Lifu Song, Mei Wang, Xiaoping Liao, Junli Wu, Zhenkun Shi, Ruoyu Wang, Haoran Li, Lulu Liu, Botao He, Xiaomeng Ni, Qinggang Li, Jinshan Li, Hongwu Ma, Ping Zheng, Jibin Sun, Yanhe Ma
Although several tools exist for prokaryotic genome correction, most depend on external read mappers that impose computational overhead and usability barriers, leaving a need for fast, accurate, standalone solutions. We present ProGenFixer, a mapping-free tool that identifies and corrects errors in prokaryotic genomes using next-generation sequencing data. ProGenFixer compares k-mer profiles between an input genome and sequencing reads to pinpoint discrepancies, then applies a local assembly-based algorithm with a conservative correction strategy suited to refining high-quality reference genomes. In benchmarking, ProGenFixer ran over 5× faster than the existing tools tested while maintaining higher accuracy. Against slower but highly accurate tools, it showed comparable accuracy while running >17× faster, with better performance on long indels (>10 bp). ProGenFixer is designed primarily for conservative updating of already curated reference sequences rather than aggressive polishing of draft assemblies; accordingly, it does not automatically replace a reference base when substantial read support exists for both the reference and alternate alleles. Comparisons using 3 real bacterial resequencing datasets showed that most tool-specific disagreements occurred at mixed-support or low-depth sites, emphasizing that correction aggressiveness and accuracy are not equivalent in this application. ProGenFixer is implemented in C and runs as a standalone program with no external dependencies. The software is freely available at https://github.com/Scilence2022/ProGenFixer. A companion web platform is available at https://progenfixer.biodesign.ac.cn.