gnodevae 0.0.3__tar.gz → 0.0.4__tar.gz
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- {gnodevae-0.0.3 → gnodevae-0.0.4}/GNODEVAE/__init__.py +1 -1
- {gnodevae-0.0.3/gnodevae.egg-info → gnodevae-0.0.4}/PKG-INFO +11 -11
- {gnodevae-0.0.3 → gnodevae-0.0.4}/README.md +10 -10
- {gnodevae-0.0.3 → gnodevae-0.0.4/gnodevae.egg-info}/PKG-INFO +11 -11
- {gnodevae-0.0.3 → gnodevae-0.0.4}/pyproject.toml +1 -1
- {gnodevae-0.0.3 → gnodevae-0.0.4}/GNODEVAE/CLEAR.py +0 -0
- {gnodevae-0.0.3 → gnodevae-0.0.4}/GNODEVAE/GODEVAE.py +0 -0
- {gnodevae-0.0.3 → gnodevae-0.0.4}/GNODEVAE/GODEVAE_agent.py +0 -0
- {gnodevae-0.0.3 → gnodevae-0.0.4}/GNODEVAE/GODEVAE_env.py +0 -0
- {gnodevae-0.0.3 → gnodevae-0.0.4}/GNODEVAE/GODEVAE_trainer.py +0 -0
- {gnodevae-0.0.3 → gnodevae-0.0.4}/GNODEVAE/SC3.py +0 -0
- {gnodevae-0.0.3 → gnodevae-0.0.4}/GNODEVAE/agent.py +0 -0
- {gnodevae-0.0.3 → gnodevae-0.0.4}/GNODEVAE/env.py +0 -0
- {gnodevae-0.0.3 → gnodevae-0.0.4}/GNODEVAE/mixin.py +0 -0
- {gnodevae-0.0.3 → gnodevae-0.0.4}/GNODEVAE/module.py +0 -0
- {gnodevae-0.0.3 → gnodevae-0.0.4}/GNODEVAE/module_refined.py +0 -0
- {gnodevae-0.0.3 → gnodevae-0.0.4}/GNODEVAE/partition_time.py +0 -0
- {gnodevae-0.0.3 → gnodevae-0.0.4}/GNODEVAE/scGCC.py +0 -0
- {gnodevae-0.0.3 → gnodevae-0.0.4}/GNODEVAE/scGNN.py +0 -0
- {gnodevae-0.0.3 → gnodevae-0.0.4}/GNODEVAE/trainer.py +0 -0
- {gnodevae-0.0.3 → gnodevae-0.0.4}/GNODEVAE/utils.py +0 -0
- {gnodevae-0.0.3 → gnodevae-0.0.4}/GNODEVAE/utils_ODE.py +0 -0
- {gnodevae-0.0.3 → gnodevae-0.0.4}/LICENSE +0 -0
- {gnodevae-0.0.3 → gnodevae-0.0.4}/MANIFEST.in +0 -0
- {gnodevae-0.0.3 → gnodevae-0.0.4}/gnodevae.egg-info/SOURCES.txt +0 -0
- {gnodevae-0.0.3 → gnodevae-0.0.4}/gnodevae.egg-info/dependency_links.txt +0 -0
- {gnodevae-0.0.3 → gnodevae-0.0.4}/gnodevae.egg-info/requires.txt +0 -0
- {gnodevae-0.0.3 → gnodevae-0.0.4}/gnodevae.egg-info/top_level.txt +0 -0
- {gnodevae-0.0.3 → gnodevae-0.0.4}/setup.cfg +0 -0
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
Metadata-Version: 2.4
|
|
2
2
|
Name: gnodevae
|
|
3
|
-
Version: 0.0.
|
|
3
|
+
Version: 0.0.4
|
|
4
4
|
Summary: A Graph-Based ODE-VAE Enhances Clustering for Single-Cell Data
|
|
5
5
|
Author-email: Zeyu Fu <fuzeyu99@126.com>
|
|
6
6
|
License: MIT License
|
|
@@ -61,35 +61,35 @@ Dynamic: license-file
|
|
|
61
61
|
|
|
62
62
|
## Introduction
|
|
63
63
|
|
|
64
|
-
GNODEVAE is
|
|
64
|
+
GNODEVAE is a computational framework that integrates Graph Attention Networks (GAT), Neural Ordinary Differential Equations (NODE), and Variational Autoencoders (VAE). It targets three common challenges in single-cell RNA sequencing data analysis:
|
|
65
65
|
|
|
66
66
|
1. Capturing complex topological relationships between cells
|
|
67
67
|
2. Modeling continuous dynamic processes of cell differentiation
|
|
68
68
|
3. Handling high levels of technical noise and biological variation
|
|
69
69
|
|
|
70
|
-
|
|
70
|
+
The integration is intended to support identification of cell subpopulations, reconstruction of developmental trajectories, and analysis of cellular heterogeneity.
|
|
71
71
|
|
|
72
72
|
## Key Contributions
|
|
73
73
|
|
|
74
74
|
### 1. Dynamic Attention Weighting for Biological Significance
|
|
75
75
|
|
|
76
|
-
The GAT
|
|
76
|
+
The GAT attention mechanism adaptively weights gene expression profiles, prioritizing biological relationships while reducing technical noise in heterogeneous cell populations.
|
|
77
77
|
|
|
78
78
|
### 2. Continuous-Time Developmental Modeling via Neural ODEs
|
|
79
79
|
|
|
80
|
-
|
|
80
|
+
Neural ordinary differential equations transform static representations into dynamic systems, with time variables providing a continuous parameterization of developmental processes.
|
|
81
81
|
|
|
82
82
|
### 3. Biologically Consistent Latent Space Representations
|
|
83
83
|
|
|
84
|
-
The model's latent space
|
|
84
|
+
The model's latent space is designed to capture variation in cell differentiation rates, while attention weights can be inspected alongside known developmental relationships between cell types.
|
|
85
85
|
|
|
86
|
-
### 4.
|
|
86
|
+
### 4. Benchmarking Summary
|
|
87
87
|
|
|
88
|
-
|
|
88
|
+
The accompanying study compares GNODEVAE with several single-cell analysis methods, including scVI, DIP-VAE, TC-VAE, beta-VAE, Info-VAE, and scTour, across 13 datasets.
|
|
89
89
|
|
|
90
|
-
### 5.
|
|
90
|
+
### 5. Gene Trend Analysis
|
|
91
91
|
|
|
92
|
-
|
|
92
|
+
The accompanying study also reports stronger Calinski-Harabasz scores than selected baselines for gene trend analysis.
|
|
93
93
|
|
|
94
94
|
|
|
95
95
|
## Installation
|
|
@@ -307,7 +307,7 @@ If you use GNODEVAE in your research, please cite:
|
|
|
307
307
|
```bibtex
|
|
308
308
|
@article{fu2025gnodevae,
|
|
309
309
|
title={GNODEVAE: a graph-based ODE-VAE enhances clustering for single-cell data},
|
|
310
|
-
author={Fu,
|
|
310
|
+
author={Fu, Zeyu and Chen, Chunlin and Wang, Song and Wang, Junping and Chen, Shilei},
|
|
311
311
|
journal={BMC Genomics},
|
|
312
312
|
volume={26},
|
|
313
313
|
pages={767},
|
|
@@ -6,35 +6,35 @@
|
|
|
6
6
|
|
|
7
7
|
## Introduction
|
|
8
8
|
|
|
9
|
-
GNODEVAE is
|
|
9
|
+
GNODEVAE is a computational framework that integrates Graph Attention Networks (GAT), Neural Ordinary Differential Equations (NODE), and Variational Autoencoders (VAE). It targets three common challenges in single-cell RNA sequencing data analysis:
|
|
10
10
|
|
|
11
11
|
1. Capturing complex topological relationships between cells
|
|
12
12
|
2. Modeling continuous dynamic processes of cell differentiation
|
|
13
13
|
3. Handling high levels of technical noise and biological variation
|
|
14
14
|
|
|
15
|
-
|
|
15
|
+
The integration is intended to support identification of cell subpopulations, reconstruction of developmental trajectories, and analysis of cellular heterogeneity.
|
|
16
16
|
|
|
17
17
|
## Key Contributions
|
|
18
18
|
|
|
19
19
|
### 1. Dynamic Attention Weighting for Biological Significance
|
|
20
20
|
|
|
21
|
-
The GAT
|
|
21
|
+
The GAT attention mechanism adaptively weights gene expression profiles, prioritizing biological relationships while reducing technical noise in heterogeneous cell populations.
|
|
22
22
|
|
|
23
23
|
### 2. Continuous-Time Developmental Modeling via Neural ODEs
|
|
24
24
|
|
|
25
|
-
|
|
25
|
+
Neural ordinary differential equations transform static representations into dynamic systems, with time variables providing a continuous parameterization of developmental processes.
|
|
26
26
|
|
|
27
27
|
### 3. Biologically Consistent Latent Space Representations
|
|
28
28
|
|
|
29
|
-
The model's latent space
|
|
29
|
+
The model's latent space is designed to capture variation in cell differentiation rates, while attention weights can be inspected alongside known developmental relationships between cell types.
|
|
30
30
|
|
|
31
|
-
### 4.
|
|
31
|
+
### 4. Benchmarking Summary
|
|
32
32
|
|
|
33
|
-
|
|
33
|
+
The accompanying study compares GNODEVAE with several single-cell analysis methods, including scVI, DIP-VAE, TC-VAE, beta-VAE, Info-VAE, and scTour, across 13 datasets.
|
|
34
34
|
|
|
35
|
-
### 5.
|
|
35
|
+
### 5. Gene Trend Analysis
|
|
36
36
|
|
|
37
|
-
|
|
37
|
+
The accompanying study also reports stronger Calinski-Harabasz scores than selected baselines for gene trend analysis.
|
|
38
38
|
|
|
39
39
|
|
|
40
40
|
## Installation
|
|
@@ -252,7 +252,7 @@ If you use GNODEVAE in your research, please cite:
|
|
|
252
252
|
```bibtex
|
|
253
253
|
@article{fu2025gnodevae,
|
|
254
254
|
title={GNODEVAE: a graph-based ODE-VAE enhances clustering for single-cell data},
|
|
255
|
-
author={Fu,
|
|
255
|
+
author={Fu, Zeyu and Chen, Chunlin and Wang, Song and Wang, Junping and Chen, Shilei},
|
|
256
256
|
journal={BMC Genomics},
|
|
257
257
|
volume={26},
|
|
258
258
|
pages={767},
|
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
Metadata-Version: 2.4
|
|
2
2
|
Name: gnodevae
|
|
3
|
-
Version: 0.0.
|
|
3
|
+
Version: 0.0.4
|
|
4
4
|
Summary: A Graph-Based ODE-VAE Enhances Clustering for Single-Cell Data
|
|
5
5
|
Author-email: Zeyu Fu <fuzeyu99@126.com>
|
|
6
6
|
License: MIT License
|
|
@@ -61,35 +61,35 @@ Dynamic: license-file
|
|
|
61
61
|
|
|
62
62
|
## Introduction
|
|
63
63
|
|
|
64
|
-
GNODEVAE is
|
|
64
|
+
GNODEVAE is a computational framework that integrates Graph Attention Networks (GAT), Neural Ordinary Differential Equations (NODE), and Variational Autoencoders (VAE). It targets three common challenges in single-cell RNA sequencing data analysis:
|
|
65
65
|
|
|
66
66
|
1. Capturing complex topological relationships between cells
|
|
67
67
|
2. Modeling continuous dynamic processes of cell differentiation
|
|
68
68
|
3. Handling high levels of technical noise and biological variation
|
|
69
69
|
|
|
70
|
-
|
|
70
|
+
The integration is intended to support identification of cell subpopulations, reconstruction of developmental trajectories, and analysis of cellular heterogeneity.
|
|
71
71
|
|
|
72
72
|
## Key Contributions
|
|
73
73
|
|
|
74
74
|
### 1. Dynamic Attention Weighting for Biological Significance
|
|
75
75
|
|
|
76
|
-
The GAT
|
|
76
|
+
The GAT attention mechanism adaptively weights gene expression profiles, prioritizing biological relationships while reducing technical noise in heterogeneous cell populations.
|
|
77
77
|
|
|
78
78
|
### 2. Continuous-Time Developmental Modeling via Neural ODEs
|
|
79
79
|
|
|
80
|
-
|
|
80
|
+
Neural ordinary differential equations transform static representations into dynamic systems, with time variables providing a continuous parameterization of developmental processes.
|
|
81
81
|
|
|
82
82
|
### 3. Biologically Consistent Latent Space Representations
|
|
83
83
|
|
|
84
|
-
The model's latent space
|
|
84
|
+
The model's latent space is designed to capture variation in cell differentiation rates, while attention weights can be inspected alongside known developmental relationships between cell types.
|
|
85
85
|
|
|
86
|
-
### 4.
|
|
86
|
+
### 4. Benchmarking Summary
|
|
87
87
|
|
|
88
|
-
|
|
88
|
+
The accompanying study compares GNODEVAE with several single-cell analysis methods, including scVI, DIP-VAE, TC-VAE, beta-VAE, Info-VAE, and scTour, across 13 datasets.
|
|
89
89
|
|
|
90
|
-
### 5.
|
|
90
|
+
### 5. Gene Trend Analysis
|
|
91
91
|
|
|
92
|
-
|
|
92
|
+
The accompanying study also reports stronger Calinski-Harabasz scores than selected baselines for gene trend analysis.
|
|
93
93
|
|
|
94
94
|
|
|
95
95
|
## Installation
|
|
@@ -307,7 +307,7 @@ If you use GNODEVAE in your research, please cite:
|
|
|
307
307
|
```bibtex
|
|
308
308
|
@article{fu2025gnodevae,
|
|
309
309
|
title={GNODEVAE: a graph-based ODE-VAE enhances clustering for single-cell data},
|
|
310
|
-
author={Fu,
|
|
310
|
+
author={Fu, Zeyu and Chen, Chunlin and Wang, Song and Wang, Junping and Chen, Shilei},
|
|
311
311
|
journal={BMC Genomics},
|
|
312
312
|
volume={26},
|
|
313
313
|
pages={767},
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|
|
File without changes
|