High resolution comparative modeling with RosettaCM

By Yifan Song, Frank DiMaio, R. Wang, David E. Kim, Chris Miles, T. Brunette, James M. Thompson, David Baker. 2013

1,047 citations recorded by the source corpus.

We describe an improved method for comparative modeling, RosettaCM, which optimizes a physically realistic all-atom energy function over the conformational space defined by homologous structures. Given a set of sequence alignments, RosettaCM assembles topologies by recombining aligned segments in Cartesian-space and building unaligned regions de novo in torsion space. The junctions between segments are regularized using a loop-closure method combining fragment superposition with gradient-based minimization. The energies of the resulting models are optimized by all-atom refinement, and the most representative low energy model is selected. The CASP10 experiment suggests RosettaCM yields models with more accurate sidechain and backbone conformations than other methods when the sequence identity to the templates is greater than ∼15%.

Subfields: Materials Chemistry, Molecular Biology, Computational Biology & Bioinformatics

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Materials Chemistry

1. David Baker
2. Yang Zhang
3. Jeffrey J. Gray
4. Jinbo Xu
5. Hahnbeom Park
6. Tanja Kortemme
7. Sergey Ovchinnikov
8. Frank DiMaio

In Molecular Biology

1. David Baker
2. Yang Zhang
3. Jeffrey J. Gray
4. Tanja Kortemme
5. Jinbo Xu
6. Hahnbeom Park
7. Sergey Ovchinnikov
8. Jens Meiler

In Computational Biology & Bioinformatics

1. David Baker
2. Jinbo Xu
3. Yang Zhang
4. Sergey Ovchinnikov
5. Robin Pearce
6. Jeffrey J. Gray
7. Hahnbeom Park
8. Tanja Kortemme

Citations in the corpus

• Improving homology modeling from low-sequence identity templates in Rosetta: A case study in GPCRs (2020, methodology): “With all input files in hand, target sequences were threaded onto the pre-aligned templates using Rosetta’s partial_thread application [23].”
• Distance-based protein folding powered by deep learning (2018, methodology): “RaptorX-TBM used our new threading program DeepThreader (29) to build sequence–template alignments and then Rosetta-CM (40) to build 3D models from alignment.”
• Protein Structure Determination using Metagenome sequence data (2017, neutral): “for​ sep_x (0,1,2) for​ sep_y (1,2,4, 8,16,32) ini_mtx = ​initialize_matrix ​(sep_x ​, ​sep_y) for​ gap_e (-0.”
• Robust and accurate prediction of residue–residue interactions across protein interfaces using evolutionary information (2014, methodology): “Comparative models were built using RosettaCM ( Song et al., 2013 ) based on alignments to homologous structures generated using HHsearch ( Remmert et al., 2011 ).”
• Large-scale determination of previously unsolved protein structures using evolutionary information (2015, methodology): “The models generated by Rosetta ab initio were refined with an iterative version of the RosettaCM ( Song et al., 2013 ) hybridization protocol used to refine models generated with contact information in CASP10 Figure 16 .”
• Prediction of homo- and hetero-protein complexes by protein docking and template-based modeling: a CASP-CAPRI experiment (2016, neutral): “…non-native contacts in the predicted interface; the root mean square displacement (rmsd) of the backbone atoms of the ligand ( L-rms ), the mis-orientation angle h L and the residual displacement d L of the ligand center of mass, after the receptor in the model and experimental structures were…”
• Comparative Protein Structure Modeling Using MODELLER (2016, methodology): “Three other programs that implement this method are 3D-JIGSAW (Bates et al., 2001), RosettaCM (Song et al., 2013), and SWISS-MODEL (Schwede et al., 2003).”
• InterEvDock2: an expanded server for protein docking using evolutionary and biological information from homology models and multimeric inputs (2018, methodology): “(iii) Once a template and a query-template alignment are available for each protein with no user-provided structure, comparative modeling using a RosettaScripts (41) protocol based on RosettaCM (42) is performed to build a 3D model for (at least part of) the input sequence.”
• SE(3)-Equivariant Energy-based Models for End-to-End Protein Folding (2021, methodology): “Most methods predict protein structures by minimizing statistics energy functions [23, 24] and/or performing comparative modeling from structural templates [37].”