Merge pull request #139 from vib-singlecell-nf/develop

Fix bug FDRs missing in final loom Former-commit-id: 8be99f3
vib-singlecell-nf · Feb 25, 2020 · 1b783b1 · 1b783b1
2 parents ba501de + b00865a
commit 1b783b1
Show file tree

Hide file tree

Showing 2 changed files with 7 additions and 13 deletions.
diff --git a/nextflow.config b/nextflow.config
@@ -3,7 +3,7 @@ manifest {
     name = 'vib-singlecell-nf/vsn-pipelines'
     description = 'A repository of pipelines for single-cell data in Nextflow DSL2'
     homePage = 'https://github.com/vib-singlecell-nf/vsn-pipelines'
-    version = '0.13.1'
+    version = '0.13.2'
     mainScript = 'main.nf'
     defaultBranch = 'master'
     nextflowVersion = '!19.12.0-edge' // with ! prefix, stop execution if current version does not match required version.

diff --git a/src/utils/bin/h5ad_to_loom.py b/src/utils/bin/h5ad_to_loom.py
@@ -299,7 +299,7 @@ def read_h5ad(file_path, backed='r'):
         gene_names = adatas[adata_idx].uns['rank_genes_groups']['names'][i]
         pvals_adj = adatas[adata_idx].uns['rank_genes_groups']['pvals_adj'][i]
         logfoldchanges = adatas[adata_idx].uns['rank_genes_groups']['logfoldchanges'][i]
-        num_genes = len(pvals_adj)
+        num_genes = len(gene_names)
         sig_genes_mask = pvals_adj < args.markers_fdr_threshold
         deg_genes_mask = np.logical_and(
             np.logical_or(
@@ -318,7 +318,7 @@ def read_h5ad(file_path, backed='r'):
 
         marker_genes_along_raw_adata_mask = np.in1d(
             raw_filtered_adata.var.index,
-            gene_names[sig_and_deg_genes_mask]
+            marker_names
         )
         marker_genes_along_raw_adata = cluster_markers.index[marker_genes_along_raw_adata_mask]
 
@@ -345,24 +345,18 @@ def read_h5ad(file_path, backed='r'):
         cluster_markers_avg_logfc.loc[
             marker_genes_along_raw_adata_mask,
             i
-        ] = np.around(
-            logfoldchanges_df["logfc"][marker_genes_along_raw_adata],
-            decimals=6
-        )
+        ] = logfoldchanges_df["logfc"][marker_genes_along_raw_adata]
 
         # Populate the marker gene false discovery rates
         pvals_adj_df = pd.DataFrame(
-            logfoldchanges[sig_and_deg_genes_mask],
+            pvals_adj[sig_and_deg_genes_mask],
             index=marker_names,
             columns=["fdr"]
         )
-        cluster_markers_avg_logfc.loc[
+        cluster_markers_pval.loc[
             marker_genes_along_raw_adata_mask,
             i
-        ] = np.around(
-            pvals_adj_df["fdr"][marker_genes_along_raw_adata],
-            decimals=6
-        )
+        ] = pvals_adj_df["fdr"][marker_genes_along_raw_adata]
 
     # Update row attribute Dict
     row_attrs_cluster_markers = {