biology genomics single-cell model_cls_name:SCANVI scvi_version:0.20.3 anndata_version:0.8.0 modality:rna tissue:nose tissue:respiratory airway tissue:lung parenchyma annotated:True

Description

The first integrated, universal transcriptomic reference of the human lung on the single-cell level. For more details, see https: //github.com/LungCellAtlas/HLCA.

Model properties

Many model properties are in the model tags. Some more are listed below.

model_init_params:

{
    "n_hidden": 128,
    "n_latent": 30,
    "n_layers": 2,
    "dropout_rate": 0.1,
    "dispersion": "gene",
    "gene_likelihood": "nb",
    "encode_covariates": true,
    "deeply_inject_covariates": false,
    "use_layer_norm": "both",
    "use_batch_norm": "none"
}

model_setup_anndata_args:

{
    "labels_key": "scanvi_label",
    "unlabeled_category": "unlabeled",
    "layer": null,
    "batch_key": "dataset",
    "size_factor_key": null,
    "categorical_covariate_keys": null,
    "continuous_covariate_keys": null
}

model_summary_stats:

Summary Stat Key Value
n_batch 14
n_cells 584884
n_extra_categorical_covs 0
n_extra_continuous_covs 0
n_labels 29
n_latent_qzm 30
n_latent_qzv 30
n_vars 2000

model_data_registry:

Registry Key scvi-tools Location
X adata.X
batch adata.obs['_scvi_batch']
labels adata.obs['_scvi_labels']
latent_qzm adata.obsm['_scanvi_latent_qzm']
latent_qzv adata.obsm['_scanvi_latent_qzv']
minify_type adata.uns['_scvi_adata_minify_type']
observed_lib_size adata.obs['_scanvi_observed_lib_size']

model_parent_module: scvi.model

data_is_minified: True

Training data

This is an optional link to where the training data is stored if it is too large to host on the huggingface Model hub.

<!-- If your model is not uploaded with any data (e.g., minified data) on the Model Hub, then make sure to provide this field if you want users to be able to access your training data. See the scvi-tools documentation for details. -->

Training data url: https://cellxgene.cziscience.com/e/066943a2-fdac-4b29-b348-40cede398e4e.cxg/

Training code

This is an optional link to the code used to train the model.

Training code url: https://github.com/LungCellAtlas/HLCA_reproducibility

References

An integrated cell atlas of the human lung in health and disease. L Sikkema, D Strobl, L Zappia, E Madissoon, NS Markov, L Zaragosi, M Ansari, M Arguel, L Apperloo, C Bécavin, M Berg, E Chichelnitskiy, M Chung, A Collin, ACA Gay, B Hooshiar Kashani, M Jain, T Kapellos, TM Kole, C Mayr, M von Papen, L Peter, C Ramírez-Suástegui, J Schniering, C Taylor, T Walzthoeni, C Xu, LT Bui, C de Donno, L Dony, M Guo, AJ Gutierrez, L Heumos, N Huang, I Ibarra, N Jackson, P Kadur Lakshminarasimha Murthy, M Lotfollahi, T Tabib, C Talavera-Lopez, K Travaglini, A Wilbrey-Clark, KB Worlock, M Yoshida, Lung Biological Network Consortium, T Desai, O Eickelberg, C Falk, N Kaminski, M Krasnow, R Lafyatis, M Nikolíc, J Powell, J Rajagopal, O Rozenblatt-Rosen, MA Seibold, D Sheppard, D Shepherd, SA Teichmann, A Tsankov, J Whitsett, Y Xu, NE Banovich, P Barbry, TE Duong, KB Meyer, JA Kropski, D Pe’er, HB Schiller, PR Tata, JL Schultze, AV Misharin, MC Nawijn, MD Luecken, F Theis. bioRxiv 2022.03.10.483747; doi: https: //doi.org/10.1101/2022.03.10.483747