filter/filter_with_quantile

Description

Filter observations or features from an .h5mu file based on quantile thresholds of numeric values.

This component removes observations or features that fall outside specified quantile ranges in .obs or .var. columns.

For example, you can filter out observations with total UMI counts in the bottom 5% (quantile < 0.05) or
top 5% (quantile > 0.95), or retain an intermediate range (0.1 < quantile < 0.9).

Type

python_script

License

MIT

Links

Repository

Contributors

Launch

Run this component

Run the following command to execute this component with Nextflow:

cat > params.yaml <<'EOM'  
modality: [ "rna" ]  
obs_log1p_transform: [ true ]  
var_log1p_transform: [ true ]  
min_values_for_quantile: [ 10 ]  
output: "$id.$key.output.h5mu"  
obs_name_filter: [ "filter_with_quantile" ]  
var_name_filter: [ "filter_with_quantile" ]  
id: "run"  
publish_dir: "output/"  
EOM

nextflow run https://packages.viash-hub.com/vsh/openpipeline.git \  
  -revision v4.1.0 \  
  -main-script target/nextflow/filter/filter_with_quantile/main.nf \  
  -params-file params.yaml

Inputs

Name	Type & Properties	Description
--input	file required	Input h5mu file
--modality	string	Which modality from the input MuData file to process.
--obs_column	string	Name of the column in .obs to use for quantile-based filtering. If not specified, no .obs filtering will be performed.
--var_column	string	Name of the column in .var to use for quantile-based filtering. If not specified, no .var filtering will be performed.

Name

Type & Properties

--input

file

required

--modality

string

--obs_column

string

--var_column

string

Filters

Name	Type & Properties	Description
--do_subset	boolean_true	Whether to subset before storing the output.
--obs_log1p_transform	boolean	Apply log1p transformation to the obs_column values before calculating quantiles. The transformed values will be stored as 'log1p_{obs_column}' in .obs. This is useful for highly skewed data where log transformation can improve quantile-based filtering.
--obs_min_quantile	double	Minimum quantile threshold for .obs filtering. Observations with values below this quantile will be removed. Value should be between 0.0 and 1.0.
--obs_max_quantile	double	Maximum quantile threshold for .obs filtering. Observations with values above this quantile will be removed. Value should be between 0.0 and 1.0.
--var_log1p_transform	boolean	Apply log1p transformation to the var_column values before calculating quantiles. The transformed values will be stored as 'log1p_{var_column}' in .var. This is useful for highly skewed data where log transformation can improve quantile-based filtering.
--var_min_quantile	double	Minimum quantile threshold for .var filtering. Features with values below this quantile will be removed. Value should be between 0.0 and 1.0.
--var_max_quantile	double	Maximum quantile threshold for .var filtering. Features with values above this quantile will be removed. Value should be between 0.0 and 1.0.
--min_values_for_quantile	integer	Minimum number of values required to apply quantile filtering. If fewer values are present, quantile filtering will be skipped to prevent data loss.

Name

Type & Properties

--do_subset

boolean_true

--obs_log1p_transform

boolean

--obs_min_quantile

double

--obs_max_quantile

double

--var_log1p_transform

boolean

--var_min_quantile

double

--var_max_quantile

double

--min_values_for_quantile

integer

Outputs

Name	Type & Properties	Description
--output	file output	Output h5mu file.
--obs_name_filter	string	In which .obs slot to store a boolean array corresponding to which observations should be removed.
--obs_log1p_column	string	Name for the column where log1p transformed obs values will be stored. If not provided, defaults to 'log1p_{obs_column}'.
--var_log1p_column	string	Name for the column where log1p transformed var values will be stored. If not provided, defaults to 'log1p_{var_column}'.
--var_name_filter	string	In which .var slot to store a boolean array corresponding to which variables should be removed.
--output_compression	string	Compression format to use for the output AnnData and/or Mudata H5 files. By default no compression is applied.

Name

Type & Properties

--output

file

output

--obs_name_filter

string

--obs_log1p_column

string

--var_log1p_column

string

--var_name_filter

string

--output_compression

string