diff deconvolve.xml @ 0:4f6d716e5da2 draft

planemo upload for repository https://github.com/goeckslab/tools-st/tree/main/tools/bayesTME commit 26edd05a863adac97cb54a9fb2ae5613ce95df50

--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/deconvolve.xml	Wed Dec 03 19:49:22 2025 +0000
@@ -0,0 +1,242 @@
+<tool id="bayestme_deconvolve" name="BayesTME: Deconvolve" version="@TOOL_VERSION@+galaxy@VERSION_SUFFIX@" profile="@PROFILE@">
+    <description>Reference-free and Reference-based Spot Deconvolution for Spatial Transcriptomics with BayesTME</description>
+    <macros>
+        <import>macros.xml</import>
+    </macros>
+	<expand macro="bayestme_requirements"/>
+	<command detect_errors="aggressive"><![CDATA[
+		
+		## Filtering Step (optional)
+		#if str($filter_genes.pre_filter) == 'filter_genes':
+		
+	        filter_genes
+	            --adata '${adata}'
+	            --output 'dataset_filtered.h5ad'
+	            --n-top-by-standard-deviation '${n_top_by_standard_deviation}'
+	            #if str($spot_threshold)
+	                --spot-threshold '${spot_threshold}'
+	            #end if
+			    #if $filter_ribosomal_genes
+	                --filter-ribosomal-genes
+	            #end if
+				#if $mode.deconv_mode == 'reference_based'
+				    --expression-truth '${mode.expression_truth}'
+			    #end if
+			    #if $verbose
+	                -v
+	            #end if
+				&&
+        #else
+		
+		    ## Just use unfiltered
+	        ln -s '${adata}' 'dataset_filtered.h5ad' &&
+			
+		#end if
+		
+		## Deconvolve Step
+		deconvolve
+	        --adata 'dataset_filtered.h5ad'
+			--adata-output 'dataset_deconvolved.h5ad'
+			--output 'deconvolution_samples.h5'
+			#if $mode.deconv_mode == 'reference_free'
+			    --n-components '${mode.n_components}'
+		    #end if
+			#if $mode.deconv_mode == 'reference_based'
+			    --expression-truth '${mode.expression_truth}'
+				--reference-scrna-celltype-column '${mode.reference_scrna_celltype_column}'
+				--reference-scrna-sample-column '${mode.reference_scrna_sample_column}'
+			#end if
+			--spatial-smoothing-parameter '${deconv_params.spatial_smoothing_parameter}'
+			--n-samples '${deconv_params.n_samples}'
+			--n-svi-steps '${deconv_params.n_svi_steps}'
+	        #if $deconv_params.use_spatial_guide
+	            --use-spatial-guide
+	        #else
+	            --no-use-spatial-guide
+	        #end if
+			#if $additional_cli.verbose
+	            -v
+	        #end if
+	        #if str($additional_cli.seed)
+	            --seed '${additional_cli.seed}'
+	        #end if
+			
+	]]></command>
+    
+	<inputs>
+		
+	    <!-- Required Input ST Data -->
+	    <param argument="--adata" type="data" format="h5ad" label="Input AnnData in h5 format, expected to be already bleed corrected"/>
+	    
+		<!-- Filter Genes Options (from macros) -->
+		<conditional name="filter_genes">
+	        <param name="pre_filter" type="select" label="Optional: Filter Genes By Spatial Variance Prior To Deconvolution" >
+	            <option value="no_filtering" selected="true">No Filtering</option>
+				<option value="filter_genes">Filter Genes Prior To Deconvolution</option>
+	        </param>
+			<when value="no_filtering">
+	        </when>
+	        <when value="filter_genes">				
+				<expand macro="filter_genes_options" />			
+	        </when>
+	    </conditional>
+		
+		<!-- Deconvolution Strategy -->
+		<conditional name="mode">
+			
+		    <param name="deconv_mode" type="select" label="Deconvolution Strategy">
+		        <option value="reference_free" selected="true">Reference-free (unsupervised)</option>
+		        <option value="reference_based">Reference-based (with scRNA reference)</option>
+		    </param>
+
+		    <!-- Reference-free: Specify Number of Cell Types -->
+		    <when value="reference_free">
+		        <param argument="--n-components" type="integer" optional="false" label="Number of cell types to deconvolve into"/>
+		    </when>
+
+		    <!-- Reference-based: Provide scRNA Ref -->
+		    <when value="reference_based">
+				<param argument="--expression-truth" type="data" format="h5ad" optional="true" label="Matched scRNA data in h5ad format, will be used to enforce a prior on celltypes and expression"/>
+		        <param argument="--reference-scrna-celltype-column" type="text" label="The name of the column with celltype id in the matched scRNA anndata"/>
+		        <param argument="--reference-scrna-sample-column" type="text" label="The name of the column with sample id in the matched scRNA anndata"/>
+		    </when>
+			
+		</conditional>
+		
+		<!-- Deconvolution Parameters -->
+		<section name="deconv_params" title="Deconvolution Parameters" expanded="false"> 
+			<param argument="--spatial-smoothing-parameter" type="float" value="0.5" label="Spatial smoothing parameter (referred to as lambda in paper)"/>
+		   	<param argument="--n-samples" type="integer" value="100" label="Number of samples from the posterior distribution or variational family"/>
+		   	<param argument="--n-svi-steps" type="integer" value="20000" label="Number of steps for fitting variational family"/>
+			<param name="use_spatial_guide" argument="--use-spatial-guide, --no-use-spatial-guide" type="boolean" checked="true" label="Use spatial guide (variational family with spatial priors) for SVI"/>
+		</section>
+		
+	    <!-- Additional CLI Options -->
+		<section name="additional_cli" title="Additional Arguments" expanded="false"> 
+	        <param argument="--seed" type="integer" optional="true" label="Seed value for random number generator"/>
+	   	    <param argument="--verbose" type="boolean" checked="false" label="Enable verbose logging"/>
+		</section>	   		
+		
+	</inputs>
+	
+    <outputs>
+        <data name="adata_output" format="h5ad" label="BayesTME Deconvolve on ${on_string}: dataset_deconvolved.h5ad" from_work_dir="dataset_deconvolved.h5ad"/>
+		<data name="output" format="h5" label="BayesTME Deconvolve on ${on_string}: deconvolution_samples.h5" from_work_dir="deconvolution_samples.h5"/>
+		<data name="deconvolution_loss" format="pdf" label="BayesTME Deconvolve on ${on_string}: deconvolution_loss.pdf" from_work_dir="deconvolution_loss.pdf"/>
+    </outputs>
+	
+    <tests>
+		
+		<!-- Reference-free Deconvolution Test -->
+		<test>
+			<param name="adata" value="visium_test_dataset.h5ad"/>
+			<conditional name="mode" >
+				<param name="deconv_mode" value="reference_free"/>
+				<param name="n_components" value="3"/>
+			</conditional>
+		    <section name="deconv_params" >
+			    <param name="spatial_smoothing_parameter" value="0.5"/>
+			   	<param name="n_samples" value="10"/>
+			   	<param name="n_svi_steps" value="10"/>
+			</section>
+			<section name="additional_cli" >
+				<param name="seed" value="93"/>
+			</section>
+			<conditional name="filter_genes" >
+				<param name="pre_filter" value="filter_genes"/>
+			    <param name="n_top_by_standard_deviation" value="40"/>
+				<section name="additional_filtering_options" >
+				    <param name="filter_ribosomal_genes" value="true"/>
+				</section>
+			</conditional>
+			<output name="adata_output" file="refFree_test_data/dataset_deconvolved.h5ad" compare="sim_size" delta="20000">
+				<assert_contents>
+					<has_h5_keys keys="obsm/bayestme_cell_type_counts,obsm/bayestme_cell_type_probabilities,varm/bayestme_omega,varm/bayestme_omega_difference,varm/bayestme_relative_expression,varm/bayestme_relative_mean_expression,uns/bayestme_n_cell_types"/>
+				</assert_contents>
+			</output>
+			<output name="output" file="refFree_test_data/deconvolution_samples.h5" compare="sim_size" delta="500000">
+				<assert_contents>
+					<has_h5_keys keys="beta_trace,cell_num_total_trace,cell_prob_trace,expression_trace,losses,reads_trace"/>
+				</assert_contents>
+			</output>	
+			<output name="deconvolution_loss" file="refFree_test_data/deconvolution_loss.pdf" compare="sim_size" delta="20000" ftype="pdf"/>			
+		</test>
+		
+        <!-- Reference-based Deconvolution Test -->
+        <test>
+            <param name="adata" value="visium_test_dataset.h5ad"/>
+			<conditional name="mode" >
+				<param name="deconv_mode" value="reference_based"/>
+	            <param name="expression_truth" value="refBased_test_data/scRef_test_dataset.h5ad"/>
+	            <param name="reference_scrna_celltype_column" value="cell_type"/>
+	            <param name="reference_scrna_sample_column" value="donor_id"/>
+			</conditional>
+            <section name="deconv_params" >
+	            <param name="spatial_smoothing_parameter" value="0.5"/>
+	            <param name="n_samples" value="10"/>
+	            <param name="n_svi_steps" value="10"/>
+			</section>
+			<section name="additional_cli" >
+				<param name="seed" value="93"/>
+			</section>
+			<conditional name="filter_genes" >
+				<param name="pre_filter" value="filter_genes"/>
+			    <param name="n_top_by_standard_deviation" value="40"/>
+				<section name="additional_filtering_options" >
+				    <param name="filter_ribosomal_genes" value="true"/>
+				</section>
+			</conditional>
+			<output name="adata_output" file="refBased_test_data/dataset_deconvolved.h5ad" compare="sim_size" delta="20000">
+                <assert_contents>
+                    <has_h5_keys keys="obsm/bayestme_cell_type_counts,obsm/bayestme_cell_type_probabilities,varm/bayestme_omega,varm/bayestme_omega_difference,varm/bayestme_relative_expression,varm/bayestme_relative_mean_expression,uns/bayestme_n_cell_types"/>
+                </assert_contents>
+            </output>
+            <output name="output" file="refBased_test_data/deconvolution_samples.h5" compare="sim_size" delta="500000">
+                <assert_contents>
+                    <has_h5_keys keys="beta_trace,cell_num_total_trace,cell_prob_trace,expression_trace,losses,reads_trace"/>
+                </assert_contents>
+            </output>
+			<output name="deconvolution_loss" file="refBased_test_data/deconvolution_loss.pdf" compare="sim_size" delta="20000" ftype="pdf"/>
+        </test>
+    
+	</tests>
+    <help>
+<![CDATA[
+
+usage: deconvolve [-h] [--adata ADATA] [--adata-output ADATA_OUTPUT] [-i] [--output OUTPUT] [--seed SEED] [--n-components N_COMPONENTS]
+					[--spatial-smoothing-parameter SPATIAL_SMOOTHING_PARAMETER] [--n-samples N_SAMPLES] [--expression-truth EXPRESSION_TRUTH]
+					[--reference-scrna-celltype-column REFERENCE_SCRNA_CELLTYPE_COLUMN] [--reference-scrna-sample-column REFERENCE_SCRNA_SAMPLE_COLUMN]
+					[--n-svi-steps N_SVI_STEPS] [--use-spatial-guide | --no-use-spatial-guide] [-v]
+
+Deconvolve data
+
+options:
+	-h, --help            show this help message and exit
+	--adata ADATA         Input AnnData in h5 format, expected to be already bleed corrected
+	--adata-output ADATA_OUTPUT
+						A new AnnData in h5 format created with the deconvolution summary results appended.
+	-i, --inplace         If provided, append deconvolution summary results to the --adata archive in place
+	--output OUTPUT       Path where DeconvolutionResult will be written h5 format
+	--seed SEED           Seed value for random number generator.
+	--n-components N_COMPONENTS
+						Number of cell types to deconvolve into.
+	--spatial-smoothing-parameter SPATIAL_SMOOTHING_PARAMETER
+						Spatial smoothing parameter (referred to as lambda in paper)
+	--n-samples N_SAMPLES
+						Number of samples from the posterior distribution or variational family.
+	--expression-truth EXPRESSION_TRUTH
+						Matched scRNA data in h5ad format, will be used to enforce a prior on celltypes and expression.
+	--reference-scrna-celltype-column REFERENCE_SCRNA_CELLTYPE_COLUMN
+						The name of the column with celltype id in the matched scRNA anndata.
+	--reference-scrna-sample-column REFERENCE_SCRNA_SAMPLE_COLUMN
+						The name of the column with sample id in the matched scRNA anndata.
+	--n-svi-steps N_SVI_STEPS
+						Number of steps for fitting variational family
+	--use-spatial-guide, --no-use-spatial-guide
+						Use spatial guide (variational family with spatial priors) for SVI
+	-v, --verbose         Enable verbose logging
+
+]]>
+    </help>
+    <expand macro="citations"/>
+</tool>
\ No newline at end of file