Princeton_TIGRESS: proTeIn Geometry REfinement using Simulations and Support vector machines

This webserver is an automated refinement method that intends to refine an input protein structure by addressing both sampling and selection. First, the method generates many local minima near the starting structure using torsion-angle dynamics in Cyana. Each of these local minima are next relaxed using Rosetta Fast Relax with constraints on the coordinates to remain near the start in order to repack the side-chains and to make movements that will enhance the number of hyrogen bonds. This ensemble of structures is next filtered to remove structures deviating too far away from the input and the remaining structures are passed through a support vector machines model which classifies which structures have moved in the more native-like direction. The models selected by the SVM are then scored using the dDFIRE energy function and the lowest energy structure is selected. This structure is then refined via all atom molecular dynamics simulatons in CHARMM using the FACTS implicit solvent model. The final refined structure is sent to the user which has subtantially lower clashes and aims to improve GDT_TS based on benchmarking on all CASP7,8,9, and 10 refinement targets and typical use runs.

Due to limited resources, the TIGRESS web service will be unavailable during CASP11. Please contact us if you have a specific urgent refinement to perform

How to Cite this Method

If using this webserver, please cite the associated publication, which describes the details of our method.

Khoury, G.A., Tamamis, P., Pinnaduwage, N., Smadbeck, J., Kieslich, C.A., and Floudas, C. A. Princeton_TIGRESS: ProTeIn Geometry REfinement using Simulations and Support vector machines. Proteins: Structure, Function, Bioinformatics DOI: 10.1002/prot.24459


CAF acknowledges support from The National Institutes of Health grant number R01GM052032, and the National Science Foundation. GAK is grateful for support by a National Science Foundation Graduate Research Fellowship under grant number DGE-1148900. We thank Eric First for help making this webserver.


Although every consideration has been made during the design, generation, and posting of this data, we provide all data 'as-is'. We make no warranties regarding its correctness, and disclaim liability for damages resulting from its use. This website and associated data and services are for research only, not for clinical or commercial use. We provide this information as a non-profit service to the academic community. The user utilizes the server and its predictions completely at their own risk. Princeton University reserves the right to change the website or discontinue its service at any time.