Merge pull request #132 from nf-core/scvi-parameters

Add scVI parameters
nf-core · Dec 9, 2024 · 4045b22 · 4045b22
2 parents 6da26ba + 15a57cf
commit 4045b22
Show file tree

Hide file tree

Showing 7 changed files with 94 additions and 8 deletions.
diff --git a/conf/modules.config b/conf/modules.config
@@ -247,6 +247,13 @@ process {
     // Integration
 
     withName: SCVITOOLS_SCVI {
+        ext.n_hidden = { params.scvi_n_hidden }
+        ext.n_layers = { params.scvi_n_layers }
+        ext.n_latent = { params.scvi_n_latent }
+        ext.dispersion = { params.scvi_dispersion }
+        ext.gene_likelihood = { params.scvi_gene_likelihood }
+        ext.max_epochs = { params.scvi_max_epochs }
+
         publishDir = [
             path: { "${params.outdir}/combine/integrate/scvi" },
             mode: params.publish_dir_mode,
@@ -255,6 +262,13 @@ process {
     }
 
     withName: SCVITOOLS_SCANVI {
+        ext.n_hidden = { params.scvi_n_hidden }
+        ext.n_layers = { params.scvi_n_layers }
+        ext.n_latent = { params.scvi_n_latent }
+        ext.dispersion = { params.scvi_dispersion }
+        ext.gene_likelihood = { params.scvi_gene_likelihood }
+        ext.max_epochs = { params.scvi_max_epochs }
+
         publishDir = [
             path: { "${params.outdir}/combine/integrate/scanvi" },
             mode: params.publish_dir_mode,

diff --git a/modules/local/scvitools/scanvi/main.nf b/modules/local/scvitools/scanvi/main.nf
@@ -25,5 +25,11 @@ process SCVITOOLS_SCANVI {
 
     script:
     prefix = task.ext.prefix ?: "${meta.id}"
+    n_hidden = task.ext.n_hidden ?: 128
+    n_layers = task.ext.n_layers ?: 2
+    n_latent = task.ext.n_latent ?: 30
+    dispersion = task.ext.dispersion ?: 'gene'
+    gene_likelihood = task.ext.gene_likelihood ?: 'zinb'
+    max_epochs = task.ext.max_epochs ?: null
     template 'scanvi.py'
 }
diff --git a/modules/local/scvitools/scanvi/templates/scanvi.py b/modules/local/scvitools/scanvi/templates/scanvi.py
@@ -54,12 +54,18 @@ def format_yaml_like(data: dict, indent: int = 0) -> str:
         )
     else:
         SCANVI.setup_anndata(adata, batch_key="batch", labels_key="label", unlabeled_category="unknown")
-        model = SCANVI(adata)
+        model = SCANVI(adata,
+                        n_hidden=int("${n_hidden}"),
+                        n_layers=int("${n_layers}"),
+                        n_latent=int("${n_latent}"),
+                        dispersion="${dispersion}",
+                        gene_likelihood="${gene_likelihood}")
 
 if "${task.ext.use_gpu}" == "true":
     model.to_device(0)
 
-model.train(early_stopping=True)
+model.train(early_stopping=True,
+            max_epochs=int("${max_epochs}") if "${max_epochs?:''}" else None)
 adata.obsm["X_emb"] = model.get_latent_representation()
 adata.obs["label:scANVI"] = model.predict()
 

diff --git a/modules/local/scvitools/scvi/main.nf b/modules/local/scvitools/scvi/main.nf
@@ -24,5 +24,11 @@ process SCVITOOLS_SCVI {
 
     script:
     prefix = task.ext.prefix ?: "${meta.id}"
+    n_hidden = task.ext.n_hidden ?: 128
+    n_layers = task.ext.n_layers ?: 2
+    n_latent = task.ext.n_latent ?: 30
+    dispersion = task.ext.dispersion ?: 'gene'
+    gene_likelihood = task.ext.gene_likelihood ?: 'zinb'
+    max_epochs = task.ext.max_epochs ?: null
     template 'scvi.py'
 }
diff --git a/modules/local/scvitools/scvi/templates/scvi.py b/modules/local/scvitools/scvi/templates/scvi.py
@@ -43,12 +43,18 @@ def format_yaml_like(data: dict, indent: int = 0) -> str:
     model = SCVI.load_query_data(adata, reference_model_path)
 else:
     SCVI.setup_anndata(adata, batch_key = "batch")
-    model = SCVI(adata)
+    model = SCVI(adata,
+                    n_hidden=int("${n_hidden}"),
+                    n_layers=int("${n_layers}"),
+                    n_latent=int("${n_latent}"),
+                    dispersion="${dispersion}",
+                    gene_likelihood="${gene_likelihood}")
 
 if "${task.ext.use_gpu}" == "true":
     model.to_device(0)
 
-model.train(early_stopping=True)
+model.train(early_stopping=True,
+            max_epochs=int("${max_epochs}") if "${max_epochs?:''}" else None)
 
 adata.obsm["X_emb"] = model.get_latent_representation()
 

diff --git a/nextflow.config b/nextflow.config
@@ -41,6 +41,14 @@ params {
     use_gpu                      = false
     prep_cellxgene               = false
 
+    // scVI options
+    scvi_n_latent               = 30
+    scvi_n_hidden               = 128
+    scvi_n_layers               = 2
+    scvi_dispersion             = 'gene'
+    scvi_gene_likelihood        = 'zinb'
+    scvi_max_epochs             = null
+
     // MultiQC options
     multiqc_config               = null
     multiqc_title                = null

diff --git a/nextflow_schema.json b/nextflow_schema.json
@@ -205,6 +205,47 @@
                 }
             }
         },
+        "scVI_options": {
+            "title": "scVI/scANVI options",
+            "type": "object",
+            "fa_icon": "fas fa-microscope",
+            "description": "Options for the scVI and scANVI integration methods",
+            "properties": {
+                "scvi_n_latent": {
+                    "type": "integer",
+                    "default": 30,
+                    "description": "Number of latent dimensions for scVI/scANVI"
+                },
+                "scvi_n_hidden": {
+                    "type": "integer",
+                    "default": 128,
+                    "description": "Number of hidden units in the neural network for scVI/scANVI"
+                },
+                "scvi_n_layers": {
+                    "type": "integer",
+                    "default": 2,
+                    "description": "Number of layers in the neural network for scVI/scANVI"
+                },
+                "scvi_dispersion": {
+                    "type": "string",
+                    "default": "gene",
+                    "enum": ["gene", "gene-batch", "gene-label", "gene-cell"],
+                    "description": "Dispersion parameter for scVI/scANVI",
+                    "help_text": "Dispersion parameter for scVI/scANVI. Can be 'gene', 'gene-batch', 'gene-label', or 'gene-cell'."
+                },
+                "scvi_gene_likelihood": {
+                    "type": "string",
+                    "default": "zinb",
+                    "enum": ["zinb", "nb", "poisson", "normal"],
+                    "description": "Gene likelihood for scVI/scANVI",
+                    "help_text": "Gene likelihood for scVI/scANVI. Can be 'zinb', 'nb', 'poisson', or 'normal'."
+                },
+                "scvi_max_epochs": {
+                    "type": "integer",
+                    "description": "Maximum number of epochs for training scVI/scANVI. If not set, a heuristic provided by scVI/scANVI will be used."
+                }
+            }
+        },
         "institutional_config_options": {
             "title": "Institutional config options",
             "type": "object",
@@ -349,16 +390,15 @@
         {
             "$ref": "#/$defs/input_output_options"
         },
-
-        {
-            "$ref": "#/$defs/institutional_config_options"
-        },
         {
             "$ref": "#/$defs/reference_mapping_options"
         },
         {
             "$ref": "#/$defs/pipeline_options"
         },
+        {
+            "$ref": "#/$defs/scVI_options"
+        },
         {
             "$ref": "#/$defs/institutional_config_options"
         },