Bump version to 1.0.1: restore stranded hg19/hg38 coords in miRNA/RBP

docs/conf.py

@@ -10,8 +10,8 @@
 copyright = '2026, Barry Digby'
 author = 'Barry Digby'
 
-release = '0.1'
-version = '0.1.5'
+release = '1.0'
+version = '1.0.1'
 
 # -- General configuration
 

docs/outputs.rst

@@ -527,19 +527,21 @@ miRNA output are a combination of CircNET and CSCD - for this reason, not all Ta
 
 .. code-block:: text
 
-    circRNA                     miRNA                  MSA_start  MSA_end  Site_type  Score   Energy  Algorithm
-    chr15:100048856-100054054|+   miR-1208               30.0       37.0     8mer-1a    144.0   -13.62  miRanda,TargetScan
-    chr15:100048856-100054054|+   miR-1237-5p/4488-5p                                   140.0   -20.01  miRanda
-    chr15:100048856-100054054|+   miR-1289                                              141.0   -22.95  miRanda
-    chr15:100048856-100054054|+   miR-141-5p             39.0       45.0     7mer-m8    178.0   -26.76  miRanda,TargetScan
+    hg19                          hg38                          miRNA                  MSA_start  MSA_end  Site_type  Score   Energy  Algorithm
+    chr15:100048856-100054054|+   chr15:100048856-100054054|+   miR-1208               30.0       37.0     8mer-1a    144.0   -13.62  miRanda,TargetScan
+    chr15:100048856-100054054|+   chr15:100048856-100054054|+   miR-1237-5p/4488-5p                                   140.0   -20.01  miRanda
+    chr15:100048856-100054054|+   chr15:100048856-100054054|+   miR-1289                                              141.0   -22.95  miRanda
+    chr15:100048856-100054054|+   chr15:100048856-100054054|+   miR-141-5p             39.0       45.0     7mer-m8    178.0   -26.76  miRanda,TargetScan
 
 .. list-table::
    :header-rows: 1
    :widths: 50 50
 
    * - Column Name
      - Description
-   * - circRNA
+   * - hg19
+     - The circRNA coordinates in the hg19 reference genome.
+   * - hg38
      - The circRNA coordinates in the hg38 reference genome.
    * - miRNA
      - The miRNA ID.
@@ -561,18 +563,20 @@ RBP
 
 .. code-block:: text
 
-    circRNA                 Chr     Start       End         RBP_symbol  Information        RBP_numbers  source
-    chr18:12493071-12546904|+ chr18   12503834    12546883    AGO2        SBDH105-59581      24           common
-    chr18:12493071-12546904|+ chr18   12498779    12546872    FBL         SBDH120-145469     10           common
-    chr18:12493071-12546904|+ chr18   12493071    12546757    U2AF2       SBDH19-189369      100          common
+    hg19                       hg38                      Chr     Start       End         RBP_symbol  Information        RBP_numbers  source
+    chr18:12493071-12546904|+  chr18:12493071-12546904|+ chr18   12503834    12546883    AGO2        SBDH105-59581      24           common
+    chr18:12493071-12546904|+  chr18:12493071-12546904|+ chr18   12498779    12546872    FBL         SBDH120-145469     10           common
+    chr18:12493071-12546904|+  chr18:12493071-12546904|+ chr18   12493071    12546757    U2AF2       SBDH19-189369      100          common
 
 .. list-table::
    :header-rows: 1
    :widths: 50 50
 
    * - Column Name
      - Description
-   * - circRNA
+   * - hg19
+     - The circRNA coordinates in the hg19 reference genome.
+   * - hg38
      - The circRNA coordinates in the hg38 reference genome.
    * - Chr
      - Chromosome of the RBP binding site.

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "pycircdb"
-version = "1.0.0"
+version = "1.0.1"
 description = "pycircdb: integrated circRNA database annotation for computational workflows."
 readme = "README.md"
 authors = [{ name = "Barry Digby", email = "b.digby237@gmail.com" }]

utils/mirna/mirna_driver.py

@@ -30,16 +30,18 @@ def broadcast_mirna(
     for sample_name, lookup_hits in lookup_dict.items():
         # Combine all hg38 IDs for this sample across all databases
         hg38_series = []
-        # Map strand-stripped hg38 key -> full (stranded) hg38 coordinate,
-        # so strand can be restored on the output after the strand-less match.
+        # Map strand-stripped hg38 key -> stranded hg38/hg19 coordinates,
+        # so both can be restored on the output after the strand-less match.
         strand_frames = []
         for db_name, pl_hits in lookup_hits.items():
             if pl_hits.is_empty() or "hg38" not in pl_hits.columns:
                 continue
             hg38_series.append(pl_hits["hg38"].str.split("|").list.first())
+            hg19_col = pl.col("hg19") if "hg19" in pl_hits.columns else pl.lit(None, dtype=pl.Utf8).alias("hg19")
             strand_frames.append(pl_hits.select(
                 pl.col("hg38").str.split("|").list.first().alias("_hg38_key"),
-                pl.col("hg38").alias("circRNA"),
+                pl.col("hg38"),
+                hg19_col,
             ))
 
         if not hg38_series:
@@ -50,7 +52,7 @@ def broadcast_mirna(
         # Prefer stranded coordinates when de-duplicating the strand-less key.
         strand_map = (
             pl.concat(strand_frames)
-            .sort("circRNA")
+            .sort("hg38")
             .unique(subset="_hg38_key", keep="last")
         )
 
@@ -96,16 +98,20 @@ def mirna_hits(broadcast_mirna: miRNABroadcast) -> None:
 
     if not df.is_empty():
         # The miRNA table's circRNA column is the strand-stripped hg38 key.
-        # Join back to restore the strand onto the circRNA coordinate.
+        # Join back to restore stranded hg38/hg19 coordinates onto the output.
         if strand_map is not None and strand_map.height > 0:
             df = (
                 df.rename({"circRNA": "_hg38_key"})
                 .join(strand_map, on="_hg38_key", how="left")
-                # Fall back to the strand-less key if a coordinate is unmapped.
-                .with_columns(pl.col("circRNA").fill_null(pl.col("_hg38_key")))
+                # Fall back to the strand-less key if hg38 is unmapped.
+                .with_columns(pl.col("hg38").fill_null(pl.col("_hg38_key")))
             )
-            ordered = ["circRNA"] + [c for c in df.columns if c not in ("circRNA", "_hg38_key")]
-            df = df.select(ordered)
+        else:
+            # No strand info available: fall back to the strand-less key for hg38.
+            df = df.rename({"circRNA": "hg38"}).with_columns(pl.lit(None, dtype=pl.Utf8).alias("hg19"))
+
+        ordered = ["hg19", "hg38"] + [c for c in df.columns if c not in ("hg19", "hg38", "_hg38_key")]
+        df = df.select(ordered)
 
         p = Path(output_dir)
         if p.is_absolute():

utils/rbp/rbp_driver.py

@@ -29,15 +29,18 @@ def broadcast_rbp(
     for sample_name, lookup_hits in lookup_dict.items():
         # Combine all hg38 IDs for this sample across all databases
         hg38_series = []
-        # Same code as mirna_driver below
+        # Map strand-stripped hg38 key -> stranded hg38/hg19 coordinates,
+        # so both can be restored on the output after the strand-less match.
         strand_frames = []
         for db_name, pl_hits in lookup_hits.items():
             if pl_hits.is_empty() or "hg38" not in pl_hits.columns:
                 continue
             hg38_series.append(pl_hits["hg38"].str.split("|").list.first())
+            hg19_col = pl.col("hg19") if "hg19" in pl_hits.columns else pl.lit(None, dtype=pl.Utf8).alias("hg19")
             strand_frames.append(pl_hits.select(
                 pl.col("hg38").str.split("|").list.first().alias("_hg38_key"),
-                pl.col("hg38").alias("circRNA"),
+                pl.col("hg38"),
+                hg19_col,
             ))
 
         if not hg38_series:
@@ -48,7 +51,7 @@ def broadcast_rbp(
         # Prefer stranded coordinates when de-duplicating the strand-less key.
         strand_map = (
             pl.concat(strand_frames)
-            .sort("circRNA")
+            .sort("hg38")
             .unique(subset="_hg38_key", keep="last")
         )
 
@@ -88,16 +91,20 @@ def rbp_hits(broadcast_rbp: RBPBroadcast) -> None:
 
     if not df.is_empty():
         # The RBP table's circRNA column is the strand-stripped hg38 key.
-        # Join back to restore the strand onto the circRNA coordinate.
+        # Join back to restore stranded hg38/hg19 coordinates onto the output.
         if strand_map is not None and strand_map.height > 0:
             df = (
                 df.rename({"circRNA": "_hg38_key"})
                 .join(strand_map, on="_hg38_key", how="left")
-                # Fall back to the strand-less key if a coordinate is unmapped.
-                .with_columns(pl.col("circRNA").fill_null(pl.col("_hg38_key")))
+                # Fall back to the strand-less key if hg38 is unmapped.
+                .with_columns(pl.col("hg38").fill_null(pl.col("_hg38_key")))
             )
-            ordered = ["circRNA"] + [c for c in df.columns if c not in ("circRNA", "_hg38_key")]
-            df = df.select(ordered)
+        else:
+            # No strand info available: fall back to the strand-less key for hg38.
+            df = df.rename({"circRNA": "hg38"}).with_columns(pl.lit(None, dtype=pl.Utf8).alias("hg19"))
+
+        ordered = ["hg19", "hg38"] + [c for c in df.columns if c not in ("hg19", "hg38", "_hg38_key")]
+        df = df.select(ordered)
 
         p = Path(output_dir)
         if p.is_absolute():