s5_vis.py

import sys
import scanpy as sc

adata=sc.read_h5ad('d6_GSE202051_celltype.h5ad')
sys.path.insert(0, '/lustre/home/jydeng/single-cell/codes')
from util import *

sc.set_figure_params(dpi=200, figsize=(10,10), facecolor="white")

marker_dict = get_marker_gene_dict()
for ct in marker_dict.keys():        
    sc.tl.score_genes(adata, marker_dict[ct], score_name='score_' + ct)

sc.pl.umap(adata, color=['score_' + ct for ct in marker_dict.keys()], cmap='Reds', ncols=3, show=False)

save_fig_to_path('scores_for_clusters.png')

# adata.obs.filter()

'''
Seems not a good marker:
GCG
Seems a good marker:
B cell: MS4A1, BLK
'''
manual_ann=dict()
annotate(manual_ann, 'Acinar', ['18', '24']) # PRSS1, REG3A, AMY2B, REG1B, CTRB1, CTRB2
annotate(manual_ann, 'B cell', 37) #MS4A1, BANK1, ABCB4 (part)
annotate(manual_ann, 'T cell', 9) # CD69, PTPRC, CD2, IL7R, CD247
annotate(manual_ann, 'Endothelial', [2, 31, 34, 36])
annotate(manual_ann, 'Epithelial', [0, 16, 20, 22, 11, 17, 21, 25, 28, 54, 38, 42]) # ELF3, SLC15A1, KRT19, ABCA14, MACC1, EPCAM, CDH1 看不清！
annotate(manual_ann, 'Smooth muscle', [12, 19, 27]) #RGS5
#不知道怎么分：ductal
annotate(manual_ann, 'Myeloid', 3) #ITGAX, ITGAM, CD163
annotate(manual_ann, 'Schwann', 35) #CDH19, PMP22

adata.obs["manual_ann"] = adata.obs['leiden_r2'].map(manual_ann)
sc.pl.umap(adata, color=["manual_ann"], legend_fontsize="small", legend_loc="on data", show=False, save='manual_ann.png')

fil = manual_ann.keys()
unassigned = []
for i in adata.obs['leiden_r2'].unique():
    if i not in fil:
            unassigned.append(i)

# plot
sc.pl.umap(adata, color='leiden_r2', groups=unassigned, show=False, save='Unassigned.png') #!!!!
['5', '14', '1', '32', '8', '7', '30', '6', '10', '4', '26', '15', '23', '41', '13', '39', '29', '40', '33']
# 5和epi完全重叠，重新画一下marker
# 右下角也重叠，可以重新做

for ct in unassigned:
    sc.pl.umap(adata, color='leiden_r2', show=False, legend_loc='on data', legend_fontsize='small', groups=[ct], save=f'_{ct}.png')

adata = adata[adata.obs['leiden_r2'].isin(['5', '14', '1', '32', '8', '7', '30', '6', '10', '4', '26', '15', '23', '41', '13', '39', '29', '40', '33'])]

subset = adata[adata.obs['leiden_r2'].isin(['26', '30', '32', '33'])]
sc.pp.neighbors(subset)
sc.tl.umap(subset)
sc.pl.umap(subset, show=False, color='leiden_r2', save='_upper.png', legend_loc='on data', s=5)
subset.write_h5ad('subset1_upper.h5ad') # 9k? obs

subset = adata[adata.obs['leiden_r2'].isin(['1', '4', '7', '8',' 15', '23', '39'])]
sc.pp.neighbors(subset)
sc.tl.umap(subset)
sc.pl.umap(subset, show=False, color='leiden_r2', save='_right.png', legend_loc='on data', s=5)
subset.write_h5ad('subset2_right.h5ad') # 50261 obs

subset = adata[adata.obs['leiden_r2'].isin(['40', '29', '41', '14', '15', '6', '13', '10', '5'])] # 48561
sc.pp.neighbors(subset)
sc.tl.umap(subset)
sc.pl.umap(subset, show=False, color='leiden_r2', save='_left.png', legend_loc='on data', s=6)
subset.write_h5ad('subset3_left.h5ad') # 48561 obs

# the upper part should be endocrine
# let's see the marker genes
marker_dict = {
    #  'Pericyte': ['CSPG4', 'ACTA2', 'CD146', 'NES', 'PDGFRb'], # CD146, PDGFRb不在
     'Beta cell': ['ADCYAP1', 'NPTX2', 'NPY', 'MAFA', 'LIFR', 'CDKN1C', 'NKX6-1', 'BMPR1A', 'ITGA1', 'MNX1', 'IAPP', 'TFCP2L1', 'PDX1', 'PIR', 'SIX3', 'TGFBR3', 'INS', 'DLK1', 'SYT13', 'BMP5', 'SMAD9'], 
     'Beta-like cell': ['CD9'], 
     'Endocrine cell': ['CHGA', 'GCG', 'CHGB', 'PPY', 'INS', 'ISL1', 'SST', 'IAPP'], 
     'Alpha cell': ['GCG', 'ARX'],
     'Delta cell': ['SST', 'PPY'],
     'Non-beta endocrine cell': ['GCG', 'PPY', 'PDK4', 'SOD2', 'SST'],
     'Pancreatic polypeptide cell': ['SPP1', 'FOLR1', 'PPA1']
}

# label the upper part
import scanpy as sc
adata=sc.read_h5ad('subset1_upper.h5ad')
for ct in marker_dict:
    sc.tl.score_genes(adata, marker_dict[ct], score_name='score_' + ct)
sc.pl.umap(adata, color=['score_' + ct for ct in marker_dict.keys()], cmap='Reds', ncols=3, show=False, save='_upper_endocrine.png')
annotate(manual_ann, 'Endocrine', ['26', '30', '32', '33', '40'])
for ct in marker_dict.keys():
    fig = sc.pl.umap(
        adata, color=marker_dict[ct], vmin=0, vmax="p99", sort_order=False, frameon=False, cmap="Reds", show=False, save=f'_genes_{ct}.png'
    )
layer2_annotation = dict()
annotate(layer2_annotation, 'Alpha cell', 30)
annotate(layer2_annotation, 'Beta cell', 26)
annotate(layer2_annotation, 'Delta cell', 32) # 32 还可以分成两个
cluster33 = ['LINC00907', 'GRM7-AS3', 'USH2A', 'KCNB2', 'RIMBP2', 'TMEM132D'] # 看不懂这是啥

adata=sc.read_h5ad('subset2_right.h5ad')


marker_dict = {
     'Acinar cell': ['CTRB1', 'CTRB2', 'GCG', 'PPY', 'AMY2B', 'REG1B', 'PDK4', 'SOD2', 'PRSS1', 'SST', 'REG3A'], 'Alpha cell': ['GCG', 'ARX'], 'Antigen presentation cancer-associated fibroblast': ['SAA3P', 'CD74'], 'B cell': ['ABCB4', 'REG1A', 'AIDA', 'CD52', 'CD79B', 'TTR', 'AFF4', 'ABCB9', 'CD19', 'IAPP', 'PPY', 'INS', 'SST', 'IRF8', 'BLK', 'MS4A1', 'AIM2', 'GCG', 'CD79A', 'BANK1'], 'Basal cell': ['ARL13B', 'COL4A1'], 'Beta cell': ['ADCYAP1', 'NPTX2', 'NPY', 'MAFA', 'LIFR', 'CDKN1C', 'NKX6-1', 'BMPR1A', 'ITGA1', 'MNX1', 'IAPP', 'TFCP2L1', 'PDX1', 'PIR', 'SIX3', 'TGFBR3', 'INS', 'DLK1', 'SYT13', 'BMP5', 'SMAD9'], 'Beta-like cell': ['CD9'], 'Cancer cell': ['MYC', 'KRAS', 'TP53', 'KARS1'], 'Delta cell': ['SST', 'PPY'], 'Dendritic cell': ['CD1C', 'FCER1A', 'CD1A', 'CD34'], 'Ductal cell': ['MUC6', 'ANXA4', 'PDX1', 'AMBP', 'SCTR', 'BICC1', 'KRT8', 'CFTR', 'SOX9', 'HNF1B', 'BMPR1A', 'MMP7', 'KRT19', 'KRT7', 'SLC4A4'], 'Ductal epithelial cell': ['CLDN1', 'S100A14'], 'Emt-like cell': ['ZEB1', 'SNAI2', 'CDH2'], 'Endocrine cell': ['CHGA', 'GCG', 'CHGB', 'PPY', 'INS', 'ISL1', 'SST', 'IAPP'], 'Endothelial cell': ['SELE', 'CDH5', 'VWF', 'CD36', 'FLT1', 'ADGRL2', 'ICAM1', 'CLDN5', 'ECSCR', 'VEGFA', 'PLVAP', 'PECAM1', 'ACYP1', 'B2M', 'SLCO2A1', 'KDR'], 'Epithelial cell': ['SLC15A1', 'KRT19', 'DSC2', 'ANKRD30A', 'CA2', 'ADGB', 'CDH1', 'ELF3', 'EPCAM', 'BACE2', 'TJP2', 'MUC1', 'MACC1', 'SFTPB', 'SFTPC', 'ABCB10', 'MYLK', 'ABCA13'], 'Epithelial ductal cell': ['KRT19', 'CDH1', 'AQP5', 'SOX9'], 'Fibroblast': ['NECTIN1', 'SPARC', 'THY1', 'FN1', 'ACTA2', 'LRP1', 'COL1A1', 'LUM', 'C5AR2', 'DCN', 'FAP', 'COL3A1', 'PTPN13'], 'Follicular b cell': ['CR2', 'FCER2', 'CD19', 'CD22'], 'Gamma cell': ['PPY', 'SST'], 'Hematopoietic stem cell': ['CD34'], 'Immune cell': ['CD8A', 'PTPRC', 'ITGAM'], 'Inflammatory cancer-associated fibroblast': ['ACTA2', 'CXCL12', 'IL6'], 'Lymphatic endothelial cell': ['FLT4', 'PROX1', 'LYVE1', 'PDPN'], 'Lymphocyte': ['CD3D', 'CD3G', 'IL7R'], 'Macrophage': ['FCGR1A', 'CD86', 'CSF1R', 'HLA-DRB1', 'AIF1', 'CD68', 'HLA-DPA1', 'ZEB2', 'CSF2RA', 'CD74', 'HLA-DRA', 'CD14', 'PTPRC'], 'Mast cell': ['CMA1', 'CAVIN2', 'GCSAML', 'KIT', 'CAPN3', 'ASIC4', 'ADIRF', 'ENPP3', 'MAOB', 'SLC18A2', 'BACE2', 'MAML1', 'CADPS', 'CDK15'], 'Mesenchymal cell': ['INHBA', 'VIM', 'FAP', 'KDELR3', 'SERPINE2', 'GPC6'], 'Monocyte': ['VCAN', 'S100A9', 'CD68', 'FCN1', 'S100A8'], 'Monocyte derived dendritic cell': ['CLEC10A', 'CD1C', 'FCER1A', 'CD1E'], 'Monocyte-derived macrophage': ['CXCL8', 'CD68', 'APOE', 'CCL3', 'CD63', 'CCL2', 'MARCO'], 'Myeloid cell': ['FCER1G', 'TYROBP', 'CD68', 'ITGAX', 'AIF1', 'ITGAM', 'CD14', 'PTPRC', 'CD163'], 'Myofibroblastic cancer-associated fibroblast': ['ACTA2', 'TAGLN'], 'Natural killer t(nkt) cell': ['NCAM1'], 'Non-beta endocrine cell': ['GCG', 'PPY', 'PDK4', 'SOD2', 'SST'], 'Pancreatic polypeptide cell': ['SPP1', 'FOLR1', 'PPA1'], 'Pancreatic progenitor cell': ['NKX6-1'], 'Pancreatic stellate cell': ['COL1A2'], 'Pro-inflammatory macrophage': ['IL1B'], 'Progenitor cell': ['GP2', 'PDX1', 'SOX9', 'HNF1B', 'TTYH1', 'PTF1A'], 'Proliferative cell': ['MKI67'], 'Proliferative ductal cell': ['CDK1', 'AURKA', 'PLK1'], 'Schwann cell': ['PMP22', 'S100B', 'SCN7A', 'CDH19', 'CRYAB'], 'Smooth muscle cell': ['NOTCH3', 'CSRP2', 'RGS5'], 'Stellate cell': ['ADIRF', 'ACTA2', 'RGS5', 'PDGFRB', 'COL1A2'], 'T cell': ['CD8A', 'CD4', 'CD69', 'CD247', 'IL7R', 'HHLA2', 'CD2', 'CD3G', 'CXCR4', 'CD3D', 'PTPRC', 'CD3E'], 'Tissue resident macrophage': ['AMY2A', 'PRSS1', 'CELA3A', 'INS'], 'Undifferentiated pancreatic progenitor cell': ['SPP1'], 'White blood cell': ['PTPRC'], 'Astrocyte': ['ALDH1L1', 'ALDOC', 'SLC1A3', 'GFAP', 'AGT', 'AGXT2L1'], 'Erythrocyte': ['GYPA', 'ALAS2', 'HBB', 'HBE1', 'HBA1', 'HBG1', 'CA1'], 'Gmp': ['AREG', 'ANXA1', 'ASPM', 'CDKN3', 'KIT', 'CD38', 'CLSPN', 'MCM10', 'SDC4', 'RMI2', 'ADK', 'APPL1', 'AP3S1', 'CD123', 'ALDH4A1', 'MUCB2', 'APLP2', 'DEPDC7'], 'Hsc': ['ABCG2', 'PROM1', 'KIT', 'THY1', 'VCAM1', 'CD41', 'CMAH', 'ITGA5', 'ACE', 'ALDH1A1', 'BMI1', 'CD34', 'CD164'], 'Macro/mono/dc': ['BHLHE40', 'CREM', 'CSF1R', 'ADGRE2', 'CD1A', 'CD68', 'MRC1', 'CD93', 'CD209', 'ICAM4', 'CD14', 'ACPP', 'CCL18', 'ADGRE3', 'CD83', 'ACSL3'], 'Malignant cell': ['KRT8', 'KLK3', 'KRT18', 'EPCAM', 'FOLH1', 'KRT19'], 'Neuron': ['ENO2', 'MAP2', 'STMN2', 'DLG4', 'RBFOX3', 'TUBB3', 'CSF3'], 'Neutrophil': ['CXCL8', 'LGALS13', 'ADGRG3', 'AQP9', 'MNDA', 'FCGR3B', 'NFE4', 'IL5RA', 'CSF3R', 'G0S2', 'BTNL8', 'ANXA3', 'USP10'], 'Nk cell': ['KLRD1', 'KLRB1', 'NKG7', 'CD247', 'KLRK1', 'XCL1', 'XCL2', 'KLRC1', 'FCGR3A', 'NCR1', 'GZMB', 'NCR3'], 'Oligodendrocyte': ['MBP', 'OLIG2', 'MOG', 'MAG', 'PLP1', 'OLIG1', 'SOX10', 'PDGFRA'], 'Plasma cell': ['BRSK1', 'AC026202.3', 'PARM1', 'JSRP1', 'TAS1R3', 'LINC00582', 'MZB1'], 'Progenitor': ['FLT3', 'CD38', 'CD90', 'MME', 'CAR', 'CD123', 'ALDH', 'CASR', 'KDR'], 'Cdc1': ['CLEC9A'], 'Cdc2': ['CLEC10A']
}


# pancreatic stellate, mesenchymal, myofibroblast, CAF
marker_dict = {
     'Fibroblast': ['MFAP5', 'DCN', 'WNT5A', 'COL1A1', 'COL1A2', 'LUM', 'ACTA2', 'SPARC'],
     'CAF': ['PDGFRA', 'COL1A1', 'THY1', 'FAP', 'COL8A1', 'THBS2', 'CTHRC1', 'ENG', 'VIM','CDH11'], 
     'Myofibroblast': ['ACTA2', 'TAGLN', 'FAP', 'MMP11', 'POSTN', 'CTHRC1']
}
for ct in marker_dict:
    sc.tl.score_genes(adata, marker_dict[ct], score_name='score_' + ct)
sc.pl.umap(adata, color=['score_' + ct for ct in marker_dict.keys()], cmap='Reds', ncols=3, save='_right_fibro.png')
annotate(manual_ann, 'Fibroblast', ['1', '4', '7', '8',' 15', '23', '39'])

adata.obs["manual_ann"] = adata.obs['leiden_r2'].map(manual_ann)

extra_ann = {
     '40': 'Endocrine',
     '41': 'Schwann',
     '15': 'Fibroblast',
     '5': 'Epithelial',
     '6': 'Epithelial',
     '10': 'Epithelial',
     '13': 'Epithelial',
     '14': 'Epithelial',
     '29': 'Epithelial',
}
manual_ann.update(extra_ann)
adata.obs['manual_ann'] = adata.obs['leiden_r2'].map(manual_ann)