Add in selenocysteine and pyrrolysine amino acids

- Bump to v0.5.1

crates/sage-cli/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "sage-cli"
-version = "0.5.0"
+version = "0.5.1"
 authors = ["Michael Lazear <[email protected]"]
 edition = "2021"
 rust-version = "1.62"

crates/sage-cloudpath/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "sage-cloudpath"
-version = "0.5.0"
+version = "0.5.1"
 authors = ["Michael Lazear <[email protected]"]
 edition = "2021"
 rust-version = "1.62"

crates/sage/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "sage-core"
-version = "0.5.0"
+version = "0.5.1"
 authors = ["Michael Lazear <[email protected]"]
 edition = "2021"
 rust-version = "1.62"

crates/sage/src/mass.rs

@@ -54,9 +54,9 @@ impl Mass for Residue {
     }
 }
 
-pub const VALID_AA: [char; 20] = [
+pub const VALID_AA: [char; 22] = [
     'A', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'K', 'L', 'M', 'N', 'P', 'Q', 'R', 'S', 'T', 'V', 'W',
-    'Y',
+    'Y', 'U', 'O',
 ];
 
 impl Mass for char {
@@ -82,7 +82,9 @@ impl Mass for char {
             'W' => 186.079_32,
             'Y' => 163.063_32,
             'V' => 99.068_41,
-            _ => unreachable!("BUG: invalid amino acid"),
+            'U' => 150.95363,
+            'O' => 237.14773,
+            _ => unreachable!("BUG: invalid amino acid {}", self),
         }
     }
 }

crates/sage/src/ml/retention_model.rs

@@ -43,9 +43,9 @@ pub struct RetentionModel {
     pub rt_max: f64,
 }
 
-const FEATURES: usize = 63;
-const N_TERMINAL: usize = 20;
-const C_TERMINAL: usize = 40;
+const FEATURES: usize = VALID_AA.len() * 3 + 3;
+const N_TERMINAL: usize = VALID_AA.len() * 1;
+const C_TERMINAL: usize = VALID_AA.len() * 2;
 const PEPTIDE_LEN: usize = FEATURES - 3;
 const PEPTIDE_MASS: usize = FEATURES - 2;
 const INTERCEPT: usize = FEATURES - 1;