From 05157c7ef769aac370d4d87965f28441031f29a2 Mon Sep 17 00:00:00 2001
From: Elizabeth Kiernan <55763654+ekiernan@users.noreply.github.com>
Date: Tue, 14 Jan 2025 09:51:35 -0500
Subject: [PATCH] adding GTF to uns field of h5ad for optimus (#154)

updated warp-tools to 2.6.0 for modified h5ad scripts that add gtf file path into h5ad unstructured metadata
---
 tools/scripts/create_h5ad_optimus.py            | 13 +++++++++++++
 .../scripts/create_snrna_optimus_exons_h5ad.py  | 17 +++++++++++++++++
 tools/scripts/create_snrna_optimus_full_h5ad.py | 16 ++++++++++++++++
 3 files changed, 46 insertions(+)

diff --git a/tools/scripts/create_h5ad_optimus.py b/tools/scripts/create_h5ad_optimus.py
index ec8bd12e..210f62a7 100755
--- a/tools/scripts/create_h5ad_optimus.py
+++ b/tools/scripts/create_h5ad_optimus.py
@@ -390,6 +390,11 @@ def create_h5ad_files(args):
     # Set variable names
     new_data.var_names = [x for x in new_data.var["Gene"]]
     
+    # Add GTF to uns field
+
+    gtf_path = args.gtf_path
+    new_data.uns["reference_gtf_file"] = gtf_path
+    
     # Write h5ad file
     new_data.write(args.output_h5ad_path + ".h5ad")
 
@@ -458,6 +463,14 @@ def main():
         help="annotation file in GTF format",
     )
 
+    parser.add_argument(
+        "--gtf_path",
+        dest="gtf_path",
+        default=None,
+        required=False,
+        help="annotation file path",
+    )
+
     parser.add_argument(
         "--output_path_for_h5ad",
         dest="output_h5ad_path",
diff --git a/tools/scripts/create_snrna_optimus_exons_h5ad.py b/tools/scripts/create_snrna_optimus_exons_h5ad.py
index 3250ee3a..d1876e61 100644
--- a/tools/scripts/create_snrna_optimus_exons_h5ad.py
+++ b/tools/scripts/create_snrna_optimus_exons_h5ad.py
@@ -352,6 +352,15 @@ def create_h5ad_files(args):
     # Set the layer = to the exon_counts csr matrix
     new_data.layers["exon_counts"]=exon_counts
     
+    # Original path from args.annotation_file
+    # Add GTF to uns field
+
+    gtf_path = args.gtf_path
+    new_data.uns["reference_gtf_file"] = gtf_path
+    
+    # Write h5ad file
+    new_data.write(args.output_h5ad_path + ".h5ad")
+    
     # Write h5ad file
     new_data.write(args.output_h5ad_path + ".h5ad")
 
@@ -426,6 +435,14 @@ def main():
         help="annotation file in GTF format",
     )
 
+    parser.add_argument(
+        "--gtf_path",
+        dest="gtf_path",
+        default=None,
+        required=False,
+        help="annotation file path",
+    )
+
     parser.add_argument(
         "--output_path_for_h5ad",
         dest="output_h5ad_path",
diff --git a/tools/scripts/create_snrna_optimus_full_h5ad.py b/tools/scripts/create_snrna_optimus_full_h5ad.py
index d587c1f8..e74e513f 100755
--- a/tools/scripts/create_snrna_optimus_full_h5ad.py
+++ b/tools/scripts/create_snrna_optimus_full_h5ad.py
@@ -335,6 +335,14 @@ def create_h5ad_files(args):
     # Set variable names
     new_data.var_names = [x for x in new_data.var["Gene"]]
     
+    # Add GTF to uns field
+
+    gtf_path = args.gtf_path
+    new_data.uns["reference_gtf_file"] = gtf_path
+    
+    # Write h5ad file
+    new_data.write(args.output_h5ad_path + ".h5ad")
+    
     # Write h5ad file
     new_data.write(args.output_h5ad_path + ".h5ad")
     
@@ -390,6 +398,14 @@ def main():
         help="annotation file in GTF format",
     )
 
+    parser.add_argument(
+        "--gtf_path",
+        dest="gtf_path",
+        default=None,
+        required=False,
+        help="annotation file path",
+    )
+
     parser.add_argument(
         "--output_path_for_h5ad",
         dest="output_h5ad_path",