canergen commited on
Commit
2d4ce09
1 Parent(s): 58317ad

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +115 -40
README.md CHANGED
@@ -12,15 +12,76 @@ tags:
12
  - annotated:False
13
  ---
14
 
15
- # Description
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
16
 
17
  scVI model trained on synthetic IID data and uploaded with the minified data.
18
 
19
- # Model properties
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
20
 
21
- Many model properties are in the model tags. Some more are listed below.
22
 
23
- **model_init_params**:
 
 
 
 
 
 
 
 
 
 
 
24
  ```json
25
  {
26
  "n_hidden": 128,
@@ -33,7 +94,12 @@ Many model properties are in the model tags. Some more are listed below.
33
  }
34
  ```
35
 
36
- **model_setup_anndata_args**:
 
 
 
 
 
37
  ```json
38
  {
39
  "layer": null,
@@ -45,50 +111,59 @@ Many model properties are in the model tags. Some more are listed below.
45
  }
46
  ```
47
 
48
- **model_summary_stats**:
49
- | Summary Stat Key | Value |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
50
  |--------------------------|-------|
51
- | n_batch | 1 |
52
- | n_cells | 400 |
53
- | n_extra_categorical_covs | 0 |
54
- | n_extra_continuous_covs | 0 |
55
- | n_labels | 1 |
56
- | n_latent_qzm | 10 |
57
- | n_latent_qzv | 10 |
58
- | n_vars | 100 |
59
-
60
- **model_data_registry**:
61
- | Registry Key | scvi-tools Location |
62
- |-------------------|--------------------------------------|
63
- | X | adata.X |
64
- | batch | adata.obs['_scvi_batch'] |
65
- | labels | adata.obs['_scvi_labels'] |
66
- | latent_qzm | adata.obsm['scvi_latent_qzm'] |
67
- | latent_qzv | adata.obsm['scvi_latent_qzv'] |
68
- | minify_type | adata.uns['_scvi_adata_minify_type'] |
69
- | observed_lib_size | adata.obs['observed_lib_size'] |
70
-
71
- **model_parent_module**: scvi.model
72
-
73
- **data_is_minified**: True
74
-
75
- # Training data
76
-
77
- This is an optional link to where the training data is stored if it is too large
78
- to host on the huggingface Model hub.
79
 
80
  <!-- If your model is not uploaded with any data (e.g., minified data) on the Model Hub, then make
81
  sure to provide this field if you want users to be able to access your training data. See the
82
  scvi-tools documentation for details. -->
 
83
 
84
- Training data url: N/A
 
85
 
86
- # Training code
87
 
88
- This is an optional link to the code used to train the model.
89
 
90
- Training code url: N/A
91
 
92
  # References
93
 
94
- To be added...
 
12
  - annotated:False
13
  ---
14
 
15
+
16
+ ScVI is a variational inference model for single-cell RNA-seq data that can learn an underlying
17
+ latent space, integrate technical batches and impute dropouts.
18
+ The learned low-dimensional latent representation of the data can be used for visualization and
19
+ clustering.
20
+
21
+ scVI takes as input a scRNA-seq gene expression matrix with cells and genes.
22
+ We provide an extensive [user guide](https://docs.scvi-tools.org/en/1.2.0/user_guide/models/scvi.html).
23
+
24
+ - See our original manuscript for further details of the model:
25
+ [scVI manuscript](https://www.nature.com/articles/s41592-018-0229-2).
26
+ - See our manuscript on [scvi-hub](https://www.biorxiv.org/content/10.1101/2024.03.01.582887v2) how
27
+ to leverage pre-trained models.
28
+
29
+ This model can be used for fine tuning on new data using our Arches framework:
30
+ [Arches tutorial](https://docs.scvi-tools.org/en/1.0.0/tutorials/notebooks/scarches_scvi_tools.html).
31
+
32
+
33
+ # Model Description
34
 
35
  scVI model trained on synthetic IID data and uploaded with the minified data.
36
 
37
+ # Metrics
38
+
39
+ We provide here key performance metrics for the uploaded model, if provided by the data uploader.
40
+
41
+ <details>
42
+ <summary><strong>Coefficient of variation</strong></summary>
43
+
44
+ The cell-wise coefficient of variation summarizes how well variation between different cells is
45
+ preserved by the generated model expression. Below a squared Pearson correlation coefficient of 0.4
46
+ , we would recommend not to use generated data for downstream analysis, while the generated latent
47
+ space might still be useful for analysis.
48
+
49
+ **Cell-wise Coefficient of Variation**:
50
+
51
+ Not provided by uploader
52
+
53
+ The gene-wise coefficient of variation summarizes how well variation between different genes is
54
+ preserved by the generated model expression. This value is usually quite high.
55
+
56
+ **Gene-wise Coefficient of Variation**:
57
+
58
+ Not provided by uploader
59
+
60
+ </details>
61
+
62
+ <details>
63
+ <summary><strong>Differential expression metric</strong></summary>
64
+
65
+ The differential expression metric provides a summary of the differential expression analysis
66
+ between cell types or input clusters. We provide here the F1-score, Pearson Correlation
67
+ Coefficient of Log-Foldchanges, Spearman Correlation Coefficient, and Area Under the Precision
68
+ Recall Curve (AUPRC) for the differential expression analysis using Wilcoxon Rank Sum test for each
69
+ cell-type.
70
 
71
+ **Differential expression**:
72
 
73
+ Not provided by uploader
74
+
75
+ </details>
76
+
77
+ # Model Properties
78
+
79
+ We provide here key parameters used to setup and train the model.
80
+
81
+ <details>
82
+ <summary><strong>Model Parameters</strong></summary>
83
+
84
+ These provide the settings to setup the original model:
85
  ```json
86
  {
87
  "n_hidden": 128,
 
94
  }
95
  ```
96
 
97
+ </details>
98
+
99
+ <details>
100
+ <summary><strong>Setup Data Arguments</strong></summary>
101
+
102
+ Arguments passed to setup_anndata of the original model:
103
  ```json
104
  {
105
  "layer": null,
 
111
  }
112
  ```
113
 
114
+ </details>
115
+
116
+ <details>
117
+ <summary><strong>Data Registry</strong></summary>
118
+
119
+ Registry elements for AnnData manager:
120
+ | Registry Key | scvi-tools Location |
121
+ |--------------------------|--------------------------------------|
122
+ | X | adata.X |
123
+ | batch | adata.obs['_scvi_batch'] |
124
+ | labels | adata.obs['_scvi_labels'] |
125
+ | latent_qzm | adata.obsm['scvi_latent_qzm'] |
126
+ | latent_qzv | adata.obsm['scvi_latent_qzv'] |
127
+ | minify_type | adata.uns['_scvi_adata_minify_type'] |
128
+ | observed_lib_size | adata.obs['observed_lib_size'] |
129
+
130
+ - **Data is Minified**: True
131
+
132
+ </details>
133
+
134
+ <details>
135
+ <summary><strong>Summary Statistics</strong></summary>
136
+
137
+ | Summary Stat Key | Value |
138
  |--------------------------|-------|
139
+ | n_batch | 1 |
140
+ | n_cells | 400 |
141
+ | n_extra_categorical_covs | 0 |
142
+ | n_extra_continuous_covs | 0 |
143
+ | n_labels | 1 |
144
+ | n_latent_qzm | 10 |
145
+ | n_latent_qzv | 10 |
146
+ | n_vars | 100 |
147
+
148
+ </details>
149
+
150
+
151
+ <details>
152
+ <summary><strong>Training</strong></summary>
 
 
 
 
 
 
 
 
 
 
 
 
 
 
153
 
154
  <!-- If your model is not uploaded with any data (e.g., minified data) on the Model Hub, then make
155
  sure to provide this field if you want users to be able to access your training data. See the
156
  scvi-tools documentation for details. -->
157
+ **Training data url**: Not provided by uploader
158
 
159
+ If provided by the original uploader, for those interested in understanding or replicating the
160
+ training process, the code is available at the link below.
161
 
162
+ **Training Code URL**: Not provided by uploader
163
 
164
+ </details>
165
 
 
166
 
167
  # References
168
 
169
+ To be added...