To evaluate and compare the experimental parameters and resultant differences of two magnetic bead-based nucleic acid extraction reagents in pathogen-targeted next-generation sequencing (ptNGS) for the detection of bronchoalveolar lavage fluid (BF) and peripheral blood (PB) samples, and to identify the more effective nucleic acid extraction reagent and assess its applicability in ptNGS detection, in order to offer guidance for the selection of appropriate nucleic acid extraction reagents tailored to various sample types.

Two representative magnetic bead-based nucleic acid extraction kits, designated as KitA and KitB were employed to conduct ptNGS detection on 12 BF samples and 8 PB samples, all of which had established culture or real-time fluorescence quantitative PCR (qPCR) results. An equal volume was utilized for each sample. The methodology encompassed several key stages, including nucleic acid extraction, library construction, sequencing and subsequent result analysis. The experimental parameters, including total nucleic acid extraction mass, absorbance (A) _260/280, library concentration and sample sequencing data quality index Q30; and pathogen detection copy numbers between KitA and KitB during the detection process of samples (BF and PB) were systematically compared by Wilcoxon signed-rank tests, and the magnetic bead extraction reagent which was more suitable for ptNGS detection of BF and PB samples were clarified. The suitability of the extraction kit was further evaluated by the accuracy, precision, detection limit and conventional anti-interference ability of ptNGS detection.

The total amount of nucleic acid extracted in BF by KitA was significantly higher than that of KitB (W=-66, P=0.001), but without significant difference in PB (W=19, P=0.063). However, the absorbance A_260/280 of KitB nucleic acid in BF and PB samples were significantly better than those of KitA (BF: W=54, P=0.014; PB: W=21, P=0.031). But the library concentration and fragment size of the libraries constructed by KitA and KitB nucleic acid were not significantly different; sequencing data quality metrics including Q30 scores, primer dimer percentage and percentage of Real aligne were not significantly different between the two extraction methods (all P > 0.05). The number and species of KitA and KitB detected in the pathogen were consistent. The pathogen copy number detected by KitA in BF was significantly higher than that of KitB (W=-301, P < 0.001), however, there was no statistically significant difference in pathogen copy number in PB between the two extraction methods (W=-3, P=0.844). Further, a more suitable nucleic acid extraction kit KitA was selected to evaluate its suitability for ptNGS detection in BF and PB. The positive concordance rate of ptNGS detected by KitA was 97.14%, the missed detection rate was 2.86%, and the overall concordance rate was 95%. Both intra-batch and inter-batch precision exhibited coefficients of variation below 10%. The limit of detection (LOD) was established at 100 copies/ml for bacterial and fungal targets, and 1 000 copies/ml for viral targets. Notably, non-target nucleic acids (including human-derived DNA/RNA) in the samples did not interfere with the qualitative results of ptNGS analysis.

For ptNGS detection in BF samples, Kit A demonstrated overall superior performance compared with Kit B across all evaluated parameters. In PB samples, both kits were consistent in overall experimental parameters, including total nucleic acid yield, library concentration and sequencing data quality indicators Q30. KitA meets the requirements of clinical testing in terms of accuracy, precision, detection limit and routine anti-interference capability in the performance evaluation of ptNGS detection