Composite Genome Quality Index for Pathogenic Bacterial Genomes

High-quality bacterial genomes are essential for robust comparative genomics, reliable taxonomic assignment, and accurate pathogen and antimicrobial resistance (AMR) surveillance. Yet, public repositories still contain highly heterogeneous assemblies, and genome quality is often judged using single metrics in isolation. Here we develop an integrative Genome Quality Index (GQI) that combines four complementary metrics—including BUSCO single-copy completeness, contig number, N50, and unmapped read percentage—into a composite, interpretable score. We re-assembled and evaluated 474 pathogenic bacterial genomes submitted from South Korea using a standardized Illumina-based pipeline and validated the framework on an independent Enterobacteriaceae dataset (n = 5781). Species-level analyses and unsupervised clustering revealed pronounced variation in genome quality (one-way ANOVA, p < 1.05 × 10−33), with Cronobacter sakazakii and Listeria monocytogenes showing consistently high GQI scores, whereas Mycobacterium tuberculosis exhibited broad variability, including clear low-quality outliers. After log-transforming skewed variables, contig count and N50 remained strongly negatively correlated (r = −0.83), while BUSCO completeness showed moderate positive association with N50 and negative association with unmapped reads. GQI scores spanned 0.23–0.96, with most genomes clustering between 0.70 and 0.85. A Random Forest classifier trained on the four raw metrics predicted GQI-based quality tiers (low, medium, high) with 97% accuracy. From the top-decile genomes, we derived empirical thresholds like BUSCO ≥ 98.6%, contigs ≤ 30, N50 ≥ 1 Mb, and unmapped reads ≤ 0.82% that refine existing recommendations and provide actionable curation criteria. Our framework complements tools such as CheckM, gVolante, and Hybracter by offering a platform-agnostic composite scoring system that can be integrated into submission workflows and surveillance pipelines to systematically flag low-quality genomes and improve the reliability of microbial genomics.