Source-linked AI summary
Serverification of Molecular Modeling Applications: the Rosetta Online Server that Includes Everyone (ROSIE)
Sergey Lyskov, Fang-Chieh Chou, Shane Ó Conchúir, Bryan S. Der, Kevin Drew, Daisuke Kuroda, Jianqing Xu, Brian D. Weitzner, P. Douglas Renfrew, Parin Sripakdeevong, Benjamin Borgo, James J. Havranek, Brian Kuhlman, Tanja Kortemme, Richard Bonneau, Jeffrey J. Gray, Rhiju Das
TL;DR
Rosetta’s difficult setup, computational demands, and limited server availability restricted access beyond developers and collaborators. ROSIE unified Rosetta protocols into a maintainable web framework, enabling rapid deployment of diverse applications and broader access to Rosetta modeling.
Problem
Rosetta applications lacked broad server access, while their difficult use and computational demands limited availability beyond developers and collaborators.
Method
ROSIE combines a common interface, developer API, shared cluster backend, and centralized administration to serverify Rosetta applications.
Results
Serverification took four weeks for RNA de novo modeling and three weeks for ERRASER, followed by seven additional diverse Rosetta functionalities.
Takeaways & Limitations
ROSIE provides a general, speedy framework for deploying diverse Rosetta applications through shared infrastructure.
Takeaways & Limitations
Additional applications and users may eventually strain ROSIE’s computational resources.
Abstract
from arXiv · showhide
The Rosetta molecular modeling software package provides experimentally tested and rapidly evolving tools for the 3D structure prediction and high-resolution design of proteins, nucleic acids, and a growing number of non-natural polymers. Despite its free availability to academic users and improving documentation, use of Rosetta has largely remained confined to developers and their immediate collaborators due to the code's difficulty of use, the requirement for large computational resources, and the unavailability of servers for most of the Rosetta applications. Here, we present a unified web framework for Rosetta applications called ROSIE (Rosetta Online Server that Includes Everyone). ROSIE provides (a) a common user interface for Rosetta protocols, (b) a stable application programming interface for developers to add additional protocols, (c) a flexible back-end to allow leveraging of computer cluster resources shared by RosettaCommons member institutions, and (d) centralized administration by the RosettaCommons to ensure continuous maintenance. This paper describes the ROSIE server infrastructure, a step-by-step 'serverification' protocol for use by Rosetta developers, and the deployment of the first nine ROSIE applications by six separate developer teams: Docking, RNA de novo, ERRASER, Antibody, Sequence Tolerance, Supercharge, Beta peptide design, NCBB design, and VIP redesign. As illustrated by the number and diversity of these applications, ROSIE offers a general and speedy paradigm for serverification of Rosetta applications that incurs negligible cost to developers and lowers barriers to Rosetta use for the broader biological community. ROSIE is available at http://rosie.rosettacommons.org.
RNA de novo modeling and ERRASER as ‘external’ test cases
RNA de novo modeling and ERRASER served as external Stanford test cases demonstrating that ROSIE could rapidly serverify applications developed outside its originating Johns Hopkins site. RNA de novo required four weeks, while ERRASER was developed nearly independently and integrated after three weeks.
- External test cases: RNA de novo modeling and ERRASER, both developed by Stanford’s Rosetta group, provided the first external test cases for ROSIE.This tested ROSIE’s ability to support laboratories at different universities and with different modeling focuses.
- RNA de novo modeling: 4 weeks was the total development time for serverifying RNA de novo modeling, including high-resolution refinement, which had not previously been available via a server.The Stanford team specified application inputs, outputs, and Rosetta command lines for modeling, clustering, and simple testing.
- ERRASER: ERRASER was the first ROSIE server implemented and published concomitantly with a new Rosetta application and the first to accept e…It provided a test case for more rapid, independent implementation by application developers.
- ERRASER: ERRASER optimizes local geometries of RNA crystallographic structures using electron-density-map constraints and the Rosetta scoring function.The application developers used the previously implemented docking and RNA de novo applications as templates and developed the server nearly independently from ROSIE administrators.
- ERRASER: 3 weeks was the overall development time for ERRASER: two weeks for developers’ initial deployment and one week for central integration.A virtual-machine image enabled local testing before public Web deployment.
- Rapid creation of additional server functionalities: Seven additional Rosetta functionalities were serverified in parallel with or after deployment of the first three applications.These applications are introduced as a diverse set of additional ROSIE functionalities.
β-peptides
β-peptides have an extra backbone carbon atom that creates an additional dihedral angle and increases the distance between adjacent side chains; structured non-biological-backbone polymers are often called foldamers.
- β-peptides: β-peptides contain an additional backbone carbon atom, producing an extra dihedral angle and greater spacing between adjacent side chains.Structured β-peptides are non-biological-backbone polymers often referred to as foldamers.
Adding functionality: NMR chemical shifts in RNA de novo
ROSIE incorporated NMR chemical shift-guided modeling directly into its RNA de novo server rather than creating a separate server. This extends Rosetta RNA de novo modeling with nonexchangeable 1H chemical shift data, which has produced high-resolution RNA structures.
- Motivation: NMR chemical shifts provide important structural information for functional macromolecules and are widely used in structural analysis and model refinement.Backbone chemical shifts support protein secondary-structure and backbone-torsion determination, as well as three-dimensional model refinement.
- NMR chemical shift-guided modeling: Integrating nonexchangeable 1H chemical shift data with Rosetta RNA de novo modeling has produced high-resolution RNA structures.
- NMR chemical shift-guided modeling: ROSIE added NMR chemical shift-guided modeling to the RNA de novo server instead of creating a separate server.The passage identifies this as a modeling mode within the existing RNA de novo server.
RosettaVIP
RosettaVIP is an iterative protocol that improves protein-core packing by proposing mutations, rescuing designs with packing defects and potentially improving wild-type protein stability. It operates on PDB structural models for a user-selected maximum number of mutation iterations.
- Protocol outcomes: RosettaVIP rescues protein designs with packing defects and suggests mutations predicted to improve wild-type protein stability.Packing defects are associated with reduced stability and conformational specificity, motivating computational core-packing correction.
- Iterative protocol: Each iteration proposes a mutation predicted to improve protein-core packing or terminates when no such mutation can be identified.The relaxed model containing the new mutation becomes the input for the next iteration.
- Inputs and parameters: The protocol takes a structural model in PDB format as input and lets users select the maximum number of iterations, corresponding to the maximum number of mutations tried.Each iteration outputs a new mutation and a relaxed structural model for continued processing.
I. Install a local ROSIE test server
Installing a local ROSIE test server requires configuring the VM for localhost, launching separate back-end and front-end processes, and accessing the service through a browser. Adding an application involves implementing protocol, server, interface, documentation, testing, database, and integration changes.
- I. Install a local ROSIE test server: Download the ROSIE VM, open it with VirtualBox, and optionally update the ROSIE and Rosetta directories before rebuilding Rosetta.
- I. Install a local ROSIE test server: Comment out host = 192.168.0.64 and enable host = 127.0.0.1 in development.ini.
- I. Install a local ROSIE test server: Run the ROSIE daemon and server from separate terminals, then open localhost:8080 and log in as admin with password managepass.
- II. Add a new application XXX: Create an application under rosie.back/protocols/XXX with submit.py and analyze.py, configure machine-dependent files, and add its front-end controller and root registration.
- II. Add a new application XXX: Add application pages, main-page links, optional icons, documentation, bootstrap registration, and an updated protocol schema, then test, document, validate, commit, and notify administrators.