Feat/ims lfq

CHANGELOG.md

@@ -6,6 +6,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [v0.15.0-alpha] (unreleased)
 ### Added
+- Initial support for LFQ on data with ion mobility.
+- Speedup on the generation of databases when large number of peptides are redundant.
 - Initial support for searching diaPASEF data
 - `override_precursor_charge` setting that forces multiple charge states to be searched
 ### Breaking Changes

crates/sage-cli/src/input.rs

@@ -80,6 +80,7 @@ pub struct LfqOptions {
     pub integration: Option<sage_core::lfq::IntegrationStrategy>,
     pub spectral_angle: Option<f64>,
     pub ppm_tolerance: Option<f32>,
+    pub mobility_pct_tolerance: Option<f32>,
     pub combine_charge_states: Option<bool>,
 }
 
@@ -91,13 +92,22 @@ impl From<LfqOptions> for LfqSettings {
             integration: value.integration.unwrap_or(default.integration),
             spectral_angle: value.spectral_angle.unwrap_or(default.spectral_angle).abs(),
             ppm_tolerance: value.ppm_tolerance.unwrap_or(default.ppm_tolerance).abs(),
+            mobility_pct_tolerance: value
+                .mobility_pct_tolerance
+                .unwrap_or(default.mobility_pct_tolerance),
             combine_charge_states: value
                 .combine_charge_states
                 .unwrap_or(default.combine_charge_states),
         };
         if settings.ppm_tolerance > 20.0 {
             log::warn!("lfq_settings.ppm_tolerance is higher than expected");
         }
+        if settings.mobility_pct_tolerance > 4.0 {
+            log::warn!("lfq_settings.mobility_pct_tolerance is higher than expected");
+        }
+        if settings.mobility_pct_tolerance < 0.05 {
+            log::warn!("lfq_settings.mobility_pct_tolerance is smaller than expected");
+        }
         if settings.spectral_angle < 0.50 {
             log::warn!("lfq_settings.spectral_angle is lower than expected");
         }
@@ -221,6 +231,7 @@ impl Input {
     fn check_tolerances(tolerance: &Tolerance) {
         let (lo, hi) = match tolerance {
             Tolerance::Ppm(lo, hi) => (*lo, *hi),
+            Tolerance::Pct(_, _) => unreachable!("Pct tolerance should never be used on mz"),
             Tolerance::Da(lo, hi) => (*lo, *hi),
         };
         if hi.abs() > lo.abs() {

crates/sage-cli/src/output.rs

@@ -1,10 +1,10 @@
 use rayon::prelude::*;
-use sage_core::spectrum::ProcessedSpectrum;
+use sage_core::spectrum::MS1Spectra;
 use sage_core::{scoring::Feature, tmt::TmtQuant};
 
 #[derive(Default)]
 pub struct SageResults {
-    pub ms1: Vec<ProcessedSpectrum>,
+    pub ms1: MS1Spectra,
     pub features: Vec<Feature>,
     pub quant: Vec<TmtQuant>,
 }
@@ -19,7 +19,34 @@ impl FromParallelIterator<SageResults> for SageResults {
             .reduce(SageResults::default, |mut acc, x| {
                 acc.features.extend(x.features);
                 acc.quant.extend(x.quant);
-                acc.ms1.extend(x.ms1);
+                match (acc.ms1, x.ms1) {
+                    (MS1Spectra::NoMobility(mut a), MS1Spectra::NoMobility(b)) => {
+                        a.extend(b);
+                        acc.ms1 = MS1Spectra::NoMobility(a);
+                    }
+                    (MS1Spectra::WithMobility(mut a), MS1Spectra::WithMobility(b)) => {
+                        a.extend(b);
+                        acc.ms1 = MS1Spectra::WithMobility(a);
+                    }
+                    (MS1Spectra::Empty, MS1Spectra::Empty) => {
+                        acc.ms1 = MS1Spectra::Empty;
+                    }
+                    (MS1Spectra::Empty, MS1Spectra::WithMobility(a))
+                    | (MS1Spectra::WithMobility(a), MS1Spectra::Empty) => {
+                        acc.ms1 = MS1Spectra::WithMobility(a);
+                    }
+                    (MS1Spectra::Empty, MS1Spectra::NoMobility(a))
+                    | (MS1Spectra::NoMobility(a), MS1Spectra::Empty) => {
+                        acc.ms1 = MS1Spectra::NoMobility(a);
+                    }
+                    _ => {
+                        // In theory this can happen if someone is searching
+                        // together files of different types, mixing the ones
+                        // that support IMS and the ones that dont.
+                        // ... I dont think this should be run-time recoverable.
+                        unreachable!("Found combination of MS1 spectra with and without mobility.")
+                    }
+                };
                 acc
             })
     }
@@ -35,7 +62,30 @@ impl FromIterator<SageResults> for SageResults {
             .fold(SageResults::default(), |mut acc, x| {
                 acc.features.extend(x.features);
                 acc.quant.extend(x.quant);
-                acc.ms1.extend(x.ms1);
+                match (acc.ms1, x.ms1) {
+                    (MS1Spectra::NoMobility(mut a), MS1Spectra::NoMobility(b)) => {
+                        a.extend(b);
+                        acc.ms1 = MS1Spectra::NoMobility(a);
+                    }
+                    (MS1Spectra::WithMobility(mut a), MS1Spectra::WithMobility(b)) => {
+                        a.extend(b);
+                        acc.ms1 = MS1Spectra::WithMobility(a);
+                    }
+                    (MS1Spectra::Empty, MS1Spectra::Empty) => {
+                        acc.ms1 = MS1Spectra::Empty;
+                    }
+                    (MS1Spectra::Empty, MS1Spectra::WithMobility(a))
+                    | (MS1Spectra::WithMobility(a), MS1Spectra::Empty) => {
+                        acc.ms1 = MS1Spectra::WithMobility(a);
+                    }
+                    (MS1Spectra::Empty, MS1Spectra::NoMobility(a))
+                    | (MS1Spectra::NoMobility(a), MS1Spectra::Empty) => {
+                        acc.ms1 = MS1Spectra::NoMobility(a);
+                    }
+                    _ => {
+                        unreachable!("Found combination of MS1 spectra with and without mobility.")
+                    }
+                };
                 acc
             })
     }