基于高通量测序的生物制品微生物污染检测工具Kraken2与Sylph 的性能比较与分析

Performance Comparison and Analysis of Kraken2 and Sylph Tools for Microbial Contamination Detection of Biological Products Based on Next-Generation Sequencing

关键词:  高通量测序  外源因子污染  生物制品  检测限

Key words:  next-generation sequencing  adventitious agents  biological products  detection limit

摘要:

目的 评估高通量测序技术(Next-Generation Sequencing, NGS)对生物制品污染(病毒、细菌及真菌)检测中的应用潜力，系统比较生物信息学工具Kraken2与Sylph的检测性能差异。方法 以市售人血白蛋白为基质，梯度浓度加标乙肝病毒(Hepatitis B virus, HBV)、副铜绿假单胞菌(Pseudomonas paraeruginosa)、斯皮宰曾氏芽孢杆菌(Bacillus spizizenii)及巴西曲霉(Aspergillus brasiliensis)，构建不同浓度污染样本，进行高通量测序，分别采用Kraken2和Sylph进行物种分类和丰度分析，并系统评估检测限、灵敏度和特异性。结果 在0.5M的数据量下，Kraken2工具对于HBV的检测限为1×105 copies·mL-1，对P. paraeruginosa、B. spizizenii和A. brasiliensis的检测限分别为1 CFU·mL-1、100 CFU·mL-1和100 CFU·mL-1。Sylph工具对HBV的检测限为1×107 copies·mL-1；对P. paraeruginosa、B. spizizenii和A. brasiliensis的检测限分别为1×104 CFU·mL-1、1×104 CFU·mL-1和1×105 CFU·mL-1。阴性对照中检测到非特异性污染。在多种数据量下，Kraken2工具对于HBV的检测限为1×105 copies·mL-1，对P. paraeruginosa、B. spizizenii和A. brasiliensis的检测限分别为1 CFU·mL-1、10 CFU·mL-1和10 CFU·mL-1。Sylph工具对HBV的检测限为1×105 copies·mL-1；对P. paraeruginosa、B. spizizenii和A. brasiliensis的检测限分别为100CFU·mL-1、1×103 CFU·mL-1和1×104 CFU·mL-1。结论 Kraken2灵敏度较Sylph更高，但同时其检测的结果中也存在更多的非特异性污染。总体来说，NGS技术可同步检测多类别生物制品污染，通过优化测序流程及结合多工具交叉验证策略，显著提升检测可靠性，NGS技术有望成为生物制品质控的有效手段。

Abstract:

OBJECTIVE The application potential of Next-Generation Sequencing (NGS) in the detection of biological products contamination(viruses, bacteria and fungi) was evaluated. The detection performance differences between Kraken2 and Sylph were systematically compared. METHODS Hepatitis B virus (HBV), Pseudomonas paraeruginosa, Bacillus spizizenii and Aspergillus brasiliensis were spiked at different concentrations with commercially available human albumin as the matrix, and NGS was performed. Kraken2 and Sylph were used for species classification and abundance analysis, respectively. The detection limit, sensitivity and specificity were systematically evaluated. RESULTS The detection limit of Kraken2 for HBV was 1×105 copies·mL-1, and the detection limits for P.paraeruginosa, B.spizizenii and A.brasiliensis were 1 CFU·mL-1, 100 CFU·mL-1 and 100 CFU·mL-1, respectively. The detection limit of HBV by Sylph was 1×107 copies·mL-1. The detection limits for P.paraeruginosa, B.spizizenii and A.brasiliensis were 1×104 CFU·mL-1, 1×104 CFU·mL-1 and 1×105 CFU·mL-1, respectively. Non-specific contamination was detected in the negative control. In a variety of sequencing data volume, The detection limit of Kraken2 for HBV was 1×105 copies·mL-1, and the detection limits for P.paraeruginosa, B.spizizenii and A.brasiliensis were 1 CFU·mL-1, 10 CFU·mL-1 and 10 CFU·mL-1, respectively. The detection limit of HBV by Sylph was 1×105 copies·mL-1. The detection limits for P.paraeruginosa, B.spizizenii and A.brasiliensis were 100 CFU·mL-1, 1×103 CFU·mL-1 and 1×104 CFU·mL-1, respectively. CONCLUSION The sensitivity of Kraken2 was higher than that of Sylph, but there was more non-specific contamination in the detection results. In general, NGS technology can simultaneously detect multi-category adventitious agents contamination. By optimizing the sequencing process and combining the multi-tool cross-validation strategy, the detection reliability is significantly improved. NGS technology is expected to become an effective means of quality control of biological products.