PBMC CITE-Seq Unbiased Feature Selection

[1]:

import scanpy as sc
import os
import pandas as pd
import numpy as np
import pickle as pkl
import matplotlib as mpl
import matplotlib.pyplot as plt
import scipy.stats

sc.settings.verbosity = 3

[2]:

rna_adata = sc.read("../../CITE-seq/rna.h5ad")

[3]:

data = pd.read_csv("../../CITE-seq/GSE100866_PBMC_vs_flow_10X-ADT_umi.csv", index_col=0)
data

[3]:
ACCGTAAGTGTAATGA CGTGAGCTCGAGAACG CACATTTAGAATTCCC TACGGTATCTGGGCCA TCAGGTAGTAAGTTCC TGATTTCGTTCTCATT ACACTGAAGGCCCTCA ACGGGTCGTCACACGC AGCTTGACATCCCATC ACGTCAATCCGTCATC ... TTCTCCTAGATCGATA GGAAAGCGTCGACTAT GACTAACCAACACCCG AGCGTCGTCCTCGCAT TCTCATAAGTTTGCGT GTCGGGTAGAGCTGGT GTCGGGTAGGTAGCCA GTCGGGTAGTCTTGCA ATGTGTGGTCCGTTAA CGTATGCCGTCTTCTG
CD3 9 17 18 12 18 16 15 23 18 9 ... 40 184 2 88 8 115 53 10 5 5
CD4 8 17 18 12 20 29 27 20 28 19 ... 93 157 15 47 6 112 69 14 7 5
CD8 13 21 16 16 18 25 11 23 21 10 ... 180 26 9 4 8 8 5 32 7 9
CD2 9 10 7 8 5 15 10 4 3 4 ... 38 184 1 50 2 245 87 21 3 7
CD45RA 56 37 36 53 74 69 46 44 75 41 ... 66 87 22 302 31 10 12 182 28 22
CD57 19 40 32 21 33 55 41 30 52 32 ... 15 15 24 7 15 9 3 10 19 7
CD16 15 11 4 7 14 30 20 22 25 8 ... 17 8 9 1 4 5 0 9 4 2
CD14 10 7 13 9 15 25 24 21 28 9 ... 152 6 8 3 4 3 2 24 6 3
CD11c 7 6 5 10 14 18 8 18 23 11 ... 203 7 7 3 2 4 1 29 5 1
CD19 15 14 12 19 17 22 17 12 18 4 ... 7 5 12 1 6 1 3 3 5 4

10 rows × 7985 columns

[4]:

adata = sc.AnnData(data.T)

[5]:

adata = adata[rna_adata.obs.index, ]

[6]:

sc.settings.set_figure_params(dpi=80, facecolor='white')
sc.pl.highest_expr_genes(adata, n_top=20)

normalizing counts per cell
    finished ({time_passed})

C:\Users\SLiang3\Miniconda3\envs\scanpy37\lib\site-packages\scanpy\preprocessing\_normalization.py:138: UserWarning: Revieved a view of an AnnData. Making a copy.
  view_to_actual(adata)
_images/pbmc-cite-seq_6_2.png 
[7]:

sc.pp.calculate_qc_metrics(adata, percent_top=None, log1p=False, inplace=True)
sc.pl.violin(adata, ['n_genes_by_counts', 'total_counts'], jitter=0.4, multi_panel=True)

C:\Users\SLiang3\Miniconda3\envs\scanpy37\lib\site-packages\seaborn\_core.py:1303: UserWarning: Vertical orientation ignored with only `x` specified.
  warnings.warn(single_var_warning.format("Vertical", "x"))
C:\Users\SLiang3\Miniconda3\envs\scanpy37\lib\site-packages\seaborn\_core.py:1303: UserWarning: Vertical orientation ignored with only `x` specified.
  warnings.warn(single_var_warning.format("Vertical", "x"))
C:\Users\SLiang3\Miniconda3\envs\scanpy37\lib\site-packages\seaborn\_core.py:1303: UserWarning: Vertical orientation ignored with only `x` specified.
  warnings.warn(single_var_warning.format("Vertical", "x"))
C:\Users\SLiang3\Miniconda3\envs\scanpy37\lib\site-packages\seaborn\_core.py:1303: UserWarning: Vertical orientation ignored with only `x` specified.
  warnings.warn(single_var_warning.format("Vertical", "x"))
_images/pbmc-cite-seq_7_1.png 
[8]:

sc.pl.scatter(adata, x='total_counts', y='n_genes_by_counts')
_images/pbmc-cite-seq_8_0.png 
[9]:

adata.var.highly_variable = True

[ ]:



[10]:

#sc.pp.normalize_total(adata, target_sum=1e4)
import scipy.stats.mstats
def clr(adata):
    temp = adata.X + 1
    adata.X = np.log(temp / scipy.stats.mstats.gmean(temp, axis=1).reshape([-1, 1]))

clr(adata)

[11]:

sc.pp.scale(adata, max_value=2)

[12]:

np.isnan(adata.X).sum()

[12]:

0

[13]:

sc.tl.pca(adata, svd_solver='arpack')

computing PCA
    with n_comps=9
    finished (0:00:00)

[14]:

sc.pl.pca(adata, color=data.index)
_images/pbmc-cite-seq_15_0.png 
[15]:

sc.pl.pca_variance_ratio(adata, log=True)
_images/pbmc-cite-seq_16_0.png 
[16]:

sc.pp.neighbors(adata, n_neighbors=50)
sc.tl.umap(adata)
sc.tl.leiden(adata, resolution=2.)

computing neighbors
    using data matrix X directly
    finished: added to `.uns['neighbors']`
    `.obsp['distances']`, distances for each pair of neighbors
    `.obsp['connectivities']`, weighted adjacency matrix (0:00:04)
computing UMAP
    finished: added
    'X_umap', UMAP coordinates (adata.obsm) (0:00:15)
running Leiden clustering
    finished: found 24 clusters and added
    'leiden', the cluster labels (adata.obs, categorical) (0:00:06)

[17]:

sc.pl.umap(adata, color=['leiden'], ncols=3, legend_loc="on data", legend_fontsize=8.)
_images/pbmc-cite-seq_18_0.png 
[18]:

sc.pl.umap(adata, color=data.index, ncols=3, legend_loc="on data", legend_fontsize=8.)
_images/pbmc-cite-seq_19_0.png 
[19]:

filtered_adata = adata.copy()
proteins = filtered_adata.var.index.to_series().apply(lambda x: 'PROTEIN_' + x).tolist()
rnas = rna_adata.raw.var.index.to_series().apply(lambda x: '' + x).tolist()
merged_adata = sc.AnnData(pd.DataFrame(np.hstack([filtered_adata.X, rna_adata.raw.X]),
                                    columns=proteins+rnas))
merged_adata.obs = filtered_adata.obs
merged_adata.obsm = filtered_adata.obsm

C:\Users\SLiang3\Miniconda3\envs\scanpy37\lib\site-packages\anndata\_core\anndata.py:119: ImplicitModificationWarning: Transforming to str index.
  warnings.warn("Transforming to str index.", ImplicitModificationWarning)

[20]:

#print(*merged_adata.var_names)

[21]:

sc.pl.umap(merged_adata, color=['leiden',
                         'GZMB', 'GZMA', 'NCAM1', 'CD244', # NK (Granzyme)
                         'CD3D', 'CD3E', 'CD3G', 'CD4', 'CD8A', 'CD8B', # T CD4/CD8
                         'CD14', 'FCGR3A', 'FCGR3B', # MONO CD14 / (CD16 = FCGR)
                         'CD19', 'PTPRC', 'CD79A', # B (CD45R = PTPRC)
                         'HLA-DRA', 'CD68', 'IL3RA', 'CLEC4C', 'NRP1',  # DC
                         'CXCL9', 'CXCL10', 'CD86'
                        ], ncols=3, legend_loc="on data", legend_fontsize=8.)
_images/pbmc-cite-seq_22_0.png

B cells:

Cluster 3, 13, and 23 are classified as B cells given high CD19 expression in protein, and high CD19 and CD79A in mRNA. Because cluster 23 also has high expression of CD3, and thus is annotated as “CD3+” B cells.

T cells:

Cluster 0, 1, 2, 8, 9, 12, 21, 22 (bottom left) and cluster 6, 11, 15, 16 are clearly high in CD3, and are all T-cells. Cluster 0, 1, 2, 8, 9, 12 are CD4 T cells; Cluster 6, 11, 16 are CD8 T cells; Cluster 21, 22 are CD4+ CD8+ (double positive) T cells, labeled as DP T; CD22 is also high in CD57 Cluster 15 are CD4- CD8- (double negative) T cells, labeled as DN T;

NK cells:

Cluster 7, 9, 10, 18 are high in CD16, GZMA/B, NCAM1 (CD56), and CD244. Cluster 18 is high in CD57.

Monocytes:

Cluster 4, 5, 14, 19, 20, 17 are high in CD11c and CD68 Cluster 4, 5, 14, 19 are high in CD14 Cluster 14 are high in CD8

[22]:

cluster2subtype = {'0': 'CD45RA+ CD4 T',
                '1': 'CD4 T',
                '2': 'CD4 T',
                '3': 'B',
                '4': 'CD14+ CD16+ Mono',
                '5': 'CD14+ Mono',
                '6': 'CD8 T',
                '7': 'NK',
                '8': 'CD45RA+ CD4 T',
                '9': 'NK',
                '10': 'NK',
                '11': 'CD8 T',
                '12': 'CD45RA+ CD4 T',
                '13': 'B',
                '14': 'CD3+ CD8+ CD14+ Mono',
                '15': 'DN T',
                '16': 'CD57+ CD8 T',
                '17': 'CD16+ Mono',
                '18': 'CD57+ NK',
                '19': 'CD3+ CD14+ Mono',
                '20': 'CD14- CD16- Mono',
                '21': 'DP T',
                '22': 'CD57+ DP T',
                '23': 'CD3+ B'}

adata.obs['cell subtype'] = [cluster2subtype[i] for i in adata.obs.leiden]
filtered_adata.obs['cell subtype'] = [cluster2subtype[i] for i in adata.obs.leiden]
merged_adata.obs['cell subtype'] = [cluster2subtype[i] for i in adata.obs.leiden]

cluster2type = {'0': 'CD4 T',
                '1': 'CD4 T',
                '2': 'CD4 T',
                '3': 'B',
                '4': 'Mono',
                '5': 'Mono',
                '6': 'CD8 T',
                '7': 'NK',
                '8': 'CD4 T',
                '9': 'NK',
                '10': 'NK',
                '11': 'CD8 T',
                '12': 'CD4 T',
                '13': 'B',
                '14': 'Mono',
                '15': 'DN T',
                '16': 'CD8 T',
                '17': 'Mono',
                '18': 'NK',
                '19': 'Mono',
                '20': 'Mono',
                '21': 'DP T',
                '22': 'DP T',
                '23': 'B'}

adata.obs['cell type'] = [cluster2type[i] for i in adata.obs.leiden]
filtered_adata.obs['cell type'] = [cluster2type[i] for i in adata.obs.leiden]
merged_adata.obs['cell type'] = [cluster2type[i] for i in adata.obs.leiden]

[23]:

protein2rna = {'PROTEIN_CD2': ['CD2'],
               'PROTEIN_CD3': ['CD3D', 'CD3E', 'CD3G'],
               'PROTEIN_CD4': ['CD4'],
               'PROTEIN_CD8': ['CD8A', 'CD8B'],
               'PROTEIN_CD11c': ['ITGAX'],
               'PROTEIN_CD14': ['CD14'],
               'PROTEIN_CD16': ['FCGR3A', 'FCGR3B'],
               'PROTEIN_CD19': ['CD19'],
               'PROTEIN_CD45RA': ['PTPRC'],
               'PROTEIN_CD57': ['B3GAT1']
              }
for k, v in protein2rna.items():
    sc.pl.umap(merged_adata, color=[k] + v, ncols=4, legend_fontsize=8.)

C:\Users\SLiang3\Miniconda3\envs\scanpy37\lib\site-packages\anndata\_core\anndata.py:1192: FutureWarning: is_categorical is deprecated and will be removed in a future version.  Use is_categorical_dtype instead
  if is_string_dtype(df[key]) and not is_categorical(df[key])
... storing 'cell subtype' as categorical
... storing 'cell type' as categorical
_images/pbmc-cite-seq_25_1.png
_images/pbmc-cite-seq_25_2.png
_images/pbmc-cite-seq_25_3.png
_images/pbmc-cite-seq_25_4.png
_images/pbmc-cite-seq_25_5.png
_images/pbmc-cite-seq_25_6.png
_images/pbmc-cite-seq_25_7.png
_images/pbmc-cite-seq_25_8.png
_images/pbmc-cite-seq_25_9.png
_images/pbmc-cite-seq_25_10.png 
[24]:

sc.pl.umap(adata, color=['cell type', 'cell subtype'], ncols=4, legend_fontsize=8.)

... storing 'cell subtype' as categorical
... storing 'cell type' as categorical
_images/pbmc-cite-seq_26_1.png 
[25]:

#import sys
#sys.path.insert(0,'..')
#import compactmarker._tsne_l1
#np.random.seed(0)
#model = compactmarker._tsne_l1.TsneL1.tune(50, rna_adata.X, filtered_adata.X) # , max_inner_iter=1, max_outer_iter=1

[26]:

#np.random.seed(0)
#model = compactmarker._tsne_l1.TsneL1.tune(50, rna_adata.X, filtered_adata.X,
#                                           min_lasso=1.2409377607517196e-05, max_lasso=1.539926526059492e-05)
#import torch
#torch.autograd.set_detect_anomaly(True)

[39]:

import sys
sys.path.insert(0,'..')
import compactmarker

model = compactmarker.UmapL1(lasso=1e-3, ridge=1e-3, n_pcs=None, perplexity=100., use_beta_in_Q=True, n_threads=6,
                             max_outer_iter=2) # , max_inner_iter=1, max_outer_iter=1
model.fit(rna_adata.X, X_teacher=filtered_adata.X)

Calculating distance matrix and scaling factors...
Computing pairwise distances...
Using 6 threads...
Mean value of sigma: 0.362250
Done. Elapsed time: 61.17 seconds. Total: 61.17 seconds.
Creating model without batches...
Optimizing using OWLQN (because lasso is nonzero)...
0 loss: 38.25368118286133 Nonzero: 69 Elapsed time: 472.48 seconds. Total: 533.65 seconds.
1 loss: 3.821220874786377 Nonzero: 69 Elapsed time: 278.84 seconds. Total: 812.49 seconds.
final loss: 3.8009145259857178 Nonzero: 69 Elapsed time: 6.67 seconds. Total: 819.16 seconds.

[39]:

<compactmarker._umap_l1.UmapL1 at 0x1c99fc46a08>

[45]:

import sys
sys.path.insert(0,'..')
import compactmarker

model = compactmarker.UmapL1(lasso=5e-4, ridge=0., n_pcs=None, perplexity=100., use_beta_in_Q=True, n_threads=6,
                             max_outer_iter=2) # , max_inner_iter=1, max_outer_iter=1
model.fit(rna_adata.X, X_teacher=filtered_adata.X)

Calculating distance matrix and scaling factors...
Computing pairwise distances...
Using 6 threads...
Mean value of sigma: 0.362250
Done. Elapsed time: 56.94 seconds. Total: 56.94 seconds.
Creating model without batches...
Optimizing using OWLQN (because lasso is nonzero)...
0 loss: 12.995180130004883 Nonzero: 27 Elapsed time: 468.43 seconds. Total: 525.37 seconds.
1 loss: 3.8472976684570312 Nonzero: 26 Elapsed time: 461.75 seconds. Total: 987.12 seconds.
final loss: 3.816372871398926 Nonzero: 26 Elapsed time: 6.67 seconds. Total: 993.79 seconds.

[45]:

<compactmarker._umap_l1.UmapL1 at 0x1c99fdf4088>

[46]:

print(*rna_adata.var_names[model.get_mask()])

AIF1 CD7 CD74 CD79A CD8B CST7 CTSS FCER1G GNLY HLA-DPA1 HLA-DRA HLA-DRB1 IL32 LEF1 LGALS1 LST1 LYZ NKG7 PRF1 RPL13 S100A11 S100A4 S100A8 S100A9 TRBC2 TYROBP

[47]:

rnas = rna_adata.raw.var.index.to_series().apply(lambda x: '' + x).tolist()
merged_adata2 = rna_adata.copy()

protein_markers = ['PROTEIN_' + i for i in filtered_adata.var.index]

for i, marker in enumerate(filtered_adata.var.index.tolist()):
    merged_adata2.obs['PROTEIN_' + marker] = filtered_adata.X[:, i]

[48]:

sc.pl.pca(adata, color='cell type')
_images/pbmc-cite-seq_33_0.png 
[51]:

new_adata = model.transform(merged_adata2.copy())
new_adata.obs['cell type'] = filtered_adata.obs['cell type']
new_adata.obs['cell subtype'] = filtered_adata.obs['cell subtype']

sc.tl.pca(new_adata, svd_solver='arpack')
sc.pl.pca(new_adata, color='cell type', frameon=False)
sc.pp.neighbors(new_adata, n_pcs=5, use_rep="X_pca")
sc.tl.umap(new_adata)

sc.pl.umap(new_adata, color=['cell type'], frameon=False, title="", size=3.)

C:\Users\SLiang3\Miniconda3\envs\scanpy37\lib\site-packages\anndata\_core\anndata.py:1094: FutureWarning: is_categorical is deprecated and will be removed in a future version.  Use is_categorical_dtype instead
  if not is_categorical(df_full[k]):
Trying to set attribute `.obs` of view, copying.

computing PCA
    on highly variable genes
    with n_comps=24
    finished (0:00:00)
_images/pbmc-cite-seq_34_2.png 
computing neighbors
    finished: added to `.uns['neighbors']`
    `.obsp['distances']`, distances for each pair of neighbors
    `.obsp['connectivities']`, weighted adjacency matrix (0:00:01)
computing UMAP
    finished: added
    'X_umap', UMAP coordinates (adata.obsm) (0:00:09)
_images/pbmc-cite-seq_34_4.png 
[ ]:



[32]:

new_adata

[32]:

AnnData object with n_obs × n_vars = 7634 × 48
    obs: 'cell subtype', 'cell type', 'n_genes', 'n_genes_by_counts', 'total_counts', 'total_counts_mt', 'pct_counts_mt', 'leiden', 'PROTEIN_CD3', 'PROTEIN_CD4', 'PROTEIN_CD8', 'PROTEIN_CD2', 'PROTEIN_CD45RA', 'PROTEIN_CD57', 'PROTEIN_CD16', 'PROTEIN_CD14', 'PROTEIN_CD11c', 'PROTEIN_CD19'
    var: 'n_cells', 'mt', 'n_cells_by_counts', 'mean_counts', 'pct_dropout_by_counts', 'total_counts', 'highly_variable', 'means', 'dispersions', 'dispersions_norm', 'mean', 'std'
    uns: 'cell subtype_colors', 'cell type_colors', 'hvg', 'leiden', 'leiden_colors', 'neighbors', 'pca', 'umap'
    obsm: 'X_pca', 'X_umap'
    varm: 'PCs'
    obsp: 'connectivities', 'distances'

[33]:

import pandas as pd
import seaborn as sn
import matplotlib.pyplot as plt

df = pd.DataFrame(new_adata.X, columns=new_adata.var.index)

corrMatrix = df.corr()

fig, ax = plt.subplots(figsize=(14, 12))
sn.heatmap(ax=ax, data=corrMatrix, xticklabels=True, yticklabels=True)
plt.show()

corr_matrix = corrMatrix.values
corr_matrix = np.abs(corr_matrix)

corr_values = []

for i in range(corr_matrix.shape[0]):
    for j in range(i):
        corr_values.append(corr_matrix[i, j])

plt.hist(corr_values, bins=50)
plt.title('Distribution of correlations of \n selected markers')
_images/pbmc-cite-seq_37_0.png 
[33]:

Text(0.5, 1.0, 'Distribution of correlations of \n selected markers')
_images/pbmc-cite-seq_37_2.png 
[34]:

import pandas as pd
import seaborn as sn
import matplotlib.pyplot as plt

df2 = pd.concat([df.reset_index(drop=True),
                new_adata.obs.loc[:, new_adata.obs.columns.str.contains('PROTEIN_')].reset_index(drop=True)],
               axis=1)

corrMatrix = df2.corr()
corrMatrix = corrMatrix.loc[corrMatrix.index.str.contains('PROTEIN_'), ~corrMatrix.index.str.contains('PROTEIN_')]

fig, ax = plt.subplots(figsize=(20, 3))
sn.heatmap(ax=ax, data=corrMatrix.abs(), xticklabels=True, yticklabels=True)
plt.show()
_images/pbmc-cite-seq_38_0.png 
[35]:

corrMatrix.T.style

[35]:
PROTEIN_CD3 PROTEIN_CD4 PROTEIN_CD8 PROTEIN_CD2 PROTEIN_CD45RA PROTEIN_CD57 PROTEIN_CD16 PROTEIN_CD14 PROTEIN_CD11c PROTEIN_CD19
AC132872.2 -0.057121 -0.033678 0.022242 -0.026684 0.015539 0.021639 0.023311 0.033757 0.043185 0.003818
AP1G2 0.003806 -0.007512 -0.019456 0.014663 0.041407 0.014168 -0.009372 -0.050276 -0.045060 0.024469
APEH -0.001326 0.003457 -0.013882 0.018534 0.027218 0.025875 0.024947 -0.045393 -0.023549 -0.005811
C2orf68 0.004219 0.018819 -0.013954 -0.001032 0.005328 -0.008308 0.006165 -0.006577 0.008854 -0.013827
CBLB -0.017304 -0.055466 0.036698 0.061028 0.052467 0.016983 -0.003008 -0.053712 -0.036736 -0.016574
CD320 -0.004388 -0.068123 0.056684 0.080928 0.054757 0.083013 0.011571 -0.081302 -0.082888 -0.041756
CD59 0.067997 0.059027 -0.028081 0.077722 -0.031213 0.019404 -0.022022 -0.075959 -0.082218 -0.006978
CD82 -0.006573 0.022886 -0.110960 0.015401 0.015857 0.007061 -0.057523 -0.079809 -0.074393 0.204989
CDK10 -0.005229 -0.007795 -0.001334 0.029529 0.023666 0.009640 0.004888 -0.038625 -0.018294 -0.003952
CITED4 0.124631 0.112310 -0.043116 0.060291 -0.018634 -0.096972 -0.061113 -0.042644 -0.035454 -0.102378
CNOT11 0.039936 0.013133 0.017130 0.063803 -0.016901 0.004260 -0.033437 -0.041727 -0.046556 -0.027419
COA5 -0.014229 -0.025784 0.018401 -0.018384 0.020379 0.017014 0.010107 -0.028425 0.002627 0.014582
COMMD10 -0.034393 -0.001196 0.013050 -0.026152 -0.017461 0.003855 0.027445 0.030550 0.045945 -0.006138
CRYL1 0.010003 0.021328 -0.011779 0.018476 -0.029595 -0.007285 -0.013955 -0.000704 0.013180 -0.005122
CST3 -0.345264 0.019570 0.130521 -0.486025 -0.304183 -0.238582 0.095101 0.742058 0.839373 -0.144727
DOCK9 0.060252 0.034208 0.003071 0.057710 -0.021485 -0.014831 -0.020848 -0.042459 -0.054808 -0.037967
DPH5 0.069149 0.042724 -0.026671 0.057574 0.000370 -0.028330 -0.049750 -0.068381 -0.070880 0.000552
FCN1 -0.306816 0.018249 0.116456 -0.469859 -0.314470 -0.221331 0.084627 0.744712 0.788196 -0.128639
GABPB1-AS1 0.022763 -0.007434 -0.047978 0.018762 0.082215 0.009561 -0.039467 -0.081629 -0.101957 0.070178
GMPR2 0.015361 0.001043 0.007426 0.040610 0.016050 0.009150 0.003124 -0.046416 -0.031352 -0.025418
GPANK1 0.027088 0.012442 -0.007912 0.029707 0.008911 0.001101 -0.018153 -0.044393 -0.041172 0.003045
HSD17B8 0.057555 0.031551 -0.001858 0.051736 -0.003797 -0.025695 -0.028289 -0.048395 -0.042747 -0.037324
KIN -0.001901 -0.001338 -0.001949 0.010628 -0.007908 0.009455 0.009179 -0.011014 0.005309 -0.001850
KRTCAP3 0.050297 0.032055 -0.023494 0.043469 -0.006392 -0.008872 -0.030755 -0.066973 -0.058224 0.015308
LINC00402 0.179249 0.128856 -0.048477 0.104298 -0.028654 -0.092499 -0.083487 -0.091408 -0.115161 -0.092257
LYZ -0.308150 0.020707 0.129248 -0.467036 -0.335649 -0.235306 0.047670 0.787555 0.802407 -0.130433
MRPL46 0.031722 0.012419 -0.010568 0.029103 0.023233 -0.001981 -0.029543 -0.052403 -0.048489 -0.001838
NAGPA -0.001235 0.005049 -0.009468 -0.003438 -0.000618 0.007869 -0.027073 0.004224 0.011885 0.005165
NAT9 0.003746 0.010192 -0.018497 -0.010489 0.031342 -0.014602 -0.019943 -0.018622 -0.010718 0.011342
NDEL1 0.018304 0.016491 -0.009681 0.026141 -0.001764 -0.012347 -0.003363 -0.017089 -0.008624 -0.018704
NKAP -0.005028 0.001098 -0.017527 0.023260 -0.005454 0.029372 0.017674 -0.025580 -0.016604 0.019324
POLDIP2 -0.022635 -0.022624 0.006805 0.018742 0.009673 0.030947 0.012325 -0.018104 0.000050 0.005588
PPP1R3E 0.027233 0.010902 -0.037340 0.022627 0.062267 -0.001434 -0.007526 -0.074333 -0.089508 0.035120
PPP1R8 -0.017425 -0.025874 0.005143 0.013556 0.032592 0.031551 0.020002 -0.039399 -0.025948 0.014832
RIF1 0.008047 -0.000676 0.015667 0.017868 -0.001064 -0.029437 -0.036240 -0.000438 -0.003309 -0.011216
RNF19A -0.016699 -0.031279 -0.001289 0.006751 0.003266 0.043778 -0.003904 -0.017940 -0.016350 0.047477
SCAMP3 -0.004628 -0.025407 0.004629 0.040217 0.030370 0.009009 -0.005542 -0.050580 -0.036885 0.015669
SLC2A4RG 0.078483 0.045201 -0.011598 0.094449 -0.026786 0.000995 -0.018170 -0.074844 -0.070343 -0.051508
SNHG17 0.002901 0.009197 0.003716 0.002854 -0.001650 -0.014684 -0.011865 0.006123 -0.005473 -0.005647
SSR3 -0.039209 0.017750 0.017859 -0.052679 -0.077998 -0.042835 -0.003835 0.113954 0.137226 -0.023115
TMEM245 0.049015 0.048746 -0.021178 0.034155 -0.026393 -0.028445 -0.017854 -0.013756 -0.017335 -0.030610
TOR1AIP1 -0.026530 -0.013282 0.010001 0.011240 0.017378 0.025602 0.017205 -0.030057 -0.006080 0.002709
USP8 -0.040339 -0.012659 -0.003770 -0.038324 -0.007250 -0.011114 -0.013279 0.040626 0.057500 0.036300
VPS26B -0.034631 -0.038457 0.040399 0.013927 0.006315 0.045906 0.025707 -0.006399 0.013654 -0.029259
WDR73 0.026131 0.010095 -0.005092 0.018848 0.009678 -0.004826 0.002599 -0.049521 -0.041935 0.003765
XXbac-BPG299F13.17 0.029379 -0.009988 0.004856 0.043457 0.017339 -0.006785 -0.018289 -0.042830 -0.064013 0.008200
ZBTB1 0.000205 0.000028 0.001345 0.034133 -0.008531 0.012861 0.000149 -0.010494 -0.022030 0.003472
ZBTB20 0.028594 0.006783 -0.031892 0.030955 0.036300 0.007254 -0.019074 -0.066323 -0.070462 0.036235 
[36]:

print(*(i for i in corrMatrix.T.index))

AC132872.2 AP1G2 APEH C2orf68 CBLB CD320 CD59 CD82 CDK10 CITED4 CNOT11 COA5 COMMD10 CRYL1 CST3 DOCK9 DPH5 FCN1 GABPB1-AS1 GMPR2 GPANK1 HSD17B8 KIN KRTCAP3 LINC00402 LYZ MRPL46 NAGPA NAT9 NDEL1 NKAP POLDIP2 PPP1R3E PPP1R8 RIF1 RNF19A SCAMP3 SLC2A4RG SNHG17 SSR3 TMEM245 TOR1AIP1 USP8 VPS26B WDR73 XXbac-BPG299F13.17 ZBTB1 ZBTB20

[37]:

adata.obs.to_csv("../../CITE-seq/human-pbmc-cell.csv")

[52]:

weight_df = pd.DataFrame(model.w[model.w > 0.], index=rna_adata.var_names[model.w > 0.])
weight_df.to_pickle("cite-seq-26-weights.pkl")
weight_df.to_csv("cite-seq-26-weigths.csv")

[56]:

merged_adata.write_loom("cite-seq.loom")

The loom file will lack these fields:
{'X_umap', 'X_pca'}
Use write_obsm_varm=True to export multi-dimensional annotations