convert_to_anndata.py

import scanpy as sc
import pandas as pd

data = pd.read_csv('path_to_file.csv')
feature_annotations = pd.read_csv('path_to_annotations.csv')

# Make sure samples are rows and features are columns. if not data = data.T
adata = sc.AnnData(data, obs=sample_names, var=feature_annotations)

sc.pp.pca(adata, n_comps=10)
sc.pp.neighbors(adata, n_neighbors = 10, n_pcs = 10)

plot_dewakss_results.py

import matplotlib.pyplot as plt

fig, ax = dewaxer.plot_global_performance()
fig.savefig('global_performance.svg', format='svg')
fig, ax = dewaxer.plot_local_performance()
fig.savefig('local_performance.svg', format='svg')

fig = plt.figure(figsize=(6,4), constrained_layout=True)
ax = fig.subplots(1, 1)
sc.pl.embedding(ddata, basis='umap', color='leiden', edges=True, ax=ax)
fig.savefig('embedding_clusters.svg', format='svg')

run_basic_dewakss.py

from dewakss import denoise as dewakss

# Load data:

# Rin dewakss
neigbours = list(range(3,10))
npcss = list(range(3,10))

dewaxer = dewakss.DEWAKSS(adata, n_neighbors=neigbours, n_pcs=npcss, use_global_err=False)
dewaxer.fit(adata)
ddata = dewaxer.transform(adata, copy=True)

sc.tl.leiden(ddata, resolution = 1)
sc.tl.umap(ddata, min_dist = 0.01)