import matplotlib.pyplot as plt
import matplotlib.lines as mlines
import seaborn as sns
from lungsc.ingest.load_dataset import DatasetLung, versions
fig_fdn = '../../figures/endomese_share/endo_paper_figure_4/'
if __name__ == '__main__':
os.makedirs(fig_fdn, exist_ok=True)
version = versions[-1]
ds0 = DatasetLung.load(preprocess=True, version=version, include_hyperoxia=True)
ds0.query_samples_by_metadata(
'(cellType == "endothelial") & (doublet == 0)',
inplace=True)
print('Total endothelial cells analyzed: {:}'.format(ds0.n_samples))
ds = ds0.query_samples_by_metadata('Treatment == "normal"')
vs = ds.samplesheet[['embed_endo_1', 'embed_endo_2']].copy()
vs.columns = ['dimension 1', 'dimension 2']
if False:
print('Plot embedding of genes')
genes = [
'Mest',
'Peg3',
'Sparcl1',
sep='\t',
header=None,
).set_index(1)
go_terms = go_terms.loc[go_terms[0] == 'Biological Process'][2]
def fun(df):
gt = go_terms.loc[df[4]]
idx = gt.str.contains('cell cycle')
#idx |= ...
return idx.sum() == 0
gby = go_long.loc[go_long[4].isin(go_terms.index)].groupby(2)
go_genes = {key for key, val in gby if fun(val)}
go_genes = sorted(go_genes)
ds = DatasetLung.load(preprocess=True, version=versions[-2])
ds.query_samples_by_metadata(
'(cellType == "immune") & (doublet == 0) & (Treatment == "normal") & (cellSubtype in ("Mac I", "Mac II", "Mac III", "Mac IV", "Mac V"))',
inplace=True)
print('Find markers')
csts = ['Mac I', 'Mac II', 'Mac III', 'Mac IV', 'Mac V']
comps = {}
for cst in csts:
print(cst)
ds.samplesheet['is_{:}'.format(cst)] = ds.samplesheet['cellSubtype'] == cst
dsp = ds.split('is_{:}'.format(cst))
dsp[True].subsample(100, inplace=True)
dsp[False].subsample(100, inplace=True)
comp = dsp[True].compare(dsp[False])
import matplotlib.pyplot as plt
import matplotlib.lines as mlines
import seaborn as sns
from lungsc.ingest.load_dataset import DatasetLung, versions
fig_fdn = '../../figures/endomese_share/endo_paper_figure_5/'
if __name__ == '__main__':
os.makedirs(fig_fdn, exist_ok=True)
version = versions[-1]
ds = DatasetLung.load(preprocess=True,
include_hyperoxia=False,
version=version)
ds.query_samples_by_metadata(
'(cellType == "endothelial") & (Treatment == "normal") & (doublet == 0)',
inplace=True)
ds.samplesheet['is_prolif'] = ds.samplesheet[
'cellSubtype'] == 'Proliferative EC'
#print('Load tSNE from file')
#vs = pd.read_csv(
# '../../data/sequencing/datasets/all_{:}/tsne_with_hyperoxia_endo.tsv'.format(version),
# sep='\t',
# index_col=0,
# )
#vs = vs.loc[ds.samplenames]
pa = argparse.ArgumentParser()
pa.add_argument(
'--cellType',
choices=['immune', 'general'],
required=True,
)
pa.add_argument(
'--includeNoCellType',
action='store_true',
)
args = pa.parse_args()
version = versions[-1]
ds = DatasetLung.load(version=version, include_doublets=True)
ds.samplesheet.loc[ds.samplesheet['doublet'] == 1,
'cellSubtype'] = 'doublet?'
if args.cellType == 'general':
print('Plot dimensionality reduction of dataset')
print('Feature selection')
features = ds.feature_selection.overdispersed_within_groups(
'Mousename',
inplace=False,
)
dsf = ds.query_features_by_name(features)
print('PCA')
dsc = dsf.dimensionality.pca(
n_dims=30,
import glob
import gzip
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
# Ensure leidenalg is correct
sys.path.insert(0, os.path.abspath('../../packages'))
from lungsc.ingest.load_dataset import versions, DatasetLung
if __name__ == '__main__':
version = versions[-1]
ds = DatasetLung.load(version=version)
print('Feature selection')
features = ds.feature_selection.overdispersed_within_groups('Mousename',
inplace=False)
dsf = ds.query_features_by_name(features)
print('PCA')
dsc = dsf.dimensionality.pca(n_dims=30,
robust=False,
return_dataset='samples')
#print('UMAP')
#vsu = dsc.dimensionality.umap()
print('tSNE')