annotate/scanvi

Description

Semi-supervised model for single-cell transcriptomics data. A scVI extension that can leverage the cell type knowledge for a subset of the cells present in the data sets to infer the states of the rest of the cells.

Inputs

Name	Type & Properties	Description
--input -i	file required	Input h5mu file.
--modality	string	Which modality to process.
--var_input_gene_names	string	.var field containing the gene names, if the .var index is not to be used.
--input_reference_gene_overlap	integer	The minimum number of genes present in both the reference and query datasets.

Name

Type & Properties

--input

-i

file

required

--modality

string

--var_input_gene_names

string

--input_reference_gene_overlap

integer

Reference model

Name	Type & Properties	Description
--scvi_reference_model	file	Pretrained SCVI reference model to initialize the SCANVI model with. The model needs to include the AnnData object used to trained the model stored.
--scanvi_reference_model	file	Pretrained SCANVI reference model.

Name

Type & Properties

--scvi_reference_model

file

--scanvi_reference_model

file

SCANVI reference model training arguments

Name	Type & Properties	Description
--reference_train_size	double	Size of training set.
--reference_max_epochs	integer	Maximum number of epochs.
--reference_learning_rate	double	Learning rate.
--reference_reduce_lr_on_plateau	boolean	Reduce learning rate on plateau.
--reference_lr_patience	integer	Patience for learning rate reduction.
--reference_lr_factor	double	Factor by which to reduce learning rate.
--reference_early_stopping	boolean	Early stopping.
--reference_early_stopping_patience	integer	Patience for early stopping.

Name

Type & Properties

--reference_train_size

double

--reference_max_epochs

integer

--reference_learning_rate

double

--reference_reduce_lr_on_plateau

boolean

--reference_lr_patience

integer

--reference_lr_factor

double

--reference_early_stopping

boolean

--reference_early_stopping_patience

integer

SCANVI query model training arguments

Name	Type & Properties	Description
--query_train_size	double	Size of training set.
--query_max_epochs	integer	Maximum number of epochs.
--query_learning_rate	double	Learning rate.
--query_reduce_lr_on_plateau	boolean	Reduce learning rate on plateau.
--query_lr_patience	integer	Patience for learning rate reduction.
--query_lr_factor	double	Factor by which to reduce learning rate.
--query_early_stopping	boolean	Early stopping.
--query_early_stopping_patience	integer	Patience for early stopping.

Name

Type & Properties

--query_train_size

double

--query_max_epochs

integer

--query_learning_rate

double

--query_reduce_lr_on_plateau

boolean

--query_lr_patience

integer

--query_lr_factor

double

--query_early_stopping

boolean

--query_early_stopping_patience

integer

Outputs

Name	Type & Properties	Description
--output	file required output	Output h5mu file.
--output_compression	string
--output_model	file output	Folder where the state of the trained model will be saved to.
--output_obs_predictions	string	In which `.obs` slots to store the predicted information.
--output_obs_probability	string	In which `.obs` slots to store the probability of the predictions.
--output_obsm_scanvi_embedding	string	In which `.obsm` slots to store the scvi embedding.
--unknown_celltype	string	Label for unknown cell types.

Name

Type & Properties

--output

file

required

output

--output_compression

string

--output_model

file

output

--output_obs_predictions

string

--output_obs_probability

string

--output_obsm_scanvi_embedding

string

--unknown_celltype

string

Run this component

Run the following command to execute this component with Nextflow:

cat > params.yaml <<'EOM'  
modality: [ "rna" ]  
input_reference_gene_overlap: [ 100 ]  
reference_train_size: [ 0.9 ]  
reference_max_epochs: [ 400 ]  
reference_learning_rate: [ 0.001 ]  
reference_reduce_lr_on_plateau: [ true ]  
reference_lr_patience: [ 25 ]  
reference_lr_factor: [ 0.5 ]  
reference_early_stopping: [ true ]  
reference_early_stopping_patience: [ 50 ]  
query_train_size: [ 0.9 ]  
query_max_epochs: [ 400 ]  
query_learning_rate: [ 0.001 ]  
query_reduce_lr_on_plateau: [ true ]  
query_lr_patience: [ 25 ]  
query_lr_factor: [ 0.5 ]  
query_early_stopping: [ true ]  
query_early_stopping_patience: [ 50 ]  
output: "$id.$key.output.h5mu"  
output_model: "$id.$key.output_model.model_dir"  
output_obs_predictions: [ "scanvi_pred" ]  
output_obs_probability: [ "scanvi_probability" ]  
output_obsm_scanvi_embedding: [ "scanvi_embedding" ]  
unknown_celltype: [ "Unknown" ]  
id: "run"  
publish_dir: "output/"  
EOM

nextflow run https://packages.viash-hub.com/vsh/openpipeline.git \  
  -revision 2.0.0 \  
  -main-script target/nextflow/annotate/scanvi/main.nf \  
  -params-file params.yaml

annotate/scanvi

Inputs

Reference model

SCANVI reference model training arguments

SCANVI query model training arguments

Outputs

Run this component

Relationships

Used by

Uses