Merge pull request #326 from bcgsc/feature/add-arriba-support

#325 Adding arriba support

.github/workflows/build.yml

@@ -15,7 +15,9 @@ jobs:
     steps:
     - uses: actions/checkout@v2
     - name: install machine dependencies
-      run: sudo apt-get install -y libcurl4-openssl-dev
+      run: |
+        sudo apt-get update
+        sudo apt-get install -y libcurl4-openssl-dev
     - name: Set up Python ${{ matrix.python-version }}
       uses: actions/setup-python@v2
       with:

docs/background/citations.md

@@ -62,6 +62,11 @@ Saunders,C.T. et al. (2012) Strelka: accurate somatic small-variant
     calling from sequenced tumor--normal sample pairs. Bioinformatics,
     28, 1811--1817.
 
+## Uhrig-2021
+
+Uhrig,S. et al. (2021) Accurate and efficient detection of gene
+    fusions from RNA sequencing data. Genome Res., 31, 448--460.
+
 ## Yates-2016
 
 Yates,A. et al. (2016) Ensembl 2016. Nucleic Acids Res., 44,

docs/glossary.md

@@ -127,6 +127,11 @@ install instructions.
 Community based standard of reccommendations for variant notation.
 See [http://varnomen.hgvs.org/](http://varnomen.hgvs.org/)
 
+## Arriba
+
+Arriba is an SV caller. Source for Arriba can be found
+[here](https://github.com/suhrig/arriba) [Uhrig-2021](../background/citations#uhrig-2021)
+
 ## BreakDancer
 
 BreakDancer is an SV caller. Source for BreakDancer can be found

docs/inputs/support.md

@@ -62,6 +62,7 @@ It should be noted however that the tool tracked will only be listed as
 
 | Name                                       | Version(s)       | MAVIS input                                   | Publication                                                 |
 | ------------------------------------------ | ---------------- | --------------------------------------------- | ----------------------------------------------------------- |
+| [Arriba](../../glossary/#arriba)           | `2.2.1`          | `fusions.tsv`                                 | [Uhrig-2021](../../background/citations#uhrig-2021)          |
 | [BreakDancer](../../glossary/#breakdancer) | `1.4.5`          | `Tools main output file(s)`                   | [Chen-2009](../../background/citations#chen-2009)           |
 | [BreakSeq](../../glossary/#breakseq)       | `2.2`            | `work/breakseq.vcf.gz`                        | [Abyzov-2015](../../background/citations#abyzov-2015)       |
 | [Chimerascan](../../glossary/#chimerascan) | `0.4.5`          | `*.bedpe`                                     | [Iyer-2011](../../background/citations#Iyer-2011)           |

mavis/tools/__init__.py

@@ -13,6 +13,7 @@
 from .cnvnator import convert_row as _parse_cnvnator
 from .vcf import convert_file as read_vcf
 from .breakdancer import convert_file as _convert_breakdancer_file
+from .arriba import convert_row as _parse_arriba
 from .starfusion import convert_row as _parse_starfusion
 from .chimerascan import convert_row as _parse_chimerascan
 
@@ -142,6 +143,10 @@ def _convert_tool_row(row, file_type, stranded, assume_no_untemplated=True):
             {k: v for k, v in row.items() if k not in {'Type', 'Chr1', 'Chr2', 'Pos1', 'Pos2'}}
         )
 
+    elif file_type == SUPPORTED_TOOL.ARRIBA:
+
+        std_row.update(_parse_arriba(row))
+
     else:
         raise NotImplementedError('unsupported file type', file_type)
 

mavis/tools/arriba.py

@@ -0,0 +1,39 @@
+from ..constants import COLUMNS, ORIENT, STRAND
+
+from .constants import TRACKING_COLUMN, SUPPORTED_TOOL
+
+
+def get_orient(string):
+    if string == "downstream":
+        return ORIENT.LEFT
+    elif string == "upstream":
+        return ORIENT.RIGHT
+    return ORIENT.NS
+
+
+def convert_row(row):
+    """
+    transforms the aribba output into the common format for expansion. Maps the input column
+    names to column names which MAVIS can read
+    """
+    std_row = {}
+
+    try:
+        std_row[COLUMNS.break1_chromosome], b1_start = row["breakpoint1"].split(":")
+        std_row[COLUMNS.break2_chromosome], b2_start = row["breakpoint2"].split(":")
+
+        std_row[COLUMNS.break1_strand] = row["strand1(gene/fusion)"].split("/")[1]
+        std_row[COLUMNS.break2_strand] = row["strand2(gene/fusion)"].split("/")[1]
+        std_row[COLUMNS.event_type] = row["type"].split("/")[0]
+        std_row[COLUMNS.break1_orientation] = get_orient(row["direction1"])
+        std_row[COLUMNS.break2_orientation] = get_orient(row["direction2"])
+
+        std_row[COLUMNS.break1_position_start] = std_row[COLUMNS.break1_position_end] = b1_start
+        std_row[COLUMNS.break2_position_start] = std_row[COLUMNS.break2_position_end] = b2_start
+    except (ValueError, TypeError):
+        raise AssertionError(
+            "Could not parse the breakpoint from the Arriba row: {}, {}".format(
+                row["breakpoint1"], row["breakpoint2"]
+            )
+        )
+    return std_row

mavis/tools/constants.py

@@ -9,6 +9,7 @@
     MAVIS='mavis',
     DEFUSE='defuse',
     BREAKDANCER='breakdancer',
+    ARRIBA='arriba',
     VCF='vcf',
     BREAKSEQ='breakseq',
     CNVNATOR='cnvnator',

tests/unit/test_tool.py

@@ -105,6 +105,113 @@ def test_convert_duplication(self):
         self.assertEqual('1', bpp.break2.chr)
 
 
+
+class TestArriba(unittest.TestCase):
+    def test_convert_standard_event(self):
+        row = {
+            'breakpoint1': '13:114529969',
+            'breakpoint2': '13:114751269',
+            'type': 'inversion',
+            'strand1(gene/fusion)': '+/+',
+            'strand2(gene/fusion)': '-/-',
+            'direction1': 'downstream',
+            'direction2': 'downstream',
+        }
+        bpp_list = _convert_tool_row(row, SUPPORTED_TOOL.ARRIBA, True)
+
+        self.assertEqual(1, len(bpp_list))
+        bpp = bpp_list[0]
+        self.assertEqual('13', bpp.break1.chr)
+        self.assertEqual('13', bpp.break2.chr)
+        self.assertEqual(114529969, bpp.break1.start)
+        self.assertEqual(114751269, bpp.break2.start)
+        self.assertEqual(SVTYPE.INV, bpp.event_type)
+        self.assertEqual('L', bpp.break1.orient)
+        self.assertEqual('L', bpp.break2.orient)
+        self.assertEqual(True, bpp.opposing_strands)
+
+    def test_convert_translocation(self):
+        row = {
+            'breakpoint1': '17:69313092',
+            'breakpoint2': '20:58272875',
+            'type': 'translocation',
+            'strand1(gene/fusion)': '-/-',
+            'strand2(gene/fusion)': '-/-',
+            'direction1': 'upstream',
+            'direction2': 'downstream',
+        }
+        bpp_list = _convert_tool_row(row, SUPPORTED_TOOL.ARRIBA, True)
+
+        self.assertEqual(1, len(bpp_list))
+        bpp = bpp_list[0]
+        self.assertEqual('chr17', bpp.break1.chr)
+        self.assertEqual('chr20', bpp.break2.chr)
+        self.assertEqual(69313092, bpp.break1.start)
+        self.assertEqual(58272875, bpp.break2.start)
+        self.assertEqual(SVTYPE.TRANS, bpp.event_type)
+        self.assertEqual('R', bpp.break1.orient)
+        self.assertEqual('L', bpp.break2.orient)
+        self.assertEqual(False, bpp.opposing_strands)
+
+    def test_convert_translocation(self):
+        row = {
+            'breakpoint1': '20:57265705',
+            'breakpoint2': '20:47786405',
+            'type': 'inversion/5\'-5\'',
+            'strand1(gene/fusion)': '-/-',
+            'strand2(gene/fusion)': '-/+',
+            'direction1': 'upstream',
+            'direction2': 'upstream',
+        }
+        bpp_list = _convert_tool_row(row, SUPPORTED_TOOL.ARRIBA, True)
+
+        self.assertEqual(1, len(bpp_list))
+        bpp = bpp_list[0]
+        self.assertEqual('chr20', bpp.break1.chr)
+        self.assertEqual('chr20', bpp.break2.chr)
+        self.assertEqual(57265705, bpp.break1.start)
+        self.assertEqual(47786405, bpp.break2.start)
+        self.assertEqual(SVTYPE.INV, bpp.event_type)
+        self.assertEqual('R', bpp.break1.orient)
+        self.assertEqual('R', bpp.break2.orient)
+        self.assertEqual(True, bpp.opposing_strands)
+
+    def test_convert_translocation(self):
+        row = {
+            'breakpoint1': '14:102877322',
+            'breakpoint2': '14:102994672',
+            'type': 'deletion/read-through/5\'-5\'',
+            'strand1(gene/fusion)': '+/+',
+            'strand2(gene/fusion)': '-/+',
+            'direction1': 'downstream',
+            'direction2': 'upstream',
+        }
+        bpp_list = _convert_tool_row(row, SUPPORTED_TOOL.ARRIBA, True)
+
+        self.assertEqual(1, len(bpp_list))
+        bpp = bpp_list[0]
+        self.assertEqual('chr14', bpp.break1.chr)
+        self.assertEqual('chr14', bpp.break2.chr)
+        self.assertEqual(102877322, bpp.break1.start)
+        self.assertEqual(102994672, bpp.break2.start)
+        self.assertEqual(SVTYPE.DEL, bpp.event_type)
+        self.assertEqual('L', bpp.break1.orient)
+        self.assertEqual('R', bpp.break2.orient)
+        self.assertEqual(False, bpp.opposing_strands)
+
+    def test_malformed(self):
+        row = {
+            'breakpoint1': '',
+            'breakpoint2': None,
+            'type': 'translocation',
+            'strand1(gene/fusion)': '-/-',
+            'strand2(gene/fusion)': '-/-',
+            'direction1': 'upstream',
+            'direction2': 'downstream',
+        }
+        with self.assertRaises(AssertionError):
+            _convert_tool_row(row, SUPPORTED_TOOL.ARRIBA, False)
+
 class TestStarFusion(unittest.TestCase):
     def test_convert_standard_event(self):
         row = {