Update pipelines for Cellranger 9.0.0

auto_process_ngs/tenx/__init__.py

@@ -83,9 +83,14 @@
     'SC3Pv1': 'Single Cell 3\' v1',
     'SC3Pv2': 'Single Cell 3\' v2',
     'SC3Pv3': 'Single Cell 3\' v3',
+    'SC3Pv4': 'Single Cell 3\' v4',
+    'SC3Pv3LT': 'Single Cell 3\' v3 LT',
+    'SC3Pv3HT': 'Single Cell 3\' v3 HT',
     'SC5P-PE': 'Single Cell 5\' paired-end (both R1 and R2 are used for alignment)',
+    'SC5P-PE-v3': 'Single Cell 5\' paired-end (both R1 and R2 are used for alignment) v3',
     'SC5P-R2': 'Single Cell 5\' R2-only (where only R2 is used for alignment)',
-    'ARC-v1': 'Single Cell Multiome (ATAC+GEX) v1', # Not documented?
+    'SC5P-R2-v3': 'Single Cell 5\' R2-only (where only R2 is used for alignment)',
+    'ARC-v1': 'Single Cell Multiome (ATAC+GEX) v1 (cannot be autodetected)'
 }
 
 # Default Cellranger version

auto_process_ngs/tenx/utils.py

@@ -568,6 +568,13 @@ def make_multi_config_template(f,reference=None,probe_set=None,
         cellranger_version = DEFAULT_CELLRANGER_VERSION
     else:
         cellranger_version = str(cellranger_version)
+    # Major version
+    try:
+        cellranger_major_version = int(cellranger_version.split(".")[0])
+    except Exception as ex:
+        raise Exception(f"'{cellranger_version}': unable to extract "
+                        f"Cellranger major  version from version string: "
+                        f"{ex}")
     # Normalise and check supplied library type
     if library_type.startswith("CellPlex"):
         library_type = "cellplex"
@@ -605,14 +612,14 @@ def make_multi_config_template(f,reference=None,probe_set=None,
                      (probe_set if probe_set
                       else "/path/to/probe/set"))
         fp.write("#force-cells,n\n")
-        if cellranger_version.startswith("7."):
+        if cellranger_major_version == 7:
             # Cellranger 7.* targetted
             if no_bam is not None:
                 fp.write("no-bam,%s\n" % str(no_bam).lower())
             else:
                 fp.write("#no-bam,true|false\n")
-        elif cellranger_version.startswith("8."):
-            # Cellranger 8.* targetted
+        elif cellranger_major_version in (8, 9):
+            # Cellranger 8.* or 9.* targetted
             if no_bam is not None:
                 fp.write("create-bam,%s\n" % str(not no_bam).lower())
             else:

auto_process_ngs/test/tenx/test_utils.py

@@ -374,6 +374,13 @@ def _make_mock_cellranger_800(self):
             fp.write("#!/bin/bash\necho cellranger cellranger-8.0.0")
         os.chmod(cellranger_800,0o775)
         return cellranger_800
+    def _make_mock_cellranger_900(self):
+        # Make a fake cellranger 9.0.0 executable
+        cellranger_900 = os.path.join(self.wd,"cellranger")
+        with open(cellranger_900,'w') as fp:
+            fp.write("#!/bin/bash\necho cellranger cellranger-9.0.0")
+        os.chmod(cellranger_900,0o775)
+        return cellranger_900
     def _make_mock_cellranger_atac_101(self):
         # Make a fake cellranger-atac 1.0.1 executable
         cellranger_atac_101 = os.path.join(self.wd,"cellranger-atac")
@@ -449,6 +456,13 @@ def test_cellranger_800(self):
         self.assertEqual(cellranger_info(path=cellranger),
                          (cellranger,'cellranger','8.0.0'))
 
+    def test_cellranger_900(self):
+        """cellranger_info: collect info for cellranger 8.0.0
+        """
+        cellranger = self._make_mock_cellranger_900()
+        self.assertEqual(cellranger_info(path=cellranger),
+                         (cellranger,'cellranger','9.0.0'))
+
     def test_cellranger_atac_101(self):
         """cellranger_info: collect info for cellranger-atac 1.0.1
         """
@@ -775,6 +789,41 @@ def test_make_multi_config_template_cellplex_scrnaseq_800(self):
                                         if not line.startswith('##')])
         self.assertEqual(expected_content,actual_content)
 
+    def test_make_multi_config_template_cellplex_900(self):
+        """
+        make_multi_config_template: check CellPlex template (9.0.0)
+        """
+        expected_content = """[gene-expression]
+reference,/data/mm10_transcriptome
+#force-cells,n
+create-bam,true
+#cmo-set,/path/to/custom/cmo/reference
+
+#[feature]
+#reference,/path/to/feature/reference
+
+[libraries]
+fastq_id,fastqs,lanes,physical_library_id,feature_types,subsample_rate
+PJB_CML,/runs/novaseq_50/fastqs,any,PJB_CML,[Gene Expression|Multiplexing Capture],
+PJB_GEX,/runs/novaseq_50/fastqs,any,PJB_GEX,[Gene Expression|Multiplexing Capture],
+
+[samples]
+sample_id,cmo_ids,description
+MULTIPLEXED_SAMPLE,CMO1|CMO2|...,DESCRIPTION
+"""
+        out_file = os.path.join(self.wd,"10x_multi_config.csv")
+        make_multi_config_template(out_file,
+                                   reference="/data/mm10_transcriptome",
+                                   fastq_dir="/runs/novaseq_50/fastqs",
+                                   samples=("PJB_CML","PJB_GEX"),
+                                   library_type="CellPlex",
+                                   cellranger_version="9.0.0")
+        self.assertTrue(os.path.exists(out_file))
+        with open(out_file,'rt') as fp:
+            actual_content = '\n'.join([line for line in fp.read().split('\n')
+                                        if not line.startswith('##')])
+        self.assertEqual(expected_content,actual_content)
+
     def test_make_multi_config_template_flex_700(self):
         """
         make_multi_config_template: check Flex template (7.1.0)
@@ -849,6 +898,43 @@ def test_make_multi_config_template_flex_800(self):
                                         if not line.startswith('##')])
         self.assertEqual(expected_content,actual_content)
 
+    def test_make_multi_config_template_flex_900(self):
+        """
+        make_multi_config_template: check Flex template (9.0.0)
+        """
+        expected_content = """[gene-expression]
+reference,/data/mm10_transcriptome
+probe-set,/data/mm10_probe_set.csv
+#force-cells,n
+create-bam,false
+#cmo-set,/path/to/custom/cmo/reference
+
+#[feature]
+#reference,/path/to/feature/reference
+
+[libraries]
+fastq_id,fastqs,lanes,physical_library_id,feature_types,subsample_rate
+PJB_Flex,/runs/novaseq_50/fastqs,any,PJB_Flex,[Gene Expression|Antibody Capture],
+
+[samples]
+sample_id,probe_barcode_ids,description
+MULTIPLEXED_SAMPLE,BC001|BC002|...,DESCRIPTION
+"""
+        out_file = os.path.join(self.wd,"10x_multi_config.csv")
+        make_multi_config_template(out_file,
+                                   reference="/data/mm10_transcriptome",
+                                   probe_set="/data/mm10_probe_set.csv",
+                                   fastq_dir="/runs/novaseq_50/fastqs",
+                                   samples=("PJB_Flex",),
+                                   no_bam=True,
+                                   library_type="Flex",
+                                   cellranger_version="9.0.0")
+        self.assertTrue(os.path.exists(out_file))
+        with open(out_file,'rt') as fp:
+            actual_content = '\n'.join([line for line in fp.read().split('\n')
+                                        if not line.startswith('##')])
+        self.assertEqual(expected_content,actual_content)
+
     def test_make_multi_config_template_immune_profiling_710(self):
         """
         make_multi_config_template: check Single Cell Immune Profiling template (7.0.0)
@@ -917,6 +1003,40 @@ def test_make_multi_config_template_immune_profiling_800(self):
                                         if not line.startswith('##')])
         self.assertEqual(expected_content,actual_content)
 
+    def test_make_multi_config_template_immune_profiling_900(self):
+        """
+        make_multi_config_template: check Single Cell Immune Profiling template (9.0.0)
+        """
+        expected_content = """[gene-expression]
+reference,/data/mm10_transcriptome
+#force-cells,n
+create-bam,true
+#cmo-set,/path/to/custom/cmo/reference
+
+#[feature]
+#reference,/path/to/feature/reference
+
+#[vdj]
+#reference,/path/to/vdj/reference
+
+[libraries]
+fastq_id,fastqs,lanes,physical_library_id,feature_types,subsample_rate
+PJB_CML,/runs/novaseq_50/fastqs,any,PJB_CML,[Gene Expression|Antibody Capture|VDJ-B|VDJ-T],
+PJB_GEX,/runs/novaseq_50/fastqs,any,PJB_GEX,[Gene Expression|Antibody Capture|VDJ-B|VDJ-T],
+"""
+        out_file = os.path.join(self.wd,"10x_multi_config.csv")
+        make_multi_config_template(out_file,
+                                   reference="/data/mm10_transcriptome",
+                                   fastq_dir="/runs/novaseq_50/fastqs",
+                                   samples=("PJB_CML","PJB_GEX"),
+                                   library_type="Single Cell Immune Profiling",
+                                   cellranger_version="9.0.0")
+        self.assertTrue(os.path.exists(out_file))
+        with open(out_file,'rt') as fp:
+            actual_content = '\n'.join([line for line in fp.read().split('\n')
+                                        if not line.startswith('##')])
+        self.assertEqual(expected_content,actual_content)
+
     def test_make_multi_config_template_unsupported_library(self):
         """
         make_multi_config_template: exception for unsupported library type