feat: update tx support to latest cdot with MANE label transfer (#245) (

#248)

Cargo.toml

@@ -33,17 +33,18 @@ async-compression = { version = "0.4", features = ["tokio", "gzip"] }
 bgzip = "0.3"
 bio = "1.3"
 biocommons-bioutils = "0.1.4"
-byte-unit = "4.0"
 byteorder = "1.4"
+byte-unit = "4.0"
 chrono = "0.4"
-clap = { version = "4.4", features = ["derive"] }
 clap-verbosity-flag = "2.0"
+clap = { version = "4.4", features = ["derive"] }
 csv = "1.3"
 derivative = "2.2"
+derive_builder = { version = "0.12", features = ["clippy"] }
 env_logger = "0.10"
 flate2 = "1.0"
-futures = "0.3.29"
-hgvs = "0.12"
+futures = "0.3"
+hgvs = "0.13"
 indexmap = { version = "2.1", features = ["serde"] }
 indicatif = "0.17"
 jsonl = "4.0"
@@ -52,10 +53,10 @@ log = "0.4"
 nom = "7.1"
 noodles-bgzf = { version = "0.25", features = ["async"] }
 noodles-core = "0.12"
-noodles-csi = "0.25"
+noodles-csi = "0.26"
 noodles-fasta = "0.30"
-noodles-tabix = "0.31"
-noodles-vcf = { version = "0.43", features = ["async"] }
+noodles-tabix = "0.32"
+noodles-vcf = { version = "0.45", features = ["async"] }
 parse-display = "0.8"
 procfs = "0.16"
 prost = "0.12"
@@ -64,16 +65,17 @@ rand = "0.8"
 rand_core = "0.6"
 rocksdb = { version = "0.21", features = ["multi-threaded-cf"] }
 rustc-hash = "1.1"
-seqrepo = "0.8"
-serde = { version = "1.0", features = ["derive"] }
+seqrepo = "0.9"
 serde_json = "1.0"
+serde = { version = "1.0", features = ["derive"] }
 serde_with = { version = "3.3", features = ["indexmap_2"] }
+serde_yaml = "0.9"
 strum = { version = "0.25", features = ["derive"] }
 tempfile = "3"
 thousands = "0.2"
 tokio = { version = "1.33", features = ["full"] }
-tracing = { version = "0.1", features = ["log"] }
 tracing-subscriber = "0.3"
+tracing = { version = "0.1", features = ["log"] }
 uuid = { version = "1.4", features = ["fast-rng", "serde"] }
 zstd = "0.13"
 
@@ -83,8 +85,9 @@ prost-build = "0.12"
 [dev-dependencies]
 async-std = { version = "1.12", features = ["attributes"] }
 csv = "1.3"
-hxdmp = "0.2.1"
+hxdmp = "0.2"
 insta = { version = "1.34", features = ["yaml"] }
 pretty_assertions = "1.4"
 rstest = "0.18"
 temp_testdir = "0.2"
+tracing-test = "0.2"

README.md

@@ -166,8 +166,9 @@ cargo run --release -- \
     -v \
     db create txs \
         --path-out /tmp/txs-out.bin.zst \
-        --path-cdot-json ../cdot-0.2.12.ensembl.grch37_grch38.json.gz \
-        --path-cdot-json ../cdot-0.2.12.refseq.grch37_grch38.json.gz \
+        --path-lable-tsv PATH_TO_MANE_LABEL.tsv \
+        --path-cdot-json ../cdot-0.2.21.ensembl.grch37_grch38.json.gz \
+        --path-cdot-json ../cdot-0.2.21.refseq.grch37_grch38.json.gz \
         --path-seqrepo-instance ../hgvs-rs-data/seqrepo-data/master/master
 ```
 

build.rs

@@ -1,6 +1,14 @@
 // The custom build script, needed as we use prost.
 
 fn main() {
-    println!("cargo:rerun-if-changed=src/db/create/txs/data.proto3");
-    prost_build::compile_protos(&["src/db/create/txs/data.proto3"], &["src/"]).unwrap();
+    println!("cargo:rerun-if-changed=src/proto/data.proto3");
+    prost_build::Config::new()
+        .protoc_arg("-Isrc/proto")
+        // Add serde serialization and deserialization to the generated code.
+        .type_attribute(".", "#[derive(serde::Serialize, serde::Deserialize)]")
+        // Skip serializing `None` values.
+        .type_attribute(".", "#[serde_with::skip_serializing_none]")
+        // Define the protobuf files to compile.
+        .compile_protos(&["src/proto/data.proto3"], &["src/"])
+        .unwrap();
 }

docs/db_build.md

@@ -164,8 +164,8 @@ $ mehari db create txs \
     \
     --path-seqrepo-instance path/to/seqrepo-data/master \
     \
-    --path-cdot-json cdot-0.2.12.refseq.grch37_grch38.json \
-    --path-cdot-json cdot-0.2.12.ensembl.grch37_grch38.json \
+    --path-cdot-json cdot-0.2.21.refseq.grch37_grch38.json \
+    --path-cdot-json cdot-0.2.21.ensembl.grch37_grch38.json \
     \
     --path-seqrepo-instance path/to/seqrepo-data/master \
     \

src/annotate/seqvars/ann.rs

@@ -361,16 +361,19 @@ impl FromStr for FeatureType {
     strum::EnumIter,
 )]
 pub enum FeatureBiotype {
+    /// Is coding transcript.
     Coding,
+    /// Is non-coding transcript.
     Noncoding,
+    /// Is in MANE Select set.
+    ManeSelect,
+    /// Is in MANE Plus Clinical set.
+    ManePlusClinical,
 }
 
 impl FeatureBiotype {
     pub fn is_coding(&self) -> bool {
-        match self {
-            FeatureBiotype::Coding => true,
-            FeatureBiotype::Noncoding => false,
-        }
+        matches!(self, FeatureBiotype::Coding)
     }
 }
 
@@ -509,7 +512,7 @@ pub struct AnnField {
     /// The feature identifier.
     pub feature_id: String,
     /// The feature biotype.
-    pub feature_biotype: FeatureBiotype,
+    pub feature_biotype: Vec<FeatureBiotype>,
     /// The exon / intron rank.
     pub rank: Option<Rank>,
     /// HGVS c. notation.
@@ -542,7 +545,7 @@ impl Default for AnnField {
                 term: SoFeature::Transcript,
             },
             feature_id: Default::default(),
-            feature_biotype: FeatureBiotype::Coding,
+            feature_biotype: vec![FeatureBiotype::Coding],
             rank: Default::default(),
             hgvs_t: Default::default(),
             hgvs_p: Default::default(),
@@ -572,7 +575,13 @@ impl FromStr for AnnField {
         let gene_id = fields.next().unwrap().to_string();
         let feature_type = fields.next().unwrap().parse()?;
         let feature_id = fields.next().unwrap().to_string();
-        let feature_biotype = fields.next().unwrap().parse()?;
+        let feature_biotype = fields
+            .next()
+            .unwrap()
+            .split('&')
+            .map(|s| s.parse())
+            .collect::<Result<Vec<_>, _>>()
+            .map_err(|e| anyhow::anyhow!("could not parse feature biotype: {}", e))?;
         let rank = fields.next().unwrap();
         let rank = if rank.is_empty() {
             None
@@ -669,7 +678,15 @@ impl std::fmt::Display for AnnField {
         write!(f, "|")?;
         write!(f, "{}", self.feature_id)?;
         write!(f, "|")?;
-        write!(f, "{}", self.feature_biotype)?;
+        write!(
+            f,
+            "{}",
+            self.feature_biotype
+                .iter()
+                .map(|t| format!("{}", t))
+                .collect::<Vec<_>>()
+                .join("&")
+        )?;
         write!(f, "|")?;
         if let Some(rank) = &self.rank {
             write!(f, "{}", rank)?;
@@ -1082,7 +1099,7 @@ mod test {
                 term: SoFeature::Transcript,
             },
             feature_id: String::from("feature_id"),
-            feature_biotype: FeatureBiotype::Coding,
+            feature_biotype: vec![FeatureBiotype::Coding],
             rank: Some(Rank { ord: 1, total: 2 }),
             hgvs_t: Some(String::from("HGVS.c")),
             hgvs_p: Some(String::from("HGVS.p")),