DEPair evaluates ligand-receptor mediated crosstalks among cell populations. A interaction score as well as its significance is given for each pair of clusters with DEPair. DEPair also identifies the variable interactions between different groups.
You can install DEPair via pip install depair.
import numpy as np,scanpy as sc,anndata as ad,pandas as pd
import os,pickle,random
import matplotlib.pyplot as plt
from dfply import *
merged= sc.read_h5ad("./merged.h5ad")
adata_normal = merged[ merged.obs.status == "normal" ]
adata_degenerated = merged[ merged.obs.status == "abnormal" ]
LRDB=pd.read_table("./CellTalkDB_human_lr_pair.txt")
import depair as dp
from depair.eval import crosstalk
crosstalk_normal=crosstalk(anndata=adata_normal,
LRDB_table=LRDB,
groupby="ann210819",
verbose=True, pgb_notebook=True,
n_perm=1000, n_jobs=48)
crosstalk_degenerated=crosstalk(anndata=adata_degenerated,
LRDB_table=LRDB,
groupby="ann210819",
verbose=True, pgb_notebook=True,
n_perm=1000, n_jobs=48)
- crosstalk intensity evaluation
- permutation-based crosstalk significance calculation
- permutation speed up
- human and mouse ligand-receptor databases
- ligand-receptor databases for other species
- a tutorial as well as a document online at readthedocs
- upload to pypi
- spatial transcriptomics support
- multi subunit analysis support
- ligand-receptor database with detailed annotation
- clustering granularity selection
- visualization
- scanpy support
- supergraph exploration
- 2021-04-08 prerelease test version online
- 2021-04-30 Qijin Yin contributed the permutation speed up codes
- 2022-12-23 DEPair is available at pypi (https://pypi.org/project/depair/).
- 2023-04-12 add
eval_significance=Falseoption in crosstalk() to skip the significance evaluation step.
contributors: Sijie Chen, Qijin Yin