Metadata-Version: 2.1
Name: genesurf
Version: 0.1
Summary: A self-supervised deep learning method for reference-free deconvolution.
Home-page: https://github.com/lllsssyyyy/SURF
Author: Shuyu Liang
Author-email: 
Classifier: Programming Language :: Python :: 3
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: pandas==1.5.3
Requires-Dist: numpy==1.23.0
Requires-Dist: scanpy==1.9.8
Requires-Dist: scipy==1.9.1
Requires-Dist: torch_sparse
Requires-Dist: torch_scatter
Requires-Dist: torch_scatter
Requires-Dist: torch_geometric
Requires-Dist: rpy2
Requires-Dist: tables
Requires-Dist: seaborn==0.13.2
Requires-Dist: matplotlib==3.7.1

# SURF
A self-supervised deep learning method for reference-free deconvolution. The overall approach is detailed in the official paper out in xxx.

![Fig1](https://github.com/user-attachments/assets/cd371dab-fa9a-474d-9bfa-32b41adb8cbe)

# Data input  
df_expr: (n_spots * n_genes), dataframe, with column names (gene names)  
df_pos: (n_spots * 2), dataframe, with column names [‘x’, ‘y’]  
barcodes: (n_spots,), a numpy array  
  
# Data output     
‘pred.csv’: The predicted cell types proportions in each spot.  
‘beta.csv’: The deconvolved gene expressions of each cell type.  
‘last.pkl’: The saved trained model. 

# Installation


