def integrate_datasets(datasets, genes, method='scanorama', merge=True):
"""
Parameters:
-----------
datasets: list of scipy.sparse.csr_mstrix or np.array()
genes: list of lists
Returns:
-----------
integrated_dim_red: np.array() cell by dimensions
corrected_expression_matrix: np.array() if merge, or lists of np.array() or scipy.sparse.csr_matrix
corrected_genes: np.array()
"""
if method == 'scanorama':
import scanorama
integrated_dim_red, corrected_expression_matrix, corrected_genes = \
scanorama.correct(datasets, genes, dimred=100, return_dimred=True)
if merge:
corrected_expression_matrix = scipy.sparse.vstack(
corrected_expression_matrix).toarray()
integrated_dim_red = np.vstack(integrated_dim_red)
return integrated_dim_red, corrected_expression_matrix, corrected_genes
def correct_scanorama(Xs, genes):
from scanorama import correct
Xs, genes = correct(Xs, [genes for _ in Xs], alpha=0, batch_size=10000)
X = vstack(Xs)
return X
dtype=int)
G = simulation_dim[1]
B = simulation_dim[2]
simulation_gene_list = ["gene_" + str(i) for i in range(1, (G + 1))]
simulation_gene_list = [simulation_gene_list for i in range(0, B)]
offset = 0
simulation_data = []
for b in range(B):
simulation_data.append(
simulation_count[offset:(offset + simulation_dim[(3 + b)]), :])
offset += simulation_dim[(3 + b)]
simulation_integrated, simulation_corrected, simulation_genes = scanorama.correct(
simulation_data, simulation_gene_list, return_dimred=True)
#Please note that *_corrected is of type "scipy.sparse.csr.csr_matrix".
for b in range(B):
np.savetxt("./data/scanorama_simulation_v1_integrated_batch" + str(b + 1) +
".txt",
simulation_integrated[b],
fmt='%10.9f',
delimiter="\t")
np.savetxt("./data/scanorama_simulation_v1_corrected_batch" + str(b + 1) +
".txt",
simulation_corrected[b].toarray(),
fmt='%10.9f',
delimiter="\t")
np.savetxt("scanorama_simulation_v1_genes.txt",
simulation_genes,
NAMESPACE = 'er_stress'
data_names = [
'data/pancreas/pancreas_inDrop',
'data/pancreas/pancreas_multi_celseq2_expression_matrix',
'data/pancreas/pancreas_multi_celseq_expression_matrix',
'data/pancreas/pancreas_multi_fluidigmc1_expression_matrix',
'data/pancreas/pancreas_multi_smartseq2_expression_matrix',
]
if __name__ == '__main__':
datasets, genes_list, n_cells = load_names(data_names)
#datasets, genes = merge_datasets(datasets, genes_list)
#datasets_dimred, genes = process_data(datasets, genes, hvg=hvg)
datasets, genes = correct(datasets, genes_list)
X = vstack(datasets).toarray()
X[X < 0] = 0
cell_labels = (
open('data/cell_labels/pancreas_cluster.txt').read().rstrip().split())
er_idx = [i for i, cl in enumerate(cell_labels) if cl == 'beta_er']
beta_idx = [i for i, cl in enumerate(cell_labels) if cl == 'beta']
gadd_idx = list(genes).index('GADD45A')
herp_idx = list(genes).index('HERPUD1')
plt.figure()
plt.boxplot([X[er_idx, gadd_idx], X[beta_idx, gadd_idx]], showmeans=True)
plt.title('GADD45A (p = {})'.format(
ttest_ind(X[er_idx, gadd_idx], X[beta_idx, gadd_idx],
import scanorama
pancreas_data = np.transpose(
np.genfromtxt("../../data/count_data_pancreas_v1.txt",
delimiter=" ",
dtype=int))
pancreas_gene_list = ["gene_" + str(i) for i in range(1, 2481)]
pancreas_gene_list = [pancreas_gene_list for i in range(0, 4)]
pancreas_data = [
pancreas_data[0:1006, 0:2480], pancreas_data[1006:3337, 0:2480],
pancreas_data[3337:4932, 0:2480], pancreas_data[4932:7095, 0:2480]
]
pancreas_integrated, pancreas_corrected, pancreas_genes = scanorama.correct(
pancreas_data, pancreas_gene_list, return_dimred=True)
#Please note that *_corrected is of type "scipy.sparse.csr.csr_matrix".
for b in range(4):
np.savetxt("scanorama_pancreas_v1_integrated_batch" + str(b + 1) + ".txt",
pancreas_integrated[b],
fmt='%10.9f',
delimiter="\t")
np.savetxt("scanorama_pancreas_v1_corrected_batch" + str(b + 1) + ".txt",
pancreas_corrected[b].toarray(),
fmt='%10.9f',
delimiter="\t")
np.savetxt("scanorama_pancreas_v1_genes.txt",
pancreas_genes,
fmt="%s",
delimiter="\t")
import numpy as np
import scanorama
hemat_data = np.transpose(
np.genfromtxt("../../data/count_data_hemat_v1.txt",
delimiter=" ",
dtype=int))
hemat_gene_list = ["gene_" + str(i) for i in range(1, 3471)]
hemat_gene_list = [hemat_gene_list for i in range(0, 2)]
hemat_data = [hemat_data[0:2729, 0:3470], hemat_data[2729:4649, 0:3470]]
hemat_integrated, hemat_corrected, hemat_genes = scanorama.correct(
hemat_data, hemat_gene_list, return_dimred=True)
#Please note that *_corrected is of type "scipy.sparse.csr.csr_matrix".
for b in range(2):
np.savetxt("scanorama_hemat_v1_integrated_batch" + str(b + 1) + ".txt",
hemat_integrated[b],
fmt='%10.9f',
delimiter="\t")
np.savetxt("scanorama_hemat_v1_corrected_batch" + str(b + 1) + ".txt",
hemat_corrected[b].toarray(),
fmt='%10.9f',
delimiter="\t")
np.savetxt("scanorama_hemat_v1_genes.txt",
hemat_genes,
fmt="%s",
delimiter="\t")
delimiter=" ",
dtype=int)
G = LUAD_dim[1]
B = LUAD_dim[2]
LUAD_gene_list = ["gene_" + str(i) for i in range(1, (G + 1))]
LUAD_gene_list = [LUAD_gene_list for i in range(0, B)]
offset = 0
LUAD_data = []
for b in range(B):
LUAD_data.append(LUAD_count[offset:(offset + LUAD_dim[(3 + b)]), :])
offset += LUAD_dim[(3 + b)]
LUAD_integrated, LUAD_corrected, LUAD_genes = scanorama.correct(
LUAD_data, LUAD_gene_list, return_dimred=True)
#Please note that *_corrected is of type "scipy.sparse.csr.csr_matrix".
for b in range(B):
np.savetxt("./data/scanorama_LUAD_v1_integrated_batch" + str(b + 1) +
".txt",
LUAD_integrated[b],
fmt='%10.9f',
delimiter="\t")
np.savetxt("./data/scanorama_LUAD_v1_corrected_batch" + str(b + 1) +
".txt",
LUAD_corrected[b].toarray(),
fmt='%10.9f',
delimiter="\t")
np.savetxt("scanorama_LUAD_v1_genes.txt", LUAD_genes, fmt="%s", delimiter="\t")