Addition of HRF_GBM energy method

database/scoring/weights/hrf_gbm.wts

@@ -0,0 +1 @@
+hrf_gbm 9.0

source/src/basic/options/options_rosetta.py

@@ -1533,6 +1533,7 @@
 
 		Option( 'covalent_labeling_input', 'File', desc='Input covalent labeling data in the form of neighbor counts'),
 		Option( 'hrf_dynamics_input', 'File', desc='Input covalent labeling data in the form of protection factors for FA'),
+		Option( 'hrf_gbm_input', 'File', desc='File that contains expected residue neighbor counts used for hrf_gbm scoring'),
 		Option( 'covalent_labeling_fa_input', 'File', desc='Input covalent labeling data in the form of neighbor counts for FA'),
 		Option( 'depc_ms_input', 'File', desc='Input DEPC covalent labeling MS data in the form of His/Lys/Ser/Thr/Tyr residue numbers and labeling status of labeled (L) or unlabeled (U)'),
 		Option( 'elec_min_dis', 'Real', desc='changes the minimum distance cut-off for hack-elec energy', default='1.6'),

source/src/core.4.src.settings

@@ -111,6 +111,7 @@ sources = {
 		"GenBornEnergy",
 		"GenericBondedEnergy",
 		"HRFDynamicsEnergy",
+		"HRF_GBM_Energy",
 		"HRF_MSLabelingEnergy",
 		"CCS_IMMSEnergy",
 		"MultipoleElecEnergy",

source/src/core/energy_methods/HRF_GBM_Energy.cc

@@ -0,0 +1,154 @@
+// -*- mode:c++;tab-width:2;indent-tabs-mode:t;show-trailing-whitespace:t;rm-trailing-spaces:t -*-
+// vi: set ts=2 noet:
+//
+// (c) Copyright Rosetta Commons Member Institutions.
+// (c) This file is part of the Rosetta software suite and is made available under license.
+// (c) The Rosetta software is developed by the contributing members of the Rosetta Commons.
+// (c) For more information, see http://www.rosettacommons.org. Questions about this can be
+// (c) addressed to University of Washington UW TechTransfer, email: license@u.washington.edu.
+
+/// @file   core/energy_methods/HRF_GBM_Energy.cc
+/// @brief  Score term assesses neighbor count agreement between decoy and expected value generated from external lightGBM model. Model uses experimental HRF data as input feature. Detailed in manuscript published in 2025.
+/// @author Elijah Day (ehday@ucla.edu)
+
+#include <core/energy_methods/HRF_GBM_Energy.hh>
+#include <core/energy_methods/HRF_GBM_EnergyCreator.hh>
+#include <core/conformation/Residue.hh>
+#include <core/pose/Pose.hh>
+#include <core/scoring/EnergyMap.hh>
+#include <core/scoring/ContextGraphTypes.hh>
+#include <core/scoring/methods/EnergyMethodOptions.hh>
+
+
+#include <utility/io/izstream.hh>
+#include <utility/vector1.hh>
+#include <numeric/NumericTraits.hh>
+
+#include <basic/options/keys/OptionKeys.hh> // AUTO IWYU For OptionKeys,
+
+namespace core {
+namespace energy_methods {
+
+
+/// @details Returns fresh instance of the HRF_GBM_Energy class.
+core::scoring::methods::EnergyMethodOP
+HRF_GBM_EnergyCreator::create_energy_method(
+	core::scoring::methods::EnergyMethodOptions const & options
+) const {
+	return utility::pointer::make_shared< HRF_GBM_Energy >( options );
+}
+
+core::scoring::ScoreTypes
+HRF_GBM_EnergyCreator::score_types_for_method() const {
+	using namespace core::scoring;
+	ScoreTypes sts;
+	sts.push_back(hrf_gbm);
+	return sts;
+}
+
+void
+HRF_GBM_Energy::setup_for_scoring(
+	pose::Pose & pose, core::scoring::ScoreFunction const &
+) const {
+	pose.update_residue_neighbors();
+
+}
+
+// constructor
+HRF_GBM_Energy::HRF_GBM_Energy( core::scoring::methods::EnergyMethodOptions const & options ) :
+	parent( utility::pointer::make_shared< HRF_GBM_EnergyCreator >() ),
+	hrf_gbm_input_file_( options.hrf_gbm_input() )
+{
+	init_from_file();
+}
+
+/// clone
+core::scoring::methods::EnergyMethodOP
+HRF_GBM_Energy::clone() const {
+	return utility::pointer::make_shared< HRF_GBM_Energy >( *this );
+}
+
+/////////////////////////////////////////////////////////////////////////////
+// scoring
+/////////////////////////////////////////////////////////////////////////////
+
+void
+HRF_GBM_Energy::residue_energy(
+	core::conformation::Residue const & residue,
+	core::pose::Pose const & pose,
+	core::scoring::EnergyMap & emap
+) const {
+
+	core::Size const number_residues (pose.total_residue());
+	core::Real pose_neighbor_count (0.0);
+	core::Real nc_from_file (0.0);
+	core::Real distance (0.0);
+	core::Real angle (0.0);
+
+
+	for ( core::Size j=1; j <= input_nc_.size(); j++ ) {
+		if ( input_nc_[j].first == residue.seqpos() ) {
+			nc_from_file = input_nc_[j].second; // Predicted neighbor count values were obtained from the external lightGBM model described in the Day 2025 manuscript. Briefly, the model takes in sequence derived features (PSSM profile, predicted SSE, etc.) and experimental mass spec data to generated an expected neighbor count for the residue. Those neighbor counts are passed as an input for the hrf_gbm score term.
+			std::string target_atom ("CB");
+			if ( residue.type().name1() == 'G' ) {
+				target_atom = "1HA";
+			}
+			for ( core::Size res_count_neighbor = 1; res_count_neighbor <= number_residues; res_count_neighbor++ ) {
+				if ( residue.seqpos() != pose.residue(res_count_neighbor).seqpos() ) {
+					std::string neighbor_atom ("CB");
+					if ( pose.residue(res_count_neighbor).type().name1() == 'G' ) {
+						neighbor_atom = "1HA";
+					}
+					distance = residue.xyz("CA").distance(pose.residue(res_count_neighbor).xyz(neighbor_atom));
+					numeric::xyzVector<core::Real> norm_CA_target_vector = (residue.xyz(target_atom) - residue.xyz("CA")).normalize()/(residue.xyz(target_atom).distance(residue.xyz("CA")));
+					numeric::xyzVector<core::Real> norm_neighbor_vector = (pose.residue(res_count_neighbor).xyz(neighbor_atom)-residue.xyz("CA"))/(pose.residue(res_count_neighbor).xyz(neighbor_atom).distance(residue.xyz("CA")));
+					angle = std::acos(norm_CA_target_vector.dot(norm_neighbor_vector));
+					pose_neighbor_count += 1.0/(1.0 + std::exp(1.0*(distance-9.0)))*1.0/(1.0 + std::exp(numeric::NumericTraits< float >::pi()*2.0*(angle-numeric::NumericTraits< float >::pi()/2.0)));
+				}
+			}
+			emap[ core::scoring::hrf_gbm] += -1.0/(1.0 + std::exp(2.9*(std::abs(pose_neighbor_count - nc_from_file)-2.1))); //2.9 and 2.1 were optimized in the Day 2025 manuscript
+		}
+	}
+}
+
+void
+HRF_GBM_Energy::indicate_required_context_graphs( utility::vector1< bool > & context_graphs_required ) const
+{
+	context_graphs_required[ core::scoring::twelve_A_neighbor_graph ] = false;
+}
+
+
+core::Size
+HRF_GBM_Energy::version() const
+{
+	return 1;
+}
+
+void HRF_GBM_Energy::init_from_file() {
+	using namespace basic::options;
+	using namespace basic::options::OptionKeys;
+
+	utility::io::izstream input(hrf_gbm_input_file_);
+	if ( !input.good() ) {
+		std::string const msg( "Error opening file: " + hrf_gbm_input_file_ );
+		utility_exit_with_message( msg );
+	}
+
+	std::string line;
+
+	while ( getline(input,line) ) {
+		if ( line.substr(0,1) == "#" ) continue;
+		std::istringstream ss(line);
+		core::Size resi;
+		core::Real nc;
+
+		ss >> resi >> nc;
+
+		runtime_assert_string_msg( !(ss.fail() || ss.bad()), "Error in HRF_GBM_Energy::init_from_file():  Could not parse line \"" + line + "\"." );
+		input_nc_.push_back( std::pair< core::Size, core::Real >(resi, nc) );
+	}
+
+}
+
+} // scoring
+} // core

source/src/core/energy_methods/HRF_GBM_Energy.fwd.hh

@@ -0,0 +1,32 @@
+// -*- mode:c++;tab-width:2;indent-tabs-mode:t;show-trailing-whitespace:t;rm-trailing-spaces:t -*-
+// vi: set ts=2 noet:
+//
+// (c) Copyright Rosetta Commons Member Institutions.
+// (c) This file is part of the Rosetta software suite and is made available under license.
+// (c) The Rosetta software is developed by the contributing members of the Rosetta Commons.
+// (c) For more information, see http://www.rosettacommons.org. Questions about this can be
+// (c) addressed to University of Washington CoMotion, email: license@uw.edu.
+
+/// @file   core/energy_methods/HRF_GBM_Energy.hh
+/// @brief  Score term using predicted neighbor counts from experimentally derived HRF-MS data
+/// @author Elijah Day (ehday@ucla.edu)
+
+
+#ifndef INCLUDED_core_energy_methods_HRF_GBM_Energy_fwd_hh
+#define INCLUDED_core_energy_methods_HRF_GBM_Energy_fwd_hh
+
+// Unit Headers
+
+// Package headers
+
+// Project headers
+
+// Utility headers
+
+
+namespace core {
+namespace energy_methods {
+class HRF_GBM_Energy;
+}
+}
+#endif // INCLUDED_core_energy_methods_HRF_GBM_Energy_fwd_hh

source/src/core/energy_methods/HRF_GBM_Energy.hh

@@ -0,0 +1,79 @@
+// -*- mode:c++;tab-width:2;indent-tabs-mode:t;show-trailing-whitespace:t;rm-trailing-spaces:t -*-
+// vi: set ts=2 noet:
+//
+// (c) Copyright Rosetta Commons Member Institutions.
+// (c) This file is part of the Rosetta software suite and is made available under license.
+// (c) The Rosetta software is developed by the contributing members of the Rosetta Commons.
+// (c) For more information, see http://www.rosettacommons.org. Questions about this can be
+// (c) addressed to University of Washington UW TechTransfer, email: license@u.washington.edu.
+
+/// @file   core/energy_methods/HRF_GBM_Energy.hh
+/// @brief  Energy term used for scoring HRF_GBM_Energy
+/// @author Elijah Day (ehday@ucla.edu)
+
+#ifndef INCLUDED_core_energy_methods_HRF_GBM_Energy_hh
+#define INCLUDED_core_energy_methods_HRF_GBM_Energy_hh
+
+#include <core/scoring/methods/ContextDependentOneBodyEnergy.hh>
+#include <core/scoring/ScoreFunction.fwd.hh>
+#include <core/pose/Pose.fwd.hh>
+#include <utility/vector1.hh>
+
+#include <core/scoring/methods/EnergyMethodOptions.fwd.hh> // AUTO IWYU For EnergyMethodOptions
+
+
+namespace core {
+namespace energy_methods {
+
+
+
+class HRF_GBM_Energy : public core::scoring::methods::ContextDependentOneBodyEnergy {
+public:
+
+	typedef core::scoring::methods::ContextDependentOneBodyEnergy parent;
+	HRF_GBM_Energy( core::scoring::methods::EnergyMethodOptions const & options );
+
+	/// clone
+	core::scoring::methods::EnergyMethodOP
+	clone() const override;
+
+	/////////////////////////////////////////////////////////////////////////////
+	// scoring
+	/////////////////////////////////////////////////////////////////////////////
+
+	void
+	residue_energy(
+		conformation::Residue const & rsd,
+		pose::Pose const & pose,
+		core::scoring::EnergyMap & emap
+	) const override;
+
+	void
+	setup_for_scoring(
+		pose::Pose & pose, core::scoring::ScoreFunction const &
+	) const override;
+
+	void
+	finalize_total_energy(
+		pose::Pose &,
+		core::scoring::ScoreFunction const &,
+		core::scoring::EnergyMap &
+	) const override {}
+
+	core::Size version() const override;
+
+	void
+	indicate_required_context_graphs(
+		utility::vector1< bool > & /*context_graphs_required*/
+	) const override;
+
+private:
+	std::string hrf_gbm_input_file_;
+	void init_from_file();
+	utility::vector1< std::pair< core::Size, core::Real > > input_nc_;
+};
+
+}
+}
+
+#endif // INCLUDED_core_energy_methods_HRF_GBM_Energy_hh

source/src/core/energy_methods/HRF_GBM_EnergyCreator.hh

@@ -0,0 +1,42 @@
+// -*- mode:c++;tab-width:2;indent-tabs-mode:t;show-trailing-whitespace:t;rm-trailing-spaces:t -*-
+// vi: set ts=2 noet:
+//
+// (c) Copyright Rosetta Commons Member Institutions.
+// (c) This file is part of the Rosetta software suite and is made available under license.
+// (c) The Rosetta software is developed by the contributing members of the Rosetta Commons.
+// (c) For more information, see http://www.rosettacommons.org. Questions about this can be
+// (c) addressed to University of Washington UW TechTransfer, email: license@u.washington.edu.
+
+/// @file   src/core/energy_methods/HRF_GBM_EnergyCreator.hh
+/// @brief  Declaration of class that connects HRF_GBM_Energy with the ScoringManager
+/// @author Elijah Day (ehday@ucla.edu)
+
+#ifndef INCLUDED_core_energy_methods_HRF_GBM_EnergyCreator_hh
+#define INCLUDED_core_energy_methods_HRF_GBM_EnergyCreator_hh
+
+#include <core/scoring/methods/EnergyMethodCreator.hh>
+
+namespace core {
+namespace energy_methods {
+
+
+class HRF_GBM_EnergyCreator : public core::scoring::methods::EnergyMethodCreator
+{
+public:
+	/// @brief Instantiate a new HRF_GBM_Energy
+	core::scoring::methods::EnergyMethodOP
+	create_energy_method(
+		core::scoring::methods::EnergyMethodOptions const &
+	) const override;
+
+	/// @brief Return the set of score types claimed by the EnergyMethod
+	/// this EnergyMethodCreator creates in its create_energy_method() function
+	core::scoring::ScoreTypes
+	score_types_for_method() const override;
+
+};
+
+}
+}
+
+#endif // INCLUDED_core_energy_methods_HRF_GBM_EnergyCreator_hh

source/src/core/init/init.cc

@@ -86,6 +86,7 @@
 #include <core/energy_methods/Burial_v2EnergyCreator.hh>
 #include <core/energy_methods/HRF_MSLabelingEnergyCreator.hh>
 #include <core/energy_methods/HRFDynamicsEnergyCreator.hh>
+#include <core/energy_methods/HRF_GBM_EnergyCreator.hh>
 #include <core/energy_methods/CCS_IMMSEnergyCreator.hh>
 #include <core/energy_methods/CCS_IMMSComplexEnergyCreator.hh>
 #include <core/energy_methods/CCS_IMMS_with_CryoEMEnergyCreator.hh>
@@ -487,6 +488,7 @@ static EnergyMethodRegistrator< energy_methods::CovalentLabelingEnergyCreator> C
 static EnergyMethodRegistrator< energy_methods::CovalentLabelingFAEnergyCreator> CovalentLabelingFAEnergyCreator_registrator;
 static EnergyMethodRegistrator< energy_methods::DEPC_MS_EnergyCreator > DEPC_MS_EnergyCreator_registrator;
 static EnergyMethodRegistrator< energy_methods::HRFDynamicsEnergyCreator > HRFDynamicsEnergyCreator_registrator;
+static EnergyMethodRegistrator< energy_methods::HRF_GBM_EnergyCreator > HRF_GBM_EnergyCreator_registrator;
 static EnergyMethodRegistrator< energy_methods::HRF_MSLabelingEnergyCreator > HRF_MSLabelingEnergyCreator_registrator;
 static EnergyMethodRegistrator< energy_methods::CCS_IMMSEnergyCreator > CCS_IMMSEnergyCreator_registrator;
 static EnergyMethodRegistrator< energy_methods::CCS_IMMSComplexEnergyCreator > CCS_IMMSComplexEnergyCreator_registrator;

source/src/core/scoring/ScoreType.hh

@@ -477,6 +477,7 @@ enum ScoreType {
 	covalent_labeling_fa, //FA form of covalent labeling
 	hrf_ms_labeling,
 	hrf_dynamics,
+	hrf_gbm,
 	ccs_imms,
 	ccs_imms_complex,
 	ccs_imms_cryoem,

source/src/core/scoring/ScoreTypeManager.cc

@@ -337,6 +337,7 @@ ScoreTypeManager::setup_score_type_names()
 	name2score_type_[ "covalent_labeling_fa" ] = covalent_labeling_fa;
 	name2score_type_[ "hrf_ms_labeling" ] = hrf_ms_labeling;
 	name2score_type_[ "hrf_dynamics" ] = hrf_dynamics;
+	name2score_type_[" hrf_gbm" ] = hrf_gbm;
 	name2score_type_[ "ccs_imms" ] = ccs_imms;
 	name2score_type_[ "ccs_imms_complex" ] = ccs_imms_complex;
 	name2score_type_[ "ccs_imms_cryoem" ] = ccs_imms_cryoem;