From 6153471ee93ba8bf4839b4aedb2593661188bfb4 Mon Sep 17 00:00:00 2001
From: Michael Lazear <michaellazear92@gmail.com>
Date: Mon, 31 Oct 2022 10:16:38 -0700
Subject: [PATCH] Add in selenocysteine and pyrrolysine amino acids

- Bump to v0.5.1
---
 Cargo.lock                            | 6 +++---
 crates/sage-cli/Cargo.toml            | 2 +-
 crates/sage-cloudpath/Cargo.toml      | 2 +-
 crates/sage/Cargo.toml                | 2 +-
 crates/sage/src/mass.rs               | 8 +++++---
 crates/sage/src/ml/retention_model.rs | 6 +++---
 6 files changed, 14 insertions(+), 12 deletions(-)

diff --git a/Cargo.lock b/Cargo.lock
index 81445ce..57b878d 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -1176,7 +1176,7 @@ checksum = "4501abdff3ae82a1c1b477a17252eb69cee9e66eb915c1abaa4f44d873df9f09"
 
 [[package]]
 name = "sage-cli"
-version = "0.5.0"
+version = "0.5.1"
 dependencies = [
  "anyhow",
  "clap",
@@ -1195,7 +1195,7 @@ dependencies = [
 
 [[package]]
 name = "sage-cloudpath"
-version = "0.5.0"
+version = "0.5.1"
 dependencies = [
  "async-compression",
  "aws-config",
@@ -1209,7 +1209,7 @@ dependencies = [
 
 [[package]]
 name = "sage-core"
-version = "0.5.0"
+version = "0.5.1"
 dependencies = [
  "async-compression",
  "base64",
diff --git a/crates/sage-cli/Cargo.toml b/crates/sage-cli/Cargo.toml
index 7737ee2..d98135c 100644
--- a/crates/sage-cli/Cargo.toml
+++ b/crates/sage-cli/Cargo.toml
@@ -1,6 +1,6 @@
 [package]
 name = "sage-cli"
-version = "0.5.0"
+version = "0.5.1"
 authors = ["Michael Lazear <michaellazear92@gmail.com"]
 edition = "2021"
 rust-version = "1.62"
diff --git a/crates/sage-cloudpath/Cargo.toml b/crates/sage-cloudpath/Cargo.toml
index dcdb739..898547e 100644
--- a/crates/sage-cloudpath/Cargo.toml
+++ b/crates/sage-cloudpath/Cargo.toml
@@ -1,6 +1,6 @@
 [package]
 name = "sage-cloudpath"
-version = "0.5.0"
+version = "0.5.1"
 authors = ["Michael Lazear <michaellazear92@gmail.com"]
 edition = "2021"
 rust-version = "1.62"
diff --git a/crates/sage/Cargo.toml b/crates/sage/Cargo.toml
index febe629..f32c6fe 100644
--- a/crates/sage/Cargo.toml
+++ b/crates/sage/Cargo.toml
@@ -1,6 +1,6 @@
 [package]
 name = "sage-core"
-version = "0.5.0"
+version = "0.5.1"
 authors = ["Michael Lazear <michaellazear92@gmail.com"]
 edition = "2021"
 rust-version = "1.62"
diff --git a/crates/sage/src/mass.rs b/crates/sage/src/mass.rs
index a74c869..8106b88 100644
--- a/crates/sage/src/mass.rs
+++ b/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),
         }
     }
 }
diff --git a/crates/sage/src/ml/retention_model.rs b/crates/sage/src/ml/retention_model.rs
index 6166f1d..898efb8 100644
--- a/crates/sage/src/ml/retention_model.rs
+++ b/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;