Merge branch 'jdaw/remove-plasmid-polya-option' into 'release-v0.6.0'

Remove plasmid polyA configurability See merge request machine-learning/dorado!903
nanoporetech · Mar 22, 2024 · ce957ae · ce957ae
2 parents ddac0ef + 91b6c4e
commit ce957ae
Show file tree

Hide file tree

Showing 3 changed files with 4 additions and 50 deletions.
diff --git a/documentation/PolyTailConfig.md b/documentation/PolyTailConfig.md
@@ -5,7 +5,6 @@ Dorado supports estimation of Poly(A/T) tails for DNA (PCS AND PCB) and RNA samp
 Dorado also supports additional features that can be customized through a configuration file (described below):
 * Custom primer sequence for cDNA tail estimation
 * Clustering of interrupted Poly(A/T) tails
-* Estimation of Poly(A/T) length in plasmids
 
 ## Poly(A/T) Reference Diagram
 
@@ -25,16 +24,6 @@ dRNA
 3' ---- ADAPTER ---- poly(A) ---- RNA ---- 5'
 ```
 
-```
-Plasmid
-
-5' ---- ADAPTER ---- DNA ---- FRONT_FLANK ---- poly(A) ---- REAR_FLANK --- DNA ---- 3'
-
-OR
-
-5' ---- ADAPTER ---- RC(DNA) ---- RC(REAR_FLANK) ---- poly(T) ---- RC(FRONT_FLANK) ---- RC(DNA) ---- 3'
-```
-
 ## Configuration Format
 
 The configuration file needs to be in the `toml` format.
@@ -43,8 +32,6 @@ The configuration file needs to be in the `toml` format.
 [anchors]
 front_primer = "ATCG"
 rear_primer = "CGTA"
-plasmid_front_flank = "CGATCG"
-plasmid_rear_flank = "TGACTGC"
 
 [threshold]
 flank_threshold = 10
@@ -59,7 +46,5 @@ tail_interrupt_length = 10
 | -- | -- |
 | front_primer | Front primer sequence for cDNA |
 | rear_primer | Rear primer sequence for cDNA |
-| plasmid_front_flank | Front flanking sequence of poly(A) in plasmid |
-| plasmid_rear_flank | Rear flanking sequence of poly(A) in plasmid |
 | flank_threshold  | The edit distance threshold to use for detection of the flank/primer sequences |
 | tail_interrupt_length | Combine tails that are within this distance of each other (default is 0, i.e. don't combine any) |
diff --git a/dorado/poly_tail/poly_tail_config.cpp b/dorado/poly_tail/poly_tail_config.cpp
@@ -29,19 +29,6 @@ PolyTailConfig prepare_config(std::istream& is) {
             config.front_primer = toml::find<std::string>(anchors, "front_primer");
             config.rear_primer = toml::find<std::string>(anchors, "rear_primer");
         }
-
-        if (anchors.contains("plasmid_front_flank") || anchors.contains("plasmid_rear_flank")) {
-            if (!(anchors.contains("plasmid_front_flank") &&
-                  anchors.contains("plasmid_rear_flank"))) {
-                throw std::runtime_error(
-                        "Both plasmid_front_flank and plasmid_rear_flank must be provided in "
-                        "the PolyA configuration file.");
-            }
-            config.plasmid_front_flank = toml::find<std::string>(anchors, "plasmid_front_flank");
-            config.plasmid_rear_flank = toml::find<std::string>(anchors, "plasmid_rear_flank");
-            config.is_plasmid = true;
-            config.flank_threshold = 10;  // reduced default for plasmids
-        }
     }
 
     if (config_toml.contains("threshold")) {

diff --git a/tests/PolyACalculatorTest.cpp b/tests/PolyACalculatorTest.cpp
@@ -106,7 +106,6 @@ TEST_CASE("PolyTailConfig: Test parsing file", TEST_GROUP) {
     }
 
     SECTION("Only one primer is provided") {
-        auto path = (tmp_dir.m_path / "only_one_primer.toml").string();
         const toml::value data{{"anchors", toml::table{{"front_primer", "ACTG"}}}};
         const std::string fmt = toml::format(data);
         std::stringstream buffer(fmt);
@@ -116,23 +115,11 @@ TEST_CASE("PolyTailConfig: Test parsing file", TEST_GROUP) {
                           "configuration file.");
     }
 
-    SECTION("Only one plasmid flank is provided") {
-        auto path = (tmp_dir.m_path / "only_one_flank.toml").string();
-        const toml::value data{{"anchors", toml::table{{"plasmid_rear_flank", "ACTG"}}}};
-        const std::string fmt = toml::format(data);
-        std::stringstream buffer(fmt);
-
-        CHECK_THROWS_WITH(dorado::poly_tail::prepare_config(buffer),
-                          "Both plasmid_front_flank and plasmid_rear_flank must be provided in the "
-                          "PolyA configuration file.");
-    }
-
     SECTION("Parse all supported configs") {
-        auto path = (tmp_dir.m_path / "only_one_flank.toml").string();
-        const toml::value data{{"anchors", toml::table{{"plasmid_front_flank", "CGTA"},
-                                                       {"plasmid_rear_flank", "ACTG"},
-                                                       {"front_primer", "AAAAAA"},
-                                                       {"rear_primer", "GGGGGG"}}},
+        const toml::value data{{"anchors",
+                                toml::table{
+
+                                        {"front_primer", "AAAAAA"}, {"rear_primer", "GGGGGG"}}},
                                {"tail", toml::table{{"tail_interrupt_length", 10}}}};
         const std::string fmt = toml::format(data);
         std::stringstream buffer(fmt);
@@ -142,11 +129,6 @@ TEST_CASE("PolyTailConfig: Test parsing file", TEST_GROUP) {
         CHECK(config.rc_front_primer == "TTTTTT");
         CHECK(config.rear_primer == "GGGGGG");
         CHECK(config.rc_rear_primer == "CCCCCC");
-        CHECK(config.plasmid_front_flank == "CGTA");
-        CHECK(config.rc_plasmid_front_flank == "TACG");
-        CHECK(config.plasmid_rear_flank == "ACTG");
-        CHECK(config.rc_plasmid_rear_flank == "CAGT");
-        CHECK(config.is_plasmid);  // Since the plasmid flanks were specified
         CHECK(config.tail_interrupt_length == 10);
     }
 }