BioForge: a fast and safe macromolecular structure preparation library in Rust.
- 1Materials and Process Simulation Center, Beckman Institute, California Institute of Technology, Pasadena, CA 91125, United States.
Abstract
Summary
BioForge is the first macromolecular structure preparation library for Rust, complementing sequence-analysis libraries such as Rust-Bio with cleaning, geometry-based missing-atom repair, protonation, and solvation as a force-field-agnostic preprocessing stage. Across 22 structures (1439-3 819 839 atoms) on a single CPU core, BioForge scales linearly in wall time (fitted N0.89) versus super-linear N1.17 for PDBFixer and N1.25 for PDB2PQR, completing the 3.8M-atom 9MJN in 5.1 s at 1.1 GB peak (PDBFixer: error, 273 GB); PDB2PQR is limited to fewer than 100 000 atoms. BioForge also compiles to WebAssembly for browser deployment and supports direct embedding into simulation and parameterization pipelines via its application programming interface.Availability and implementation
An MIT-licensed Rust library crate (bio-forge) and a WebAssembly package (bio-forge-wasm) extend the Rust bioinformatics ecosystem with structure preparation, both published on https://crates.io and NPM respectively; the same WebAssembly binary powers a client-side web application at https://bio-forge.app that runs entirely in the browser without a server or data upload. Source code and precompiled command-line binaries are freely available at https://github.com/TKanX/bio-forge; an archival snapshot of the version used for all reported experiments is deposited at Zenodo (https://doi.org/10.5281/zenodo.21627953).
Related
- A novel one-step multiplex PCR method (OS-MPCR) for efficient NGS library preparation in cancer
- HiSpoT: a novel one-tube dual-barcode PCR method for NGS spoligotyping amplicon library preparation in Mycobacterium tuberculosis complex
- DIELECTRIC CONTINUUM MODELING FOR THE DE NOVO DESIGNED CALCIUM CHANNEL CALC6.3: STRUCTURE PREPARATION, MESHGENERATION, AND CALCIUM SELECTIVITY
- LoL-align: sensitive and fast probabilistic protein structure alignment
- SSAlign: fast and sensitive protein structure search at scale
BibTeX
@article{kan2026bioforge,
title = {{BioForge: a fast and safe macromolecular structure preparation library in Rust.}},
author = {Tony Kan and Ted Yu and Soo-Kyung Kim and Gaoyuan Liu and Yixin C Xu and Sejun Kim and William A Goddard},
journal = {Bioinformatics (Oxford, England)},
year = {2026},
month = oct,
doi = {10.1093/bioinformatics/btag616},
url = {https://europepmc.org/article/MED/42841735},
note = {CC BY},
}
RIS
TY - JOUR
TI - BioForge: a fast and safe macromolecular structure preparation library in Rust.
AU - Tony Kan
AU - Ted Yu
AU - Soo-Kyung Kim
AU - Gaoyuan Liu
AU - Yixin C Xu
AU - Sejun Kim
AU - William A Goddard
PY - 2026
DA - 2026/10/01
JO - Bioinformatics (Oxford, England)
DO - 10.1093/bioinformatics/btag616
UR - https://europepmc.org/article/MED/42841735
AB - Summary BioForge is the first macromolecular structure preparation library for Rust, complementing sequence-analysis libraries such as Rust-Bio with cleaning, geometry-based missing-atom repair, protonation, and solvation as a force-field-agnostic preprocessing stage. Across 22 structures (1439-3 819 839 atoms) on a single CPU core, BioForge scales linearly in wall time (fitted N0.89) versus super-linear N1.17 for PDBFixer and N1.25 for PDB2PQR, completing the 3.8M-atom 9MJN in 5.1 s at 1.1 GB peak (PDBFixer: error, 273 GB); PDB2PQR is limited to fewer than 100 000 atoms. BioForge also compiles to WebAssembly for browser deployment and supports direct embedding into simulation and parameterization pipelines via its application programming interface.Availability and implementation An MIT-licensed Rust library crate (bio-forge) and a WebAssembly package (bio-forge-wasm) extend the Rust bioinformatics ecosystem with structure preparation, both published on https://crates.io and NPM respectively; the same WebAssembly binary powers a client-side web application at https://bio-forge.app that runs entirely in the browser without a server or data upload. Source code and precompiled command-line binaries are freely available at https://github.com/TKanX/bio-forge; an archival snapshot of the version used for all reported experiments is deposited at Zenodo (https://doi.org/10.5281/zenodo.21627953).
SN - 1367-4803
SN - 1367-4811
ER -