority voting done!
Time elapsed: 216.93211460113525 seconds
In [15]:
predictions_low = predictions_low.to_adata()
predictions_low
Out[15]:
AnnData object with n_obs × n_vars = 148682 × 38606
obs: 'orig.ident', 'nCount_RNA', 'nFeature_RNA', 'group', 'sample', 'percent_mt', 'percent_rb', 'percent_hb', 'percent_pt', 's_score', 'g2m_score', 'phase', 'cc_diff', 'is_doublet', 'doublet_score', 'harmony_clusters_0.01', 'harmony_clusters_0.05', 'harmony_clusters_0.1', 'harmony_clusters_0.2', 'harmony_clusters_0.5', 'harmony_clusters_1', 'ident', 'predicted_labels', 'over_clustering', 'majority_voting', 'conf_score'
uns: 'X_name', 'log1p', 'neighbors', 'over_clustering'
obsm: 'HARMONY', 'PCA', 'UMAP_HARMONY', 'UMAP_UNINT', 'X_umap', 'X_pca'
obsp: 'connectivities', 'distances'
layers: None (.X)
In [16]:
sc.pl.umap(
predictions_low, color=["majority_voting"], legend_loc="on data", size=1, alpha=.5,
)
In [17]:
t_start = time.time()
predictions_high = celltypist.annotate(
adata, model="Immune_All_High.pkl", majority_voting=True
)
t_end = time.time()
print(f"Time elapsed: {t_end - t_start} seconds")
🔬 Input data has 148682 cells and 38606 genes 🔗 Matching reference genes in the model 🧬 5824 features used for prediction ⚖️ Scaling input data 🖋️ Predicting labels ✅ Prediction done! 👀 Detected a neighborhood graph in the input object, will run over-clustering on the basis of it ⛓️ Over-clustering input data with resolution set to 25 🗳️ Majority voting the predictions ✅ Majority voting done!
Time elapsed: 10.706525087356567 seconds
In [18]:
predictions_high = predictions_high.to_adata()
predictions_high
Out[18]:
AnnData object with n_obs × n_vars = 148682 × 38606
obs: 'orig.ident', 'nCount_RNA', 'nFeature_RNA', 'group', 'sample', 'percent_mt', 'percent_rb', 'percent_hb', 'percent_pt', 's_score', 'g2m_score', 'phase', 'cc_diff', 'is_doublet', 'doublet_score', 'harmony_clusters_0.01', 'harmony_clusters_0.05', 'harmony_clusters_0.1', 'harmony_clusters_0.2', 'harmony_clusters_0.5', 'harmony_clusters_1', 'ident', 'predicted_labels', 'majority_voting', 'conf_score', 'over_clustering'
uns: 'X_name', 'log1p', 'neighbors', 'over_clustering', 'majority_voting_colors'
obsm: 'HARMONY', 'PCA', 'UMAP_HARMONY', 'UMAP_UNINT', 'X_umap', 'X_pca'
obsp: 'connectivities', 'distances'
layers: None (.X)
In [19]:
sc.pl.umap(
predictions_high,
color=["majority_voting"],
legend_loc="on data",
size=1,
alpha=0.5,
)
4. Save Results¶
把标签写回，供 04 对照。
In [20]:
predictions_low.obs.loc[
:, ["predicted_labels", "over_clustering", "majority_voting", "conf_score"]
].to_csv("../celltypist/predictions_low.tsv", sep="\t")
predictions_high.obs.loc[
:, ["predicted_labels", "over_clustering", "majority_voting", "conf_score"]
].to_csv("../celltypist/predictions_high.tsv", sep="\t")
In [21]:
print(session_info())
anndata 0.13.3.post0 numpy 2.5.3 scanpy 1.12.4 celltypist 1.7.1 ---- ---- Python 3.12.14 (main, Sep 2 2026, 23:27:36) [GCC 15.3.0] OS Linux-6.17.0-22-generic-x86_64-with-glibc2.39 Updated 2026-09-14 09:26