Merge pull request #67 from rusteomics/glycan-rendering

Glycan rendering

Author: douweschulte

.github/workflows/rust.yml

@@ -28,6 +28,8 @@ jobs:
         cargo build -p rustyms --no-default-features --features rand
         cargo build -p rustyms --no-default-features --features rayon
         cargo build -p rustyms --no-default-features --features mzdata
+        cargo build -p rustyms --no-default-features --features glycan-render
+        cargo build -p rustyms --no-default-features --features glycan-render-bitmap
   
   fmt:
     runs-on: ubuntu-latest

.github/workflows/scripts/update-all-databases.sh

@@ -41,11 +41,9 @@ function make-ontologies {
     curl https://raw.githubusercontent.com/HUPO-PSI/mzIdentML/master/cv/XLMOD.obo \
         > ${db_data}/XLMOD.obo
     curl -L http://purl.obolibrary.org/obo/gno.obo \
-        | sed '/(property_value: GNO:00000(022|023|041|042|101|102) .*$\n)|(def: .*$\n)/d' \
-        | gzip -c \
-        > ${db_data}/GNOme.obo.gz
+        > ${db_data}/GNOme.obo
     curl -L https://glycosmos.org/download/glycosmos_glycans_list.csv \
-        | gzip -c > ${db_data}/glycosmos_glycans_list.csv.gz
+        > ${db_data}/glycosmos_glycans_list.csv
 
 
     echo "Serializing the other databases..."

.gitignore

@@ -7,6 +7,7 @@ Cargo.lock
 errors.dat
 *.dat.Z
 *.dat
+*.html
 GNOme.obo
 glycosmos_glycans_list.csv
 .venv/

Cargo.toml

@@ -27,15 +27,17 @@ codegen-units = 1
 
 [workspace.dependencies]
 afl = "0.15"
+base64 = "0.22"
 bincode = "1.3"
 clap = { version = "4.5", features = ["derive", "cargo"] }
 directories = "6.0"
 flate2 = "1.0"
 iai-callgrind = "0.14"
 itertools = "0.14"
-mzdata = "0.44"
+mzdata = {version="0.48", default-features = false, features = ["miniz_oxide"]}
 ndarray = "0.16"
 ordered-float = { version = "4.6", features = ["serde"] }
+png = "0.17"
 probability = "0.20"
 pyo3 = "0.23"
 rand = "0.9"
@@ -45,8 +47,10 @@ roxmltree = "0.20"
 serde = { version = "1.0", features = ["derive", "rc"] }
 serde_json = "1.0"
 similar = "2.7"
+swash = "0.2"
 thin-vec = { version = "0.2", features = ["serde"] }
-uom = { version = "0.36", features = ["use_serde", "usize", "isize"] }
+uom = { version = "0.36", default-features = false, features = ["use_serde", "usize", "isize", "f64"] }
+zeno = {version = "0.3.2" }
 
 [workspace.lints.rust]
 unexpected_cfgs = { level = "allow", check-cfg = [

examples/de-novo-align/Cargo.toml

@@ -7,7 +7,7 @@ license.workspace = true
 publish = false
 
 [dependencies]
-rustyms = { path = "../../rustyms" }
+rustyms = { path = "../../rustyms", default-features=false, features = ["align", "identification"] }
 clap = { workspace = true }
 itertools = { workspace = true }
 rayon = { workspace = true }

rustyms-generate-databases/src/gnome.rs

@@ -117,7 +117,7 @@ impl std::str::FromStr for GnoSubsumption {
 }
 
 fn parse_gnome() -> HashMap<String, GNOmeModification> {
-    let obo = OboOntology::from_file("rustyms-generate-databases/data/GNOme.obo.gz")
+    let obo = OboOntology::from_file("rustyms-generate-databases/data/GNOme.obo")
         .expect("Not a valid obo file");
     let mut mods = HashMap::new();
 
@@ -221,7 +221,7 @@ fn parse_gnome_structures() -> HashMap<String, GlycosmosList> {
     let mut glycans = HashMap::new();
     let mut errors = 0;
     for line in parse_csv(
-        "rustyms-generate-databases/data/glycosmos_glycans_list.csv.gz",
+        "rustyms-generate-databases/data/glycosmos_glycans_list.csv",
         b',',
         None,
     )

rustyms/Cargo.toml

@@ -13,7 +13,6 @@ repository = "https://github.com/snijderlab/rustyms"
 readme = "README.md"
 include = [
     "src/**/*",
-    "databases/**/*.gz",
     "README.md",
     "build.rs",
     "benches/**/*",
@@ -32,12 +31,17 @@ rayon = { workspace = true, optional = true }
 regex = { workspace = true }
 serde = { workspace = true }
 similar = { workspace = true }
+swash = {workspace = true, optional = true}
 thin-vec = { workspace = true }
 uom = { workspace = true }
+zeno = { workspace = true, optional = true }
 
 [dev-dependencies]
+base64 = { workspace = true }
 iai-callgrind = { workspace = true }
+png = { workspace = true }
 serde_json = { workspace = true }
+directories = {workspace = true}
 
 [features]
 default = [
@@ -48,11 +52,15 @@ default = [
     "isotopes",
     "rand",
     "mzdata",
+    "glycan-render",
+    "glycan-render-bitmap",
 ]
 imgt = []
 align = []
 identification = []
 isotopes = ["probability", "ndarray"]
+glycan-render = []
+glycan-render-bitmap = ["zeno", "swash", "glycan-render"]
 
 [[bench]]
 name = "iai"

rustyms/README.md

@@ -80,3 +80,5 @@ It has multiple features which allow you to slim it down if needed (all are enab
 * `rand` - allows the generation of random peptides.
 * `rayon` - enables parallel iterators using rayon, mostly for `imgt` but also in consecutive align.
 * `mzdata` - enables integration with [mzdata](https://github.com/mobiusklein/mzdata) which has more advanced raw file support.
+* `glycan-render` - enables the rendering to SVGs for glycans and glycan fragments
+* `glycan-render-bitmap` - enables the rendering to bitmaps for glycans, by enabling the optional dependencies zeno and swash

rustyms/data/glycan.mgf

@@ -2,7 +2,7 @@ BEGIN IONS
 PEPMASS=660.2457879192369
 CHARGE=1+
 TITLE=MS/MS scan at 1.535 min with Intensity: 604.0
-SEQUENCE=N[GlycanStructure:Hex(Hex,HexNAc)]
+SEQUENCE=N[G:G01141WK]
 
 189.48956 5050.0
 283.62076 5050.0