def ds_ds2():
from singlet.dataset import Dataset
ds = Dataset(samplesheet='example_sheet_tsv',
counts_table='example_table_tsv')
ds2 = ds.copy()
ds.samplesheet = ds.samplesheet.iloc[:2]
ds2.samplesheet = ds2.samplesheet.iloc[2:]
return (ds, ds2)
author: Fabio Zanini
date: 07/08/17
content: Test Dataset class.
'''
import numpy as np
# Script
if __name__ == '__main__':
# NOTE: an env variable for the config file needs to be set when
# calling this script
from singlet.dataset import Dataset
ds = Dataset(samplesheet='example_sheet_tsv',
counts_table='example_table_tsv')
ds2 = ds.copy()
ds.samplesheet = ds.samplesheet.iloc[:2]
ds2.samplesheet = ds2.samplesheet.iloc[2:]
print('Bootstrap')
dsboot = ds.bootstrap()
assert ('--sampling_' in dsboot.samplenames[0])
print('Done!')
print('Test feature comparison (Mann-Whitney U)')
pvals = ds.compare(ds2, method='mann-whitney')
assert (np.isclose(pvals.values.min(), 0.193931))
print('Done!')
print('Test feature comparison (Kolmogorov-Smirnov)')
pvals = ds.compare(ds2, method='kolmogorov-smirnov')
assert (np.isclose(pvals.values.min(), 0.097027))
if __name__ == '__main__':
ds = Dataset(
samplesheet='dengue',
counts_table='dengue',
featuresheet='humanGC38',
)
data_snv = xr.open_dataset('../bigdata/allele_frequencies.nc')
#ds.counts = CountsTable(data['aaf'].to_pandas().fillna(0))
# Sync with Felix metadata
with open('../data/metadataD_SNV_with_tsne_and_tsneSNV.pkl', 'rb') as ff:
metadata_felix = pickle.load(ff)
samples = metadata_felix.index[(
~np.isnan(metadata_felix[['Dn', 'Ds']])).all(axis=1)]
ds.samplesheet = ds.samplesheet.loc[samples]
metadata_felix = metadata_felix.loc[samples]
for col in [
'coverage', 'Ds', 'Dn', 'depth', 'numSNV', 'Dn_s', 'tsne1_MOI1_10',
'tsne2_MOI1_10', 'tsne1_SNV', 'tsne2_SNV', 'clusterN_SNV'
]:
ds.samplesheet[col] = metadata_felix[col]
ds.samplesheet['log_Dn'] = np.log10(1e-6 + ds.samplesheet['Dn'])
ds.samplesheet['log_Ds'] = np.log10(1e-6 + ds.samplesheet['Ds'])
cov = ds.samplesheet['coverage']
n = ds.samplesheet['numberDengueReads'].astype(int)
ds.samplesheet['virus_reads_per_million'] = 1e6 * n / (cov + n)
ds.samplesheet['log_virus_reads_per_million'] = np.log10(0.1 + 1e6 * n /
(cov + n))
# htseq-count output, below is the explanation how I solved this.
# velocyto skips quasi-empty BAM files, final order is lexicographic as in:
# echo '' > ~/subfolders.tsv; for fdn in 10017006*; do si=$(du $fdn/star/Aligned.out.possorted.bam | cut -f1); if [ $si -ge "30" ]; then echo $fdn >> ~/subfolders.tsv; fi; done
cellnames = pd.read_csv('../bigdata/rnavelocity_cellnames.tsv', header=None).values[:, 0]
vlm = vcy.VelocytoLoom("../bigdata/rna_velocity.loom")
vlm.ca['CellID'] = cellnames
print('Load normal counts and metadata and sync them with the velocity results')
# Load and sync external metadata
ds = Dataset(
#counts_table='dengue',
samplesheet='dengue',
)
with open('../data/metadataD_SNV_with_tsne_and_tsneSNV.pkl', 'rb') as ff:
metadata_felix = pickle.load(ff)
ds.samplesheet = ds.samplesheet.loc[cellnames]
metadata_felix = metadata_felix.loc[cellnames]
vlm.ca['ClusterName'] = metadata_felix['clusterN_SNV'].fillna(6).values
vlm.set_clusters(vlm.ca["ClusterName"])
ds.samplesheet['clusterN_SNV'] = metadata_felix['clusterN_SNV'].fillna(6)
ds.samplesheet['coverage'] = metadata_felix['coverage']
ds.samplesheet['virus_reads_per_million'] = 1.0 * 1e6 * ds.samplesheet['numberDengueReads'] / (ds.samplesheet['numberDengueReads'] + ds.samplesheet['coverage'])
ds.samplesheet['log_virus_reads_per_million'] = np.log10(0.1 + ds.samplesheet['virus_reads_per_million'])
print('Filter cells, genes, etc. using the velocity tutorial')
vlm.normalize("S", size=True, log=True)
vlm.filter_cells(bool_array=vlm.initial_Ucell_size > np.percentile(vlm.initial_Ucell_size, 0.5))
ds.samplesheet = ds.samplesheet.loc[vlm.ca['CellID']]
metadata_felix = metadata_felix.loc[vlm.ca['CellID']]
