Performance Evaluation of the all_ratio Algorithm for Pathogenic Microorganism and Antibiotic Resistance Gene Detection Using Nanopore Sequencing

Objective To verify the detection performance of the four-dimensional weighted all_ratio comprehensive scoring algorithm for simultaneous identification of bacteria, antibiotic resistance genes (ARGs) and pathogenic fungi in nanopore metagenomic sequencing data, and to evaluate the cross-dataset generalizability and clinical application value of the algorithm. Methods Nanopore sequencing data of 41 lower respiratory tract infection samples and 8 fungal infection samples retrieved from public databases were included in this study. Conventional microbial culture method and classic nanopore analytical pipeline were used as controls. The all_ratio algorithm integrates four core parameters: number of matches, abundance, sequence similarity and alignment length. A comprehensive score was calculated via weighted multiplication and normalization to achieve simultaneous detection of pathogenic bacteria, ARGs and pathogenic fungi. Samples were graded according to the consistency between detection results and control methods, and the detection accuracy and improvement in reducing missed diagnosis were statistically analyzed. Results Among the 41 lower respiratory tract infection samples, 22 were classified as Same_2 group, 8 as Same_1 group, 5 as Good group, and 6 as Different group. Bacterial identification with high all_ratio scores was highly consistent with control methods. The algorithm effectively detected low-abundance pathogens missed by traditional methods, and accurately identified core resistance genes such as SdeY , TEM-12 , HmrM and mexA , achieving precise matching between resistance genes and their host bacteria. Validation on 8 additional fungal samples showed that the all_ratio algorithm could stably identify various pathogenic fungi with good signal discrimination, making it suitable for fungal species identification analysis. Conclusion The all_ratio comprehensive scoring algorithm has excellent universality and stability. It enables one-stop and accurate simultaneous detection of pathogenic bacteria, antibiotic resistance genes and pathogenic fungi, and effectively compensates for the shortcomings of traditional detection methods. This algorithm is suitable for broad-spectrum, rapid and accurate screening of clinical pathogenic microorganisms.