【文献解读】Nature Microbiology | 使用MinION快速分析早产儿肠道菌群谱并鉴定抗生素抗性致病菌

                简介                 

标题：使用MinION快速分析早产儿肠道菌群谱并鉴定抗生素抗性致病菌

杂志：Nature Microbiology

影响因子：15.54

发表时间：2019年12月16日

解读：很跩的土豆

编辑：很跩的土豆

导读：MinION测序平台使得DNA测序实时分析成为可能，这在田野和临床研究中具有很大吸引力。本文作者将MinION测序平台与NanoOK RT软件相结合，对模拟菌群和早产儿（包括健康和疾病的早产儿）的粪便样本进行宏基因组测序和分析。通过使用Nanopore数据，作者可靠的鉴定了20个种的模拟菌群，检测了不同时间的肠道菌群的多样性，同时分析了菌群的抗生素抗性。此外，基于NanoOK RT软件（一种快速分析测序结果的开源软件），作者快速实时的分析了健康和疾病状态下新生儿的肠道微生物组。在本文中，作者鉴定了多种病原菌，包括Klebsiella pneumoniae和Enterobacter cloacae。同时，在1小时内鉴定出了这些病原微生物的抗生素抗性基因。检测结果进一步使用病原体分离、WGS测序和抗生素易感性检测，以及已知抗性基因组成的模拟菌群和临床样本进行确认。这些结果显示，MinION（包括更经济实惠的Flongle芯片）与NanoOK RT同时使用可以在<5h的时间内处理宏基因组样本。在以后的研究中，通过发展这些工具和方法，可以在临床实践中为患者提供个性化的抗生素治疗方案。

                正文                 

1.    引言

NGS测序技术为环境和临床微生物组的分析带来了革命性改变。不依赖培养的、高敏感性、大数据量的宏基因组结合强大的生物信息学工具使得研究者通过菌群结构（包括病原体增殖）将病人与健康人快速区分。宏基因组数据可以允许我们探究微生物的功能特点，比如抗生素抗性基因，这在预防抗生素抵抗中尤其重要。优化宏基因组方法学和生物信息学工具有助于通过菌群分析识别高风险患者个体，并有助于在临床中鉴定病原体、制定治疗方案。

与NGS平台不同，Nanopore MinION等测序平台设备便宜、可产生长读长序列；此外，实时产生测序数据为使用者提供了快速筛选的平台。然而，这种实时检测分析的功能化、碱基误判等问题依然需要从方法学和生物信息学流程中进行改进，以最终满足临床应用的需求。

尽管在宏基因组分析和诊断过程中还具有诸多挑战，但是MinION已经在临床研究中探究低复杂性样本上成功应用，比如感染监测、分离细菌和低生物量样本的鉴定等。因序列得率和准确性较低，基于MinION宏基因组的临床诊断依然困难重重，但已经足够。到目前为止，还没有相关研究探索MinION技术在临床患者肠道宏基因组样本中的应用。在临床应用中，种水平检测、菌群中细菌种的丰度以及抗生素抗性基因（AMR）的鉴定尤其重要。

2.结果部分

2.1 MinION测序对模拟菌群的分类准确性

图1. 使用Ilumina和MinION对模拟菌群（HM-277D）的测序结果对比. a，模拟菌群中20个菌种的预期DNA比例和MinION测序获得的比例的相关性（左，log-transformed Pearson’s r = 0.94），与Illumina测序获得的DNA比例的相关性（中，log-transformed Pearson’s r =0.97），MinION与Illumina测序结果的相关性（右，log-transformed Pearson’s r = 0.98）。b，使用MEGAN展示Illumina（绿色）和ONT MinION（棕色）测序或等的模拟菌群菌种比例。

Fig. 1 | Sequencing of microbial mock community (HM-277D) with Illumina and MinION sequencing. a, Correlation plot comparing the expected proportions of DNA from each of the 20 species in the mock community with the proportions of sequence obtained by MinION (left, log-transformed Pearson’s r = 0.94) and Illuminaplatforms (log-transformed Pearson’s r = 0.97). The right-hand plot shows the correlation between MinION and Illumina (log-transformed Pearson’s r = 0.98). The grey region either side of the fit line represents the 95% CIs. b, MEGAN37 taxonomy tree representing taxa assigned from the mock community as sequenced by Illumina MiSeq (green) and ONTMinION (brown). The height of the bars indicates the relative number of reads assigned, with normalized counts given next to the taxa name. Asterisks represent species assigned by MEGAN but not specified as members of the mock community. Numbers assigned to taxa correspond to Illumina MiSeq (left) and ONTMinION (right) normalized counts, respectively.

2.2 使用MinION检测早产儿肠道菌群变化特点

图2，使用MinION和Illumina测序纵向研究早产儿P10的肠道菌群特点。a, 早产儿P10粪便样本采集（三个时间段P10N, P10R和P10V）、抗生素和益生菌使用时间及相关临床观察的时间轴线图。点表示时间轴中的某个时间点，其中蓝点示益生菌使用的时间点，红点示抗生素使用时间点，黑色方框示样本采集时间点；字母“H”表示早产儿转诊到其他医院。b, Illumina（x轴）和MinION（y轴）测序获得早产儿P10N样本标准化后菌群组成的相关性（log-transformed Pearson’s r = 0.95，taxa n =92）。P10R和P10V样本中两种测序结果的相关性分别为0.90和0.94。c, P10N, P10R, P10V三个时间点的物种组成。上面三个图为MinION测序结果，下面三个图为Illumina Hiseq测序结果。图例展示了12个丰度最高的物种丰度。

Fig. 2 | Longitudinalstudy on preterm infant P10 using MinION and Illumina sequencing. a, Timeline diagram of preterm P10 indicating the times of faecal sample collection (P10N, P10R and P10V), duration of antibiotic and probiotic treatment and relevant clinical observations. The diagram is divided into weeks and the dots represent days within the timescale. The blue dots represent the days of probiotic treatment; the red dots represent the days of antibiotic treatment; the black squares represent the time points for sample collection; and the letter H represents the transfer of the preterm infant to another hospital. b, Correlation plot of normalized species abundance in taxonomic assignments of Illumina (x axis) and MinION (y axis) data for sample P10N, with log-transformed Pearson’s r = 0.95(taxa n = 92). The values for P10R and P10V were 0.90 and 0.94, respectively. The grey region either side of the fit line represents the 95% CIs. c, Taxonomic profiles at time pointsP10N, P10R and P10V, as assigned by MEGAN. The top row corresponds to the results obtained using MinION sequencing; the bottom row displays the results obtained using Illumina HiSeq. The legend comprises the 12 most abundant taxa obtained. Further information on all taxa and corresponding read counts can be found in Supplementary Table 3.

2.3 生物信息学工具根据MinION特定的物种组成快速表征肠道细菌及抗生素抵抗的细菌谱

图3，使用MinION技术对临床诊断为疑似坏死性肠炎（NEC，P49和P205）的早产儿进行快速诊断. a,b,P49（a）和P205（b）患者测序开始后1和6小时的细菌组成谱。c, P49患者在1和6小时的菌群组成相关性（taxa n = 35, log-transformedPearson’s r = 0.97）。d, P205患者在1和6小时的菌群组成相关性（taxa n = 35, log-transformedPearson’s r = 0.97）。e,f, CARD（菌群耐药分析数据库）预测P49（e）和P205（f）样本中耐药基因数量热图。g,h，使用NanoOK RT 的“Walkout”选项获得的新生物P49（g）和P205（h）的分析结果。译者注：“Walkout”功能用于分析AMR基因的细菌来源。

Fig. 3 | Rapid diagnostic using MinION technology for preterm infants clinically diagnosed with suspected NEC (P49 and P205). a,b, Taxonomic profiles comparing the results obtained at 1h and 6 h for P49 (a) and P205 (b) after sequencing started. The pie chart legends comprise the eight most abundant taxa. Detailed counts can be found in Supplementary Table 3. c,d, Correlation plots representing normalized assigned reads at 1 h and 6 h for P49 (taxa n = 35, log-transformed Pearson’s r = 0.97) (c) and P205 (taxa n = 120, log-transformed Pearson’s r = 0.98) (d). The grey region either side of the fit line represents the 95% CIs. e,f, Heat maps displaying the number of CARDhits detected among the most common groups of antibiotic resistance genes found in preterm P49 (e) and P205 (f). Further information on all the AMR genes classified can be found in Supplementary Table 4. g,h, Walkout results for preterm infants P49 (g) and P205 (h) at 6 h, as reported by the NanoOK RT’s walkout option. Results shown are for independent bacterial hits(defined as ≥ 50 bp away from the AMRsequence), at 6 h of sequencing.

图4，对临床诊断为疑似NEC的早产儿P8进行快速诊断。a, 时间轴：采样、DNA提取、质控（红，2h15min）、文库制备和DNA修复（黄，1h 45min）、本地basecalling和NanoOK RT分析（绿）。病原体检测（K. pneumoniae）和K. pneumoniae特异的AMR基因在4h 38min首先被检测（13000条reads），时间轴左侧示根据临床症状和指南开具抗生素治疗的处方。b, MinION测序1和6h的物种组成谱。c, P8患者测序结果的Walkout分析。d, 1和6h测序结果的相关性。E，耐药基因数量热图。

Fig. 4 | Rapid diagnostic of preterm P8 clinically diagnosed with suspected NEC. a, Time frame diagram showing: sample collection, DNA extraction and quality control (red, 2 h 15 min); library preparation incorporating bead clean-up and DNA repair (yellow, 1 h 45 min); data analysis using local base calling and NanoOK RT (green). Pathogen detection (K.pneumoniae) and K. pneumoniae-specific AMR genes were first detected at 4 h and 38 min (13,000 reads analysed). The left side of the panel indicates the clinical symptoms and general guidelines for antibiotic prescription. b, Taxonomic profiles obtained using MinION at 1 and 6 h since sequencing started. The pie chart legend comprises the eight most abundant taxa classified. Further taxa can be found in Supplementary Table 3. c, Walkout study of P8 reported by the NanoOK RT software showing taxa containing AMR genes. The results shown are for independent bacterial hits (defined as ≥ 50 bp away from the AMRsequence) at 6 h of sequencing. d, Correlation plot of species-level normalized assigned read counts at 1 and 6 h, with log-transformed Pearson’s r = 0.97 (taxa n = 133). The grey region either side of the fitline represents the 95% CIs. e, Heat map displaying the number of CARD hits detected among the most common groups of antibiotic resistance genes found in pretermP8. Further information on all the AMR genes obtained can be found in Supplementary Table 4.

2.4 使用抗生素抗性模拟样本对MinION和NanoOK RT数据定标和验证

图5，使用抗生素抗性模拟样本对MinION和NanoOK RT数据定标和验证。a, 将由8中细菌和一个从P8患者中分离的K. pneumoniae组成的模拟菌群用MinION测序和用NanoOK RT分析。数据结果与Illumina和MinION测序和组装的P8患者K.pneumoniae进行对比。将AMR基因根据序列相似度进行分组。蓝色表示存在，灰色表示不存在。b, 将健康早产儿P103M（未掺入）和掺入K. pneumoniae的样本的MinION AMR谱进行对比。使用两种不同浓度的P8患者的K.pneumoniae掺入P103M样本，4%（P103M 4% spike）和40%（P103M40% spike）。使用Illumina和MinION测序和组装的P8患者K.pneumoniae进行结果对比。将AMR基因根据序列相似度进行分组。c, P103M 4% spike 和P103M 40% spike样本的菌群组成，仅展示丰度>=1%的物种。

Fig. 5 | Benchmarking and validation of MinION and NanoOK RT using mock resistome samples. a, A known mock community comprising eight bacteria and an isolate of K. pneumoniae from P8 was sequenced using MinION and analysed with NanoOK RT. Findings were compared to the results obtained from the P8 K. pneumoniae isolate sequenced and assembled by Illumina and MinION. The AMR genes were grouped according to sequence similarity. Blue indicates presence, grey indicates absence. Details of specific genes detected can be found in Supplementary Table 9. b, MinION AMR profiles obtained from a faecalsample from a healthy preterm infant (P103M no spike) were compared to the K. pneumoniae spiked forms. Two different concentrations of P8 K. pneumoniae isolate were inoculated on sample P103M; 4%(P103M 4% spike) and 40% (P103M 40% spike). Findings were compared to the results obtained from the P8 K. pneumoniae isolate sequenced and assembled from Illumina and MinION. AMR genes were grouped according to sequence similarity. c, Percentages of reads assigned to P103M spiked with 40% P8 K. pneumoniae isolate (P103M 40%spike) and 4% P8 K.pneumoniae isolate(P103M 4% spike); only taxa representing ≥ 1% are shown.

3. 总结

      使用MinION与NanoOK RT可以组建一个快速分析肠道细菌菌种的平台，包括分析潜在的病原体和相应的AMR基因谱。这种方法的准确性也被Illumina宏基因组测序数据、患者细菌分离物、已知耐药基因组成的模拟菌群等验证。这些方法的联合使用，对于临床抗生素耐药分析和抗生素治疗的管理都具有重要意义。

                参考                 

Leggett,R.M., Alcon-Giner, C., Heavens, D. et al. Rapid MinION profiling of pretermmicrobiota and antimicrobial-resistant pathogens. Nat Microbiol 5, 430–442(2020). https://doi.org/10.1038/s41564-019-0626-z

                索引                 

【往期文献】

【二三代比较】

【微生物组】

【去宿主】

【转录组】

【病毒】

GUT: 新冠病毒病毒活性与COVID-19患者肠道菌群的关系

【耐药】

                后记                 

随着测序技术的不断发展，科学研究进入了数据井喷的时代。然而，测序样本的处理流程、测序数据的分析流程甚至是数据分析过程中的数据库搭建问题，都给测序技术的普及化设置了壁垒，严重阻碍了该项技术向广大科研工作者中推广。此外，基于长读长的三代测序技术的发展更是引入了一套完全有别于二代测序数据处理的分析流程，为了让更多学者认识三代测序、在科学研究中用好三代测序，本公众号应运而生。期待与您一起学习、成长。

^_^ 边学习，边分享，每天进步一点点 ^_^