Support BAM/CRAM input without SM tags in DNAscopeLR

Sentieon · Jun 5, 2024 · fc5d181 · fc5d181
1 parent 046f148
commit fc5d181
Show file tree

Hide file tree

Showing 2 changed files with 6 additions and 0 deletions.
diff --git a/sentieon_cli/command_strings.py b/sentieon_cli/command_strings.py
@@ -231,6 +231,7 @@ def cmd_samtools_fastq_minimap2(
     model_bundle: pathlib.Path,
     cores: int,
     rg_lines: List[str],
+    sample_name: str,
     input_ref: Optional[pathlib.Path] = None,
     fastq_taglist: str = "*",
     util_sort_args: str = "--cram_write_options version=3.0,compressor=rans",
@@ -284,6 +285,9 @@ def cmd_samtools_fastq_minimap2(
     # Commands to replace the @RG lines in the header
     rg_cmds: List[List[str]] = []
     for rg_line in rg_lines:
+        # Add an SM value, if missing
+        if "\tSM:" not in rg_line:
+            rg_line += f"\tSM:{sample_name}"
         rg_cmds.append(
             [
                 "samtools",

diff --git a/sentieon_cli/dnascope_longread.py b/sentieon_cli/dnascope_longread.py
@@ -79,6 +79,7 @@ def align_inputs(
 
     res: List[pathlib.Path] = []
     suffix = "bam" if bam_format else "cram"
+    sample_name = output_vcf.name.replace(".vcf.gz", "")
     for i, input_aln in enumerate(sample_input):
         out_aln = pathlib.Path(
             str(output_vcf).replace(".vcf.gz", f"_mm2_sorted_{i}.{suffix}")
@@ -96,6 +97,7 @@ def align_inputs(
                 model_bundle,
                 cores,
                 rg_lines,
+                sample_name,
                 input_ref,
                 fastq_taglist,
                 util_sort_args,