def annotate_cell_cycle_scores_human(
adata_results_file,
cell_cycle_file='/ahg/regevdata/users/oursu/code/general_data/cellcycle/regev_lab_cell_cycle_genes.txt'
):
cell_cycle_genes = [x.strip() for x in open(cell_cycle_file)]
s_genes = cell_cycle_genes[:43]
g2m_genes = cell_cycle_genes[43:]
adata_cellcycle = sc.read(adata_results_file + '.basic.h5ad')
adata_cellcycle
sc.pp.log1p(adata_cellcycle)
sc.pp.scale(adata_cellcycle)
sc.tl.score_genes_cell_cycle(adata_cellcycle,
s_genes=s_genes,
g2m_genes=g2m_genes)
adata_annotated = sc.read(adata_results_file + '.basic.h5ad')
#now, assign the cell cycle scores from the adata_cellcycle to adata_annotated
s_scores = adata_cellcycle.obs['S_score']
g2m_scores = adata_cellcycle.obs['G2M_score']
adata_annotated.obs['S_score_added'] = s_scores.loc[
adata_annotated.obs_names]
adata_annotated.obs['G2M_score_added'] = g2m_scores.loc[
adata_annotated.obs_names]
adata_annotated
adata_annotated.write(adata_results_file + '.basic.cc.h5ad')
def __init__(self, ah5_path, scanpy_pca):
self.prefix = ah5_path.split("/")[0]
self.scanpy_pca = scanpy_pca
self.adata_dict = {}
# Create output directories
for output_dir in glob.glob(ah5_path+"/*"):
# Made change for in vitro data
if output_dir.split("/")[-1] == 'preprocessing_summary':
pass
else:
try:
if not os.path.exists(output_dir + "/principle_component_matrices"):
os.makedirs(output_dir + "/principle_component_matrices")
if not os.path.exists(output_dir + "/principle_component_analysis_figures"):
os.makedirs(output_dir + "/principle_component_analysis_figures")
except OSError:
print("Error creating directory")
for processed_file in glob.glob(ah5_path+"*/gene_matrices/*.h5ad"):
# Made changes for in vitro data
if len(processed_file.split('/')[-1].split('_')) <= 2:
print(processed_file)
self.adata_dict[processed_file.split('/')[-1].split('.')[0]] = sc.read(processed_file)
def generate_cluster_expression_output_file(sample_cluster_dict,
raw_log_scanpy_obj, output_dir):
"""
:param sample_clustered_dict:
:param output_dir:
return:
"""
for sample_key, adata in sample_cluster_dict.items():
sample_cluster_dict[sample_key].raw = sc.read(raw_log_scanpy_obj)
for cluster in list(set(adata.obs.louvain)):
cluster_specific_cells = sample_cluster_dict[sample_key].obs.loc[(
sample_cluster_dict[sample_key].obs.louvain == cluster
)].index.tolist()
sample_cluster_subset = sample_cluster_dict[sample_key][
cluster_specific_cells, :]
try:
# this works for sparse matrices
subset_df = pd.DataFrame(
data=sample_cluster_subset.raw.X.toarray(),
index=sample_cluster_subset.obs.index.tolist(),
columns=sample_cluster_subset.raw.var.index.tolist())
except AttributeError:
# AttributeError occurs for non-sparse matrices (no toarrray())
subset_df = pd.DataFrame(
data=sample_cluster_subset.raw.X,
index=sample_cluster_subset.obs.index.tolist(),
columns=sample_cluster_subset.raw.var.index.tolist())
subset_df.to_csv(output_dir + sample_key + '/' + sample_key +
'_cluster_' + cluster + '.tsv',
sep='\t',
index=True)
def read_raw_file(self):
"""
Reads the raw data file and turns it into a dense matrix, stored as
an attribute.
Returns
-------
"""
print("reading single cell data from {}".format(self.raw_file))
file_format = self.raw_file.split('.')[-1]
if file_format == 'h5':
andata = sc.read_10x_h5(self.raw_file)
elif file_format == 'h5ad':
andata = sc.read(self.raw_file)
else:
raise ValueError('Reading [ %s ] failed, the inferred file '
'format [ %s ] is not supported. Please convert '
'your file to either h5 or h5ad format.' %
(self.raw_file, file_format))
# appends -1 -2... to the name of genes that already exist
andata.var_names_make_unique()
if sp_sparse.issparse(andata.X):
andata.X = andata.X.toarray()
self.sc_raw = andata
def load_10x_scanpy(path, batch_label):
sc01 = sc.read('{}/matrix.mtx'.format(path), cache=True).T
sc01.var_names = pd.read_table('{}/genes.tsv'.format(path), header=None)[1]
sc01.obs_names = pd.read_table('{}/barcodes.tsv'.format(path),
header=None)[0]
sc01.obs_names = sc01.obs_names.str.replace('-1', '')
sc01.var_names_make_unique()
sc.pp.filter_cells(sc01, min_genes=200)
sc.pp.filter_genes(sc01, min_cells=3)
sc01.obs['n_UMI'] = np.sum(sc01.X, axis=1).A1
mito_genes = sc01.var_names[sc01.var_names.str.match(r'^mt-')]
sc01.obs['percent_mito'] = np.sum(sc01[:, mito_genes].X,
axis=1).A1 / sc01.obs['n_UMI']
ribo_genes = sc01.var_names[sc01.var_names.str.match(
r'^(Rpl|Rps|Mrpl|Mrps)')]
sc01.obs['percent_ribo'] = np.sum(sc01[:, ribo_genes].X,
axis=1).A1 / sc01.obs['n_UMI']
assgn = pd.read_csv('{}/{}_assgn.csv'.format(
os.path.join(CUR_DIR, '..', '01-cluster-sc01-sc02'),
batch_label,
),
index_col=0)
assgn.columns = ['cluster']
sc01.obs['cluster'] = assgn.cluster[sc01.obs.index]
return sc01
def read_dataset(adata, transpose=False, test_split=False, copy=False):
if isinstance(adata, sc.AnnData):
if copy:
adata = adata.copy()
elif isinstance(adata, str):
adata = sc.read(adata, first_column_names=True)
else:
raise NotImplementedError
# check if observations are unnormalized using first 10
X_subset = adata.X[:10]
norm_error = 'Make sure that the dataset (adata.X) contains unnormalized count data.'
if sp.sparse.issparse(X_subset):
assert (X_subset.astype(int) != X_subset).nnz == 0, norm_error
else:
assert np.all(X_subset.astype(int) == X_subset), norm_error
if transpose: adata = adata.transpose()
if test_split:
train_idx, test_idx = train_test_split(np.arange(adata.n_obs),
test_size=0.1,
random_state=42)
spl = pd.Series(['train'] * adata.n_obs)
spl.iloc[test_idx] = 'test'
adata.obs['dca_split'] = spl.values
else:
adata.obs['dca_split'] = 'train'
adata.obs['dca_split'] = adata.obs['dca_split'].astype('category')
print('dca: Successfully preprocessed {} genes and {} cells.'.format(
adata.n_vars, adata.n_obs))
return adata
def main() :
# check there is exactly one command-line argument provided
if(len(sys.argv) != 2):
sys.stderr.write("usage: " + __file__ + " <adata-file-path>\n")
sys.exit(1)
anndata = sc.read(sys.argv[1])
# run PCA, compute pairwise distance, k-nearest-neighbors \n# trick to replacing scanpy implementation with our own
# is to \n# update the anndata object with our intermediate values\n# \n# we replace the scanpy version of PCA by
# updating\n# .adata.obsm['X_pca'] = our PCA() output\n#\n# we replace the scanpy version of k-nearest-neighbors by
# updating\n# self.adata.uns['neighbors']['connectivities'] = our knn() output\n
# self.adata.uns['neighbors']['distances'] = out knn()
# output\n\nknng = KnnG(anndata, n_neighbors=12, runPCA=True, nPC=50)")
# umap() reduce results to 2 deminsions so that we can plot the data\nsc.tl.umap(anndata)')
# ### 1.c. [5 pts] Turn in a UMAP plot of your 12-NN graph calculated from the combined chemistry PBMC dataset
# colored by batch (the chemistry used)
scanpy.pl.umap(anndata, color=['Method'])
# ### 1.d. [5 pts] Turn in another UMAP plot of your 12-NN graph calculated from the combined chemistry PBMC dataset
# but colored by cell type
scanpy.pl.umap(anndata, color=['Cell type'])
def testAdata(self):
'''
The real deal
'''
self.logger.info("BEGIN")
anndata = sc.read("../PBMC.merged.h5ad")
xxx = anndata.uns['neighbors']['connectivities']
# anndata.uns['neighbors']['connectivities'] csr_matrix
# shape <class 'tuple'>: (15476, 15476)
#
# adata.obs['louvain']
# Series: index
# data_3p-AAACCTGAGCATCATC-0 9
# data_3p-AAACCTGAGCTAGTGG-0 5
# <class 'tuple'>: (15476,)
#knn takes about 3 or 4 min
# run our implementation of nearest neighboors and update anndata
# todo try running with out knng maybe adata has values already save time
KnnG(anndata, n_neighbors=12, runPCA=True, nPC=50)
self.logger.info("begin Louvain.runWithAdata")
start = timer()
root = Louvain.runWithAdata(anndata)
end = timer()
self.logger.info("Louvain.runWithAdata execution time:{}"\
.format(timedelta(seconds=end-start)))
self.logger.info("END\n")
def Smillie2019_processed():
"""Processed data from Smillie et al. Intra- and Inter-cellular Rewiring of the Human Colon during Ulcerative Colitis. Cell. 2019
The data consists of processed single cell expression data from colon mucosa from 7 ulcerative colitis (UC) patients and 10 healthy controls, paired samples (inlamed, non-inflamed for UC, location-matched for healthy): 34 samples. Epithelial (EPI) and lamina propria (LP) fractions enriched in a two-step digestion process. Data was filtered, batch corrected using BBKNN and celltypes were annotated.
Returns
-------
adata : :class:`~anndata.AnnData`
Annotated data matrix.
Example
-------
>>> import besca as bc
>>> adata = bc.datasets.Smillie2019_processed()
>>> adata
"""
filename = pkg_resources.resource_filename(
'besca', 'datasets/data/Smillie2019_processed.h5ad')
adata = read(filename, cache=True)
return adata
def read10xData(path,min_genes):
"""
Reads, precesses and returns single cell data.
Parameters
----------
path : Str
Diractory path, the location of single cell data.
min_genes : Int
The minimum number of genes for a cell to have in order to participate the analysis.
Returns
-------
scData : AnnData
Single cell data.
"""
result = sc.read(path + 'matrix.mtx').transpose() #, cache=True
result.var_names = np.genfromtxt(path + 'genes.tsv', dtype=str)[:, 1]
result.obs_names = np.genfromtxt(path + 'barcodes.tsv', dtype=str)
result.var_names_make_unique()
result.obs['n_counts'] = np.sum(result.X, axis=1).A1
sc.pp.filter_cells(result, min_genes=min_genes)
return result
def generate_sample_expression_output_file(sample_cluster_dict,
raw_log_scanpy_obj, output_dir):
"""
:param sample_clustered_dict:
:param output_dir:
return:
"""
for sample_key, adata in sample_cluster_dict.items():
sample_cluster_dict[sample_key].raw = sc.read(raw_log_scanpy_obj)
#sample_expression_df = pd.DataFrame(adata.X, index = adata.obs.index.tolist(), columns = adata.var.index.tolist())
try:
# this works for sparse matrices
sample_expression_df = pd.DataFrame(
data=sample_cluster_dict[sample_key].raw.X.toarray(),
index=sample_cluster_dict[sample_key].obs.index.tolist(),
columns=sample_cluster_dict[sample_key].raw.var.index.tolist())
except AttributeError:
# AttributeError occurs for non-sparse matrices (no toarrray())
sample_expression_df = pd.DataFrame(
data=sample_cluster_dict[sample_key].raw.X,
index=sample_cluster_dict[sample_key].obs.index.tolist(),
columns=sample_cluster_dict[sample_key].raw.var.index.tolist())
sample_expression_df.to_csv(output_dir + sample_key +
'_log_expression.tsv',
sep='\t',
index=True)
def __init__(self, pca_h5ad, n_pcs, n_neighbors):
self.adata_dict = {}
self.prefix = pca_h5ad
for processed_file in glob.glob(
pca_h5ad + "*/principle_component_matrices/*.h5ad"):
self.adata_dict[processed_file.split('/')[-3]] = sc.read(
processed_file)
for output_dir in glob.glob(pca_h5ad + "*/"):
if output_dir.split("/")[-1] == 'preprocessing_summary':
pass
else:
try:
if not os.path.exists(output_dir + '/cluster_matrices'):
os.makedirs(output_dir + '/cluster_matrices')
if not os.path.exists(output_dir + "/cluster_analysis"):
os.makedirs(output_dir + "/cluster_analysis")
# if not os.path.exists(output_dir + "/cluster_analysis/tSNE"):
# os.makedirs(output_dir + "/cluster_analysis/tSNE")
#
# if not os.path.exists(output_dir + "/cluster_analysis/umap"):
# os.makedirs(output_dir + "/cluster_analysis/umap")
#
# if not os.path.exists(output_dir + "/cluster_analysis/louvain"):
# os.makedirs(output_dir + "/cluster_analysis/louvain")
except OSError:
print("Error creating directory")
self.n_pcs = n_pcs
self.n_neighbors = n_neighbors
def read_h5ad(args):
#read input h5ad
dataset = sc.read(args.input)
print("File read!")
compute_entropy(dataset)
def getdata(dataset):
basedir = os.path.abspath(os.path.join(__file__, ".."))
if dataset == "green":
adata = sc.read(basedir + "/data/green/green.h5ad")
process_clusts(adata, "CellType")
elif dataset == "paul":
adata = sc.read(basedir + "/data/paul/paul.h5ad")
process_clusts(adata, "paul15_clusters")
elif dataset == "zeisel":
adata = sc.read(basedir + "/data/zeisel/zeisel.h5ad")
process_clusts(adata, "group")
elif dataset == "zheng":
adata = sc.read(basedir + "/data/zheng/fresh_68k_bulk_labels.h5ad")
process_clusts(adata)
else:
raise ValueError("No such dataset")
return adata
def check_dl(filename, url):
try:
adata = read(filename, backup_url=url, cache=True)
except Exception:
raise URLError(
f'\n\n\n {filename} could not be downloaded from {url}; \n Please download it manually and store it in your besca installation: besca/datasets/data/'
)
return adata
def train(data_name="pbmc", cell_type="CD4T", p_type="unbiased"):
train_path = f"../data/train_{data_name}.h5ad"
if data_name == "pbmc":
ctrl_key = "control"
stim_key = "stimulated"
cell_type_key = "cell_type"
elif data_name == "hpoly":
ctrl_key = "Control"
stim_key = "Hpoly.Day10"
cell_type_key = "cell_label"
elif data_name == "salmonella":
ctrl_key = "Control"
stim_key = "Salmonella"
cell_type_key = "cell_label"
data = sc.read(train_path)
print("data has been loaded!")
train = data[~((data.obs["condition"] == stim_key) &
(data.obs[cell_type_key] == cell_type))]
pca = PCA(n_components=100)
pca.fit(train.X.A)
train_real_cd = train[train.obs["condition"] == "control", :]
if p_type == "unbiased":
train_real_cd = scgen.util.balancer(train_real_cd)
train_real_stimulated = train[train.obs["condition"] == "stimulated", :]
if p_type == "unbiased":
train_real_stimulated = scgen.util.balancer(train_real_stimulated)
import scipy.sparse as sparse
if sparse.issparse(train_real_cd.X):
train_real_cd.X = train_real_cd.X.A
train_real_stimulated.X = train_real_stimulated.X.A
train_real_stimulated_PCA = pca.transform(train_real_stimulated.X)
train_real_cd_PCA = pca.transform(train_real_cd.X)
adata_list = scgen.util.extractor(data, cell_type, {
"ctrl": ctrl_key,
"stim": stim_key
})
if sparse.issparse(adata_list[1].X):
adata_list[1].X = adata_list[1].X.A
adata_list[2].X = adata_list[2].X.A
ctrl_CD4T_PCA = pca.transform(adata_list[1].X)
predicted_cells = predict(pca, train_real_cd_PCA,
train_real_stimulated_PCA, ctrl_CD4T_PCA, p_type)
all_Data = sc.AnnData(
np.concatenate([adata_list[1].X, adata_list[2].X, predicted_cells]))
all_Data.obs["condition"] = ["ctrl"] * len(adata_list[1].X) + ["real_stim"] * len(adata_list[2].X) + \
["pred_stim"] * len(predicted_cells)
all_Data.var_names = adata_list[3].var_names
if p_type == "unbiased":
sc.write(f"../data/reconstructed/PCAVecArithm/PCA_CD4T.h5ad", all_Data)
else:
sc.write(f"../data/reconstructed/PCAVecArithm/PCA_CD4T_biased.h5ad",
all_Data)
def __init__(self, cluster_h5ad, marker_gene_file, cell_type):
self.prefix = cluster_h5ad
self.cell_type = cell_type
for clusters in glob.glob(self.prefix + "/*"):
if clusters.split("/")[-1] != 'preprocessing_summary':
try:
if not os.path.exists(clusters + "/cluster_analysis"):
os.makedirs(clusters + "/cluster_analysis")
# Create marker gene analysis directory
if not os.path.exists(clusters + "/cluster_analysis/marker_gene_analysis"):
os.makedirs(clusters + "/cluster_analysis/marker_gene_analysis")
if not os.path.exists(clusters + "/cluster_analysis/marker_gene_analysis/tSNE"):
os.makedirs(clusters + "/cluster_analysis/marker_gene_analysis/tSNE")
if not os.path.exists(clusters + "/cluster_analysis/marker_gene_analysis/louvain"):
os.makedirs(clusters + "/cluster_analysis/marker_gene_analysis/louvain")
if not os.path.exists(clusters + "/cluster_analysis/marker_gene_analysis/umap"):
os.makedirs(clusters + "/cluster_analysis/marker_gene_analysis/umap")
if not os.path.exists(clusters + "/cluster_analysis/marker_gene_analysis/heatmap"):
os.makedirs(clusters + "/cluster_analysis/marker_gene_analysis/heatmap")
if not os.path.exists(clusters + "/cluster_analysis/marker_gene_analysis/dotplot"):
os.makedirs(clusters + "/cluster_analysis/marker_gene_analysis/dotplot")
if not os.path.exists(clusters + "/cluster_analysis/marker_gene_analysis/cluster_gene_rankings"):
os.makedirs(clusters + "/cluster_analysis/marker_gene_analysis/cluster_gene_rankings")
# Create cell type analysis directory
if not os.path.exists(clusters + "/cluster_analysis/cell_type_analysis"):
os.makedirs(clusters + "/cluster_analysis/cell_type_analysis")
if not os.path.exists(clusters + "/cluster_analysis/cell_type_analysis/tSNE"):
os.makedirs(clusters + "/cluster_analysis/cell_type_analysis/tSNE")
if not os.path.exists(clusters + "/cluster_analysis/cell_type_analysis/louvain"):
os.makedirs(clusters + "/cluster_analysis/cell_type_analysis/louvain")
if not os.path.exists(clusters + "/cluster_analysis/cell_type_analysis/umap"):
os.makedirs(clusters + "/cluster_analysis/cell_type_analysis/umap")
except OSError:
print("Error creating directory.")
# Create dictionary containing cell cluster cluster cluster_matrices
self.cluster_matrices_dict = {}
for cluster_matrices in glob.glob(self.prefix + "/*/cluster_matrices/*"):
self.cluster_matrices_dict[cluster_matrices.split('/')[-3]] = sc.read(cluster_matrices)
# Create marker gene list with provided txt file
self.marker_gene_file = marker_gene_file
if self.marker_gene_file:
self.marker_gene_list = list(map(lambda x: x.strip(), marker_gene_file.readlines()))
def adata_neighbors():
adata = sc.read('./data/pbmc3k_raw.h5ad',
backup_url='http://falexwolf.de/data/pbmc3k_raw.h5ad')
sc.pp.filter_genes(adata, min_cells=1)
sc.pp.normalize_per_cell(adata)
sc.pp.log1p(adata)
sc.pp.pca(adata)
sc.pp.neighbors(adata)
return adata
def reconstruct():
train_path = "../data/train_pbmc.h5ad"
data = sc.read(train_path)
ctrl_key = "control"
stim_key = "stimulated"
all_data = anndata.AnnData()
print(data.obs["cell_type"].unique().tolist())
for idx, cell_type in enumerate(data.obs["cell_type"].unique().tolist()):
pca = PCA(n_components=100)
train = data[~((data.obs["condition"] == stim_key) &
(data.obs["cell_type"] == cell_type))]
pca.fit(train.X.A)
print(cell_type, end="\t")
train_real_stimulated = data[data.obs["condition"] == stim_key, :]
train_real_stimulated = train_real_stimulated[
train_real_stimulated.obs["cell_type"] != cell_type]
train_real_stimulated = scgen.util.balancer(train_real_stimulated)
train_real_stimulated_PCA = pca.transform(train_real_stimulated.X)
train_real_cd = data[data.obs["condition"] == ctrl_key, :]
train_real_cd = scgen.util.balancer(train_real_cd)
train_real_cd_PCA = pca.transform(train_real_cd.X)
cell_type_adata = data[data.obs["cell_type"] == cell_type]
cell_type_ctrl = cell_type_adata[cell_type_adata.obs["condition"] ==
ctrl_key]
cell_type_stim = cell_type_adata[cell_type_adata.obs["condition"] ==
stim_key]
if sparse.issparse(cell_type_ctrl.X):
cell_type_ctrl_PCA = pca.transform(cell_type_ctrl.X.A)
else:
cell_type_ctrl_PCA = pca.transform(cell_type_ctrl.X)
predicted_cells = predict(pca, train_real_cd_PCA,
train_real_stimulated_PCA,
cell_type_ctrl_PCA)
if sparse.issparse(cell_type_ctrl.X):
all_Data = sc.AnnData(
np.concatenate(
[cell_type_ctrl.X.A, cell_type_stim.X.A, predicted_cells]))
else:
all_Data = sc.AnnData(
np.concatenate(
[cell_type_ctrl.X, cell_type_stim.X, predicted_cells]))
all_Data.obs["condition"] = [f"{cell_type}_ctrl"] * cell_type_ctrl.shape[0] + [f"{cell_type}_real_stim"] * \
cell_type_stim.shape[0] + \
[f"{cell_type}_pred_stim"] * len(predicted_cells)
all_Data.obs["cell_type"] = [f"{cell_type}"] * (
cell_type_ctrl.shape[0] + cell_type_stim.shape[0] +
len(predicted_cells))
all_Data.var_names = cell_type_adata.var_names
if idx == 0:
all_data = all_Data
else:
all_data = all_data.concatenate(all_Data)
print(cell_type)
sc.write("../data/reconstructed/PCAVecArithm/PCA_pbmc.h5ad", all_data)
def read_h5ad(x):
#read input h5ad
dataset = sc.read(x)
kwargs = {}
kwargs["batch_vector"] = dataset.obs["Batch"]
kwargs["cell_type_vector"] = dataset.obs["cell_type1"]
do_the_filtering(dataset, **kwargs)
def load_10x_12k_mix_mouse():
filename_data = '/data/martin/single_cell/10x_12k_mix/filtered_gene_bc_matrices/mm10/matrix.mtx'
filename_genes = '/data/martin/single_cell/10x_12k_mix/filtered_gene_bc_matrices/mm10/genes.tsv'
filename_barcodes = '/data/martin/single_cell/10x_12k_mix/filtered_gene_bc_matrices/mm10/barcodes.tsv'
data = sc.read(filename_data, cache=True).transpose()
data.var_names = np.genfromtxt(filename_genes, dtype=str)[:, 1]
data.smp_names = np.genfromtxt(filename_barcodes, dtype=str)
return data
def paul_test(n_top_gene=100):
adata = sc.read("data/paul15/paul15.h5ad")
sc.pp.filter_cells(adata, min_genes=10)
sc.pp.normalize_per_cell(adata, counts_per_cell_after=1e4)
sc.pp.filter_genes(adata, min_cells=20)
sc.pp.filter_genes_dispersion(adata, n_top_genes=1000)
sc.pp.log1p(adata)
sc.pp.scale(adata, zero_center=True, max_value=False)
return adata
def read_dataset(input_file, transpose=False):
"""
Construct a anndata object
"""
if os.path.isfile(input_file):
print("The value os", os.path.isfile(input_file))
if str(input_file).endswith('h5ad'):
adata = sc.read(input_file)
elif sum([
str(input_file).endswith(str(i))
for i in ["tsv", 'TSV', 'tab', 'data']
]):
adata = sc.read_text(input_file, sep="\t", first_column_names=True)
if transpose:
adata = adata.T
elif sum([str(input_file).endswith(str(i)) for i in ['csv', "CSV"]]):
adata = sc.read_text(input_file, sep=",", first_column_names=True)
if transpose:
adata = adata.T
else:
#ValueError 'The file must be one of *.h5ad, *.tsv,*TSV,*.tab,*data, *csv,*CSV'
print(
"The file must be one of *.h5ad, *.tsv,*TSV,*.tab,*data, *csv,*CSV"
)
else:
#read folder
mtx = sc.read_mtx(os.path.join(input_file, "matrix.mtx"))
num_lines = sum(
1 for line in open(os.path.join(input_file, 'barcodes.tsv')))
cellinfo = pd.read_csv(os.path.join(input_file, "barcodes.tsv"),
sep="\t",
header=None if num_lines == mtx.shape[1] else 0)
if not 'cellname' in cellinfo.columns:
cellinfo['cellname'] = cellinfo.iloc[:, 0]
num_lines = sum(
1 for line in open(os.path.join(input_file, 'genes.tsv')))
geneinfo = pd.read_csv(os.path.join(input_file, "genes.tsv"),
sep="\t",
header=None if num_lines == mtx.shape[0] else 0)
if not 'genename' in geneinfo.columns:
geneinfo[
'genename'] = geneinfo.iloc[:,
1] # for 10x,the second columns is the genename, and the first column is gene_id
#create anndata
adata = sc.AnnData(mtx.X.T, obs=cellinfo, var=geneinfo)
adata.obs_names = adata.obs["cellname"]
adata.var_names = adata.var["genename"]
adata.obs_names_make_unique(join="-")
adata.var_names_make_unique(join="-")
#create time
now = datetime.datetime.now()
adata.uns["ProjectName"] = "DESC created in" + str(
now.strftime("%Y-%m-%d %H:%M"))
print("Creat adata successfully! The adata infofation is", adata)
return adata
def getAnnData(matrix, genelist, barcodes):
sc.settings.verbosity = 0 # verbosity: errors (0), warnings (1), info (2), hints (3)
sc.settings.autoshow = False
print('Reading matrix...')
adata = sc.read(matrix, cache=False).T
print(adata)
print('Reading gene list...')
genes = pd.read_csv(genelist, header=None, sep='\t')
geneNames = anndata.utils.make_index_unique(pd.Index(genes[1]))
adata.var_names = geneNames
adata.var['gene_ids'] = genes[0].values
adata.obs_names = pd.read_csv(barcodes, header=None)[0]
adata.var_names_make_unique()
sc.pp.filter_cells(adata, min_genes=200)
sc.pp.filter_genes(adata, min_cells=3)
mito_genes = [name for name in adata.var_names if name.startswith('MT-')]
# for each cell compute fraction of counts in mito genes vs. all genes
# the `.A1` is only necessary as X is sparse to transform to a dense array after summing
adata.obs['percent_mito'] = np.sum(adata[:, mito_genes].X,
axis=1).A1 / np.sum(adata.X, axis=1).A1
# add the total counts per cell as observations-annotation to adata
adata.obs['n_counts'] = adata.X.sum(axis=1).A1
adata = adata[adata.obs['n_genes'] < 2500, :]
adata = adata[adata.obs['percent_mito'] < 0.05, :]
adata.raw = sc.pp.log1p(adata, copy=True)
sc.pp.normalize_per_cell(adata, counts_per_cell_after=1e4)
filter_result = sc.pp.filter_genes_dispersion(adata.X,
min_mean=0.0125,
max_mean=3,
min_disp=0.5)
adata = adata[:, filter_result.gene_subset]
sc.pp.log1p(adata)
sc.pp.regress_out(adata, ['n_counts', 'percent_mito'])
sc.pp.scale(adata, max_value=10)
sc.tl.pca(adata, svd_solver='arpack')
sc.pp.neighbors(adata, n_neighbors=10, n_pcs=40)
sc.tl.umap(adata)
sc.tl.louvain(adata)
sc.pl.umap(adata, color=['louvain'], show=False)
sc.tl.rank_genes_groups(adata, 'louvain', method='logreg')
return adata
def main():
ad_path, cs_name, output, gmt_file = parse_args()
gene_sets = read_gmt(gmt_file)
ad = sc.read(ad_path)
percentage_markers_expressed_in_cluster(ad, cs_name,
gene_sets).to_csv(output,
header=True)
def normalize(adata, copy=True, highly_genes = None, filter_min_counts=True,
size_factors=True, normalize_input=True, logtrans_input=True):
"""
Normalizes input data and retains only most variable genes
(indicated by highly_genes parameter)
Args:
adata ([type]): [description]
copy (bool, optional): [description]. Defaults to True.
highly_genes ([type], optional): [description]. Defaults to None.
filter_min_counts (bool, optional): [description]. Defaults to True.
size_factors (bool, optional): [description]. Defaults to True.
normalize_input (bool, optional): [description]. Defaults to True.
logtrans_input (bool, optional): [description]. Defaults to True.
Raises:
NotImplementedError: [description]
Returns:
[type]: [description]
"""
if isinstance(adata, sc.AnnData):
if copy:
adata = adata.copy()
elif isinstance(adata, str):
adata = sc.read(adata)
else:
raise NotImplementedError
norm_error = 'Make sure that the dataset (adata.X) contains unnormalized count data.'
assert 'n_count' not in adata.obs, norm_error
if adata.X.size < 50e6: # check if adata.X is integer only if array is small
if sp.sparse.issparse(adata.X):
assert (adata.X.astype(int) != adata.X).nnz == 0, norm_error
else:
assert np.all(adata.X.astype(int) == adata.X), norm_error
if filter_min_counts:
sc.pp.filter_genes(adata, min_counts=1)#3
sc.pp.filter_cells(adata, min_counts=1)
if size_factors or normalize_input or logtrans_input:
adata.raw = adata.copy()
else:
adata.raw = adata
if size_factors:
sc.pp.normalize_per_cell(adata)
adata.obs['size_factors'] = adata.obs.n_counts / np.median(adata.obs.n_counts)
else:
adata.obs['size_factors'] = 1.0
if logtrans_input:
sc.pp.log1p(adata)
if highly_genes != None:
sc.pp.highly_variable_genes(adata, min_mean=0.0125, max_mean=3, min_disp=0.5, n_top_genes = highly_genes, subset=True)
if normalize_input:
sc.pp.scale(adata)
return adata
def load_10x_1_3mil_subsample(opt=10):
if opt == 10:
filename_data = '/data/martin/exp_sceb/subsample_1.3mil/data_1.3mil_high10_gene.h5ad'
elif opt == 5:
filename_data = '/data/martin/exp_sceb/subsample_1.3mil/data_1.3mil_high5_gene.h5ad'
elif opt == 1:
filename_data = '/data/martin/exp_sceb/subsample_1.3mil/data_1.3mil_high1_gene.h5ad'
elif opt == 0.5:
filename_data = '/data/martin/exp_sceb/subsample_1.3mil/data_1.3mil_high0.5_gene.h5ad'
data = sc.read(filename_data)
return data
def readData(self, countsFile=""):
if countsFile == "":
countsFile = self.CountsFile
if countsFile == "":
print("please input counts file path")
return ""
self.CountsFile = countsFile
datapath = self.CountsFile
if os.path.isdir(datapath):
files = os.listdir(datapath)
for i in files:
if i.endswith(".gz"):
print(i)
target = datapath + "/*.gz"
print(target)
command = subprocess.Popen("gunzip " + target,
shell=True,
stdin=PIPE,
stdout=PIPE,
stderr=STDOUT)
output = command.stdout.read()
break
files = os.listdir(datapath)
for i in files:
if i == "features.tsv":
os.rename(datapath + "/features.tsv",
datapath + "/genes.tsv")
break
files = list(os.listdir(datapath))
if ('barcodes.tsv' in files) and ('barcodes.tsv'
in files) and ("genes.tsv"
in files):
adata = sc.read_10x_mtx(datapath, var_names='gene_symbols')
self.data = adata
self.preprocess()
else:
print("input data is not correct")
return ""
elif os.path.isfile(datapath):
if datapath.endswith(".h5ad"):
adata = sc.read(datapath)
else:
adata = sc.read_csv(datapath)
adata = adata.T
self.data = adata
self.preprocess()
else:
print("file or dir not exists")
return ""
def load_10x_ercc_1k():
"""
https://support.10xgenomics.com/single-cell-gene-expression/datasets/1.1.0/ercc
"""
filename_data = '/data/martin/single_cell/10x_ERCC_1k/filtered_matrices_mex/ercc92/matrix.mtx'
filename_genes = '/data/martin/single_cell/10x_ERCC_1k/filtered_matrices_mex/ercc92/genes.tsv'
filename_barcodes = '/data/martin/single_cell/10x_ERCC_1k/filtered_matrices_mex/ercc92/barcodes.tsv'
data = sc.read(filename_data, cache=True).transpose()
data.var_names = np.genfromtxt(filename_genes, dtype=str)[:, 1]
data.obs_names = np.genfromtxt(filename_barcodes, dtype=str)
return data
def preprocessing(data_folder,
min_genes=200,
min_cells=3,
max_genes=7000,
mito_cutoff=False,
normalize=True):
"""
Combined function for preprocessing using Scanpy. For a more complete documentation on preprocessing, please
visit
Input:
data_folder = Path to data files
min_genes = Minimum amount of genes required for a gene to be valid (default is set at 200)
min_cells = Minimum amount of cells required for a gene to be valid (default is set at 3)
max_genes = Maximum amount of genes permitted for a cell to be valid (default is set at 7000)
mito_cutoff = Percentage of genes permitted to be assigned to
mitochondrial Genes in a cell (default is set at False=0)
normalize = Normalize the Anndata object (default set at True)
Returns AnnData type from matrix - genes - barcodes. Full documentation on AnnData can be found on github.
"""
#Read data and create initial AnnData Frame
path = '{}/'.format(data_folder)
adata = sc.read(path + 'matrix.mtx', cache=True).T # transpose the data
adata.var_names = pd.read_csv(path + 'genes.tsv', header=None, sep='\t')[1]
adata.obs_names = pd.read_csv(path + 'barcodes.tsv', header=None)[0]
adata.var_names_make_unique()
#Filter data with min_genes per cell, max_genes per cell, min_cells per genes
sc.pp.filter_cells(adata, min_genes=min_genes)
sc.pp.filter_genes(adata, min_cells=min_cells)
adata = adata[adata.obs['n_genes'] < max_genes, :]
# add the total counts per cell as observations-annotation to adata
adata.obs['n_counts'] = adata.X.sum(axis=1).A1
#Create mito_genes and possible filter
mito_genes = [name for name in adata.var_names if name.startswith('MT-')]
adata.obs['percent_mito'] = np.sum(adata[:, mito_genes].X,
axis=1).A1 / np.sum(adata.X, axis=1).A1
if int(mito_cutoff) == False:
pass
else:
adata = adata[adata.obs['percent_mito'] < float(mito_cutoff), :]
#Normalize data option
if normalize == True:
sc.pp.normalize_per_cell(adata, counts_per_cell_after=1e4)
return adata
