Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Metagenomic prediction of antimicrobial resistance in critically ill patients with lower respiratory tract infections

Metagenomic prediction of antimicrobial resistance in critically ill patients with lower... Background: Antimicrobial resistance (AMR) is rising at an alarming rate and complicating the management of infec- tious diseases including lower respiratory tract infections (LRTI). Metagenomic next-generation sequencing (mNGS) is a recently established method for culture-independent LRTI diagnosis, but its utility for predicting AMR has remained unclear. We aimed to assess the performance of mNGS for AMR prediction in bacterial LRTI and demonstrate proof of concept for epidemiological AMR surveillance and rapid AMR gene detection using Cas9 enrichment and nanopore sequencing. Methods: We studied 88 patients with acute respiratory failure between 07/2013 and 9/2018, enrolled through a previous observational study of LRTI. Inclusion criteria were age ≥ 18, need for mechanical ventilation, and respira- tory specimen collection within 72 h of intubation. Exclusion criteria were decline of study participation, unclear LRTI status, or no matched RNA and DNA mNGS data from a respiratory specimen. Patients with LRTI were identi- fied by clinical adjudication. mNGS was performed on lower respiratory tract specimens. The primary outcome was mNGS performance for predicting phenotypic antimicrobial susceptibility and was assessed in patients with LRTI from culture-confirmed bacterial pathogens with clinical antimicrobial susceptibility testing (n = 27 patients, n = 32 pathogens). Secondary outcomes included the association between hospital exposure and AMR gene burden in the respiratory microbiome (n = 88 patients), and AMR gene detection using Cas9 targeted enrichment and nanopore sequencing (n = 10 patients). Results: Compared to clinical antimicrobial susceptibility testing, the performance of respiratory mNGS for predict- ing AMR varied by pathogen, antimicrobial, and nucleic acid type sequenced. For gram-positive bacteria, a combi- nation of RNA + DNA mNGS achieved a sensitivity of 70% (95% confidence interval (CI) 47–87%) and specificity of 95% (CI 85–99%). For gram-negative bacteria, sensitivity was 100% (CI 87–100%) and specificity 64% (CI 48–78%). Patients with hospital-onset LRTI had a greater AMR gene burden in their respiratory microbiome versus those with Paula Hayakawa Serpa and Xianding Deng contributed equally to this work. *Correspondence: chaz.langelier@ucsf.edu Chan Zuckerberg Biohub, San Francisco, CA, USA Full list of author information is available at the end of the article © The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Serpa et al. Genome Medicine (2022) 14:74 Page 2 of 12 community-onset LRTI (p = 0.00030), or those without LRTI (p = 0.0024). We found that Cas9 targeted sequencing could enrich for low abundance AMR genes by > 2500-fold and enabled their rapid detection using a nanopore platform. Conclusions: mNGS has utility for the detection and surveillance of resistant bacterial LRTI pathogens. Background AMR, facilitate epidemiological AMR surveillance, and Antimicrobial resistance (AMR) presents a clear threat rapidly detect clinically relevant resistance genes using to human health and is responsible for increasing rates CRISPR/Cas9 targeted enrichment  coupled with real- of treatment failure in patients with lower respiratory time nanopore sequencing. tract infections (LRTI), the leading cause of infectious disease-related mortality [1]. Implementing effective Methods and targeted therapies in patients with LRTI neces- Study design sitates not only accurate detection of a broad range We studied 70 mechanically ventilated patients with of pathogens, but also requires assessment of their LRTI and 18 with non-infectious respiratory illnesses resistance to antimicrobials. In many cases, assess- (Fig. 1, Additional File 1: Table S1) who were admitted to ment of AMR is not possible due to the need to first the University of California San Francisco (UCSF) Medi- isolate a bacterial pathogen in culture prior to antimi- cal Center between 07/2013 and 9/2018. Subjects with crobial susceptibility testing (AST), a process that can LRTI were identified by two-physician adjudication using require several days and have low yield in the setting of the United States Centers for Disease Control/National prior antibiotic use [2, 3]. In the absence of a definitive Healthcare Safety Network (CDC/NHSN) surveillance microbiologic diagnosis, LRTI treatment is by neces- case definition [13], a reference list of established respira - sity empiric, which leads to broad-spectrum antibiotic tory pathogens [6], and retrospective electronic medical overuse and selects for resistant pathogens [4, 5]. record review, blinded to mNGS results. Study inclu- Metagenomic next-generation sequencing (mNGS) sion criteria were age ≥ 18, need for mechanical ventila- holds promise for overcoming the limitations of tra- tion, and lower respiratory specimen (tracheal aspirate ditional respiratory diagnostics by affording culture- (TA) or mini-bronchoalveolar lavage (mBAL)) collected independent detection of pathogens and simultaneous within 72 h of intubation. Patients were excluded if they profiling of host gene expression signatures of infec - declined study participation, had unclear LRTI status, or tion [6]. In principle, mNGS can also be used to predict did not have matched RNA and DNA mNGS data avail- pathogen AMR by detecting bacterial resistance genes. able from a respiratory specimen (Fig. 1). While the performance of cultured bacterial isolate Primary analyses were performed for 27 patients, sec- whole genome sequencing has been extensively charac- ondary analyses for all subjects. The primary analysis terized [7], studies assessing the performance of direct focused on patients with bacterial LRTI due to culture- respiratory specimen mNGS for predicting AMR have confirmed pathogens that had been clinically tested remained more limited [8–12]. for susceptibility to antimicrobials (n = 27 patients , This is in part due to the low abundance of pathogen n = 32 pathogens) (Fig.  1, Table  1, Additional File 2: AMR genes in respiratory and other clinical body flu - Table  S2). Of these, 18 patients had respiratory sam- ids, which challenges their detection using conventional ples sequenced for a prior mNGS study by our group mNGS methods [12]. Recent work has demonstrated [6]. For secondary analyses, 43 additional patients with the potential for CRISPR/Cas9 targeted enrichment clinically adjudicated LRTI and 18 patients with no using FLASH (Finding Low Abundance Sequences by evidence of LRTI were assessed. In total, the second- Hybridization) to overcome this challenge by enhanc- ary outcome analysis of hospital exposure and AMR ing detection of low abundance AMR genes in clinical gene burden in the respiratory microbiome assessed 70 samples. Independent validation of FLASH in a clinical patients with LRTI and 18 patients with no evidence of cohort, however, has been needed. LRTI. Assessment of Cas9 targeted Illumina and nano- Here, we address these gaps by studying a cohort of pore sequencing for detecting AMR genes included 10 critically ill patients to assess the potential of both DNA patients from the primary analysis with culture-con- and RNA mNGS to predict LRTI bacterial pathogen firmed bacterial LRTI. S erpa et al. Genome Medicine (2022) 14:74 Page 3 of 12 Fig. 1 Study overview and analysis workflow. A Enrollment flow diagram for the critically ill adult cohort with acute respiratory illnesses that was studied. B Metagenomic next-generation sequencing (mNGS) approach and analysis workflow. The primary analysis assessed the performance of metagenomic next-generation sequencing (mNGS) antimicrobial resistance (AMR) prediction in 27 subjects with LRTI due to 32 culture-confirmed bacterial pathogens. Secondary analyses included mNGS epidemiological assessment of hospital exposure and AMR gene burden in the airway microbiome, and proof of concept assessment of CRISPR/Cas9 targeted mNGS using Illumina and real-time nanopore sequencing Procedures filtering with PRICESeqfilter [16], and additional filter - Nucleic acid extraction and Illumina metagenomic ing to remove non-microbial sequences. The identities of sequencing the remaining microbial reads were determined by query- RNA extraction from mBAL or TA and Illumina ing the NCBI nucleotide (NT) and non-redundant pro- metagenomic sequencing were carried out as described tein (NR) databases using GSNAP-L and RAPSEARCH2, previously [6, 14]. respectively [14]. Microbial alignments detected by RNA- seq and DNA-seq were aggregated to the genus level Pathogen detection bioinformatics and the sequencing reads comprising each genus were Detection of respiratory microbes leveraged the ID-Seq then evaluated for taxonomic assignment at the species pipeline [14] that incorporates the STAR [15] aligner to level based on species relative abundance. A recently subtract the human genome (NCBI GRC h38), quality developed rules-based model (RBM) [6] was employed Serpa et al. Genome Medicine (2022) 14:74 Page 4 of 12 Table 1 Performance of mNGS for genotypic prediction of antimicrobial susceptibility compared to a reference standard of clinical microbiologic testing. Sensitivity, specificity, and accuracy of DNA + RNA mNGS compared to a reference standard of clinical antimicrobial susceptibility testing based on Clinical & Laboratory Standards Institute (CLSI) minimum inhibitory concentration (MIC) breakpoints. A Gram-positive pathogens. B Gram-negative pathogens. AMR gene(s) detected by mNGS indicated. With respect to genotype-phenotype predictions, squares filled red indicate true positives, squares filled blue indicate true-negatives, squares with purple text = false negatives, squares with orange text are false positives *mutations in PBP1a/2x, Sens Sensitivity, Spec Specificity, TN True negative, FN False negative; n/a phenotypic susceptibility to antibiotic not tested in the clinical laboratory. 95% confidence interval (CI) listed below each sensitivity and specificity value S erpa et al. Genome Medicine (2022) 14:74 Page 5 of 12 to differentiate putative pathogens from commensal excluded due to the unclear clinical significance of the microbiota. isolated microbes. This left a primary analysis cohort of The RBM leverages previous findings demonstrating 27 patients and 32 bacterial pathogens (Fig. 1). that microbial communities in patients with LRTI are We assessed susceptibility to the most common anti- typically characterized by one or more dominant patho- biotics used for complicated infections from bacterial gens present in high abundance [6, 14]. More specifically, pathogens identified in the cohort: S. aureus, S. pneumo - the RBM ranks microbial genera present in a sample by niae, E. faecium, Enterobacteriaceae, P. aeruginosa, and descending abundance (number of taxonomic align- S. maltophila. Initial AMR gene class assignment (beta ments). The greatest difference between any two sequen - lactam, aminoglycoside, macrolide/lincosamide/ strep- tial taxa is then identified to capture genera present at togramin, glycopeptide, trimethoprim/ sulfamethoxa- disproportionately high abundance compared to the rest zole) was made using ontology in ARG-ANNOT [23] and of the lung microbiota [6, 17]. a more refined AMR phenotype assignment was made All genera with an abundance greater than this largest based on CARD [24] resistome ontological relationships. gap threshold are then evaluated at the species level, by In addition to sensitivity and specificity, we assessed very identifying the most abundant species within each genus. major error (VME; predicted susceptible but phenotypi- If the species is present within a previously curated ref- cally resistant) and major error (ME; predicted resistant erence index of established respiratory pathogens [6, 17] but phenotypically susceptible) rates. derived from landmark epidemiologic surveillance stud- ies [18–22], it is selected as a putative pathogen by the Clinically tested antimicrobials used in mNGS AMR prediction RBM. A detailed description of the principles and clinical benchmarking validation of the RBM has been previously published [6, Resistance predictions were made for antibiotics rou- 17]. tinely tested in the clinically microbiology laboratory for Staphylococcus aureus, Streptococcus pneumoniae, Ente- rococcus faecium, Pseudomonas aeruginosa, Stenotropho- Detection of AMR genes monas maltophila, and Enterobacteriaceae. For S. aureus AMR genes present in RNA-seq or DNA-seq data were these included penicillin, methicillin, clindamycin or identified using SRST2 coupled with an expanded ver - erythromycin, trimethoprim/ sulfamethoxazole (TMP/ sion of the ARG-ANNOT database [23] (Additional File SMZ), and vancomycin; for S. pneumoniae: penicillin, 3), and genes with ≥ 5% allele coverage were included in ceftriaxone, and vancomycin; for E. faecium: ampicillin analyses. Because Streptococcus pneumoniae is a leading and vancomycin; for Enterobacteriaceae: ampicillin + sul- cause of bacterial LRTI [4], we also screened for point bactam, cefazolin, ceftriaxone, gentamicin, piperacillin- mutations in pbp genes associated with Streptococcus tazobactam, TMP-SMX, ertapenem, and meropenem; for beta lactam resistance using the CARD resistance gene P. aeruginosa: ampicillin + sulbactam, ceftazidime, gen- identifier tool and the ‘loose’ setting [24]. Average read tamicin, piperacillin-tazobactam, and meropenem; and depth across each allele, normalized by gene length and for S. maltophila: ceftazidime and TMP/SMZ. For some total reads (depth per million reads sequenced, dpM), isolates, clinical susceptibility testing for certain antimi- was calculated for each sample. crobials was not performed by the clinical laboratory, and thus was unavailable for our analysis. Assessing performance of genotypic antimicrobial susceptibility prediction FLASH Cas9 targeted mNGS for AMR gene detection As a reference standard, we used clinical AST results FLASH Cas9 targeted Illumina mNGS for AMR gene performed in the UCSF Clinical Microbiology Labora- detection was carried out as described in the original tory during each patient’s admission. To calculate sen- proof of concept study [12]. Briefly, FLASHit software sitivity and specificity, which was done both by microbe [26] was first used to design guide RNAs targeting clini - and by drug (Table 1), we compared mNGS-based resist- cally relevant AMR genes derived from the CARD and ance predictions against phenotypic AST determined by ResFinder databases, merging exact duplicates [12]. In the Clinical & Laboratory Standards Institute minimum total, 2226 guide RNAs targeting 381 beta lactam and 111 inhibitory concentration breakpoints [25]. We studied MLS resistance genes, in addition to the 127 diverse AMR samples from subjects with culture-confirmed bacte - genes from the original FLASH pilot study, were utilized rial pathogens for which AST was performed. One iso- for Cas9 targeted enrichment. Guide RNAs targeted mul- late that only underwent chromogenic beta lactamase tiple sites on each AMR gene, which in total represented screening was excluded. Two subjects (252, 297) with 2226 target sequences (Additional File 4). DNA templates highly polymicrobial cultures of ≥ 4 organisms were also for producing CRISPR RNAs (crRNAs) for each AMR Serpa et al. Genome Medicine (2022) 14:74 Page 6 of 12 gene target were synthesized, pooled, transcribed, and outcome s include d the a ss o c i ation b e twe en ho s - purified according to described methods [12]. pital exposure and burden of A MR genes in the Ten nanograms of DNA was 5′ dephosphorylated respiratory microbiome and AMR gene detec - using rAPid alkaline phosphatase that was subsequently tion using Cas9 targeted enrichment and real- deactivated with sodium orthovanadate. The dephos - time nanop ore s e quenc ing . phorylated DNA was added to a master mix containing the CRISPR/Cas9 ribonucleoprotein complex and incu- Statistical analysis bated at 37  °C for 2  h. The Cas9 was deactivated with Statistical significance was defined as P less than 0.05, proteinase K and removed with SPRI bead purification. using two-tailed tests of hypotheses. Nonparamet- Samples were dA-tailed and then carried forward for ric continuous variables were analyzed by Wilcoxon Illumina Sequencing according to the NEBNext Ultra rank-sum. II library prep kit (New England Biolabs, Ipswich, MA) protocol according to previously described detailed methods [12]. AMR gene identification was carried out Results using ARG-ANNOT [23] as for the primary analyses, Cohort features and genes that were detected at a dpM of > 0.1 were Seventy subjects with LRTI and 18 with no evi- assessed for enrichment compared to DNA-seq alone. dence of LRTI were identified based on inclusion and exclusion criteria (Fig.  1, Additional File 1: Nanopore sequencing Table  S1). Primar y analyses were performed for 27 FLASH-enriched DNA libraries were quantified and patients , secondar y analyses for all subjects . Clini- 200–800  ng of DNA input was used for Nanopore 1D cal AST results were returned a median of 74  h library preparation (protocol SQK-LSK109, Oxford following sample collection (95% confidence inter- Nanopore, UK). Individual sample libraries were loaded val (CI) 49–115  h, (Additional File 5: Table  S3)). into a single flow cell of a GridION instrument, and Twenty-seven subjects with culture-confirmed sequencing reads were base called in real-time mode in bacterial LRTI, representing 32 pathogens with MINKNOW. The SURPIrt pipeline running in -a mode clinical AST data performed on ≥ 2 drugs, were was utilized to identify AMR genes every 100,000– identified and assessed in the primary analysis 200,000 reads as previously described [27, 28]. (Additional File 2: Table  S2). For secondary anal- yses, 43 additional patients with clinically adju- Mitigation of background contaminants dicated LRTI and 18 patients with no evidence of To minimize inaccurate taxonomic assignments due LRTI were assessed. to environmental contaminants, we processed nega- tive water controls with each group of samples that Metagenomic sequencing, pathogen, and AMR gene underwent nucleic acid extraction, and included these, detection as well as positive control clinical samples, with each A mean of 4.3 × 10 (interquartile range (IQR) 7 7 sequencing run. We directly subtracted alignments to 1.9–4.4 × 10 ) DNA-seq reads and 6.9 × 10 (IQR those taxa in water control samples detected by both 4.8–8.3 × 10 ) RNA-seq reads were generated from res- RNA-seq and DNA-seq analyses from the raw reads per piratory samples. In the primary AMR analysis group, we million (rpm) values in all samples [6]. To account for used a previously validated [6] metagenomic rules-based selective amplification bias of contaminants in water model (RBM) to identify bacterial respiratory pathogens controls resulting from PCR amplification of metagen - that were disproportionately abundant as compared to omic libraries to a fixed standard concentration across the rest of the lung microbiome. The RBM identified 26 all samples, prior to direct subtraction, we scaled taxa of 32 (81%) of the culture-confirmed bacterial patho - rpms in the water controls to the median percent gens from the primary analysis. Four (67%) of the missed microbial reads present across all samples as previously pathogens were present in the context of polymicro- described [6]. To address environmental contaminants bial cultures, and one (17%) was identified as a different in AMR gene analyses, resistance alleles detected in streptococcal species (Additional File 2: Table S2). A total water controls at a depth > 1 were excluded. of 138 and 234 acquired AMR genes were identified by RNA-seq and DNA-seq, respectively (Additional File 6: Study outcomes Table S4). With respect to AMR gene classes, beta lactam The pr imar y out c ome w a s p er f or manc e of mN GS resistance genes were most common (81/372 total genes, for predicting phenotypic AST. Secondary 35%). S erpa et al. Genome Medicine (2022) 14:74 Page 7 of 12 Fig. 2 A AMR genes detected in the lower respiratory microbiome of critically ill patients. Composite results of DNA and RNA mNGS. AMR genes are listed in rows and are grouped by antimicrobial class. Each column represents a patient respiratory sample and is grouped by LRTI status. B AMR gene burden in the respiratory tract, measured by averaging sequencing depth across the AMR allele per million reads sequenced (dpM) in the respiratory microbiome did not differ between LRTI-positive patients and those with non-infectious acute respiratory illnesses. C The burden of AMR genes detected in the lower respiratory tract microbiome was greater in patients with hospital-onset LRTI versus those with either community-onset LRTI or no evidence of LRTI. Legend: depth = average sequencing depth across each AMR gene allele normalized per million reads sequenced. Legend: Bla = beta lactam; AGly = aminoglycoside; Fos = Fosfomycin; Flq = fluoroquinolone; Gly = glycopeptide; Mac/ Lin/Str = macrolide, lincosamide, streptogramin; Phe = phenicol; Tet = tetracycline; Tmp-Sul = trimethoprim/sulfamethoxazole; depth = average sequencing depth across each AMR gene allele normalized per million reads sequenced. The horizontal bars in panels B and C indicate mean values Comparison of mNGS versus phenotypic antimicrobial 47–87%), specificity of 95% (CI 85–99%), and an accuracy susceptibility testing of 87% (CI 78–94%) (Table  1). This equated to a VME We assessed the performance of mNGS for predicting rate of 30% and a ME rate of 5%. For gram-negative path- resistance to clinical guideline-recommended antimicro- ogens, a combination of DNA-seq and RNA-seq yielded bials used for complicated gram-negative (n = 8 drugs) a sensitivity of 100% (CI 87–100%), specificity of 64% (CI and gram-positive (n = 6 drugs) infections. AMR genes 48–78%), and accuracy of 78% (CI 67–87%) (Table  1). unrelated to the culture-confirmed bacterial pathogen This equated to a VME rate of 0% and a ME rate of 36%. were identified through the resistome ontology annota - We also assessed the performance of RNA-seq and tions in CARD [24] and excluded from this analysis. Sen- DNA-seq performed independently (Table  1, Additional sitivity and specificity compared to a reference standard File 7: Table S5). RNA-seq performed with a sensitivity of of culture-based AST varied by pathogen, drug, patient, 52% (CI 31–73%), specificity of 100% (CI 94–100%), and and nucleic acid type sequenced (Table 1, Additional File accuracy of 86% (CI 76–93%) for gram-positive patho- 7: Table S5). For gram-positive pathogens, a combination gens, and a sensitivity of 100% (CI 89–100%), specificity of DNA-seq and RNA-seq yielded a sensitivity of 70% (CI of 64% (CI 48–78%), and accuracy of 79% (CI 68–88%) Serpa et al. Genome Medicine (2022) 14:74 Page 8 of 12 for gram-negative pathogens. DNA-seq performed with genes that were associated with the culture-confirmed a sensitivity of 39% (CI 20–61%), specificity of 95% (CI pathogen and resistance phenotype, but missed by DNA- 85–99%), and accuracy of 78% (CI 67–87%) for gram- seq alone, including mecA in two patients with methicil- positive pathogens, and a sensitivity of 58% (CI 39–75%), lin-resistant Staphylococcus aureus LRTI. FLASH also specificity of 67% (CI 50–80%), and accuracy of 63% (CI resulted in the detection of AMR genes unrelated to the 51–74%) for gram-negative pathogens. culture-confirmed pathogens in five (50%) of patients In two of seven cases with genotype to phenotype false- (Additional File 8: Table S6). positive (ME) predictions, mNGS identified AMR genes We subsequently assessed the potential for rapid unrelated to the culture-confirmed microbe but related AMR gene detection using FLASH combined with to resistant pathogens that would be cultured several an Oxford nanopore sequencing platform, which days later in the context of ventilator-associated pneu- affords real-time data generation (Fig.  3B). All AMR monia (VAP). These included SST-1 from a patient who genes identified by Illumina (median sequenc - 8 7 8 developed Serratia marcescens VAP (patient 213) 4  days ing depth of 1.20 × 10 [IQR 5.41 × 10 –1.36 × 10 ]) later, and mecA from a patient who developed Staphylo- were also identified by FLASH-nanopore sequencing coccus aureus VAP 7 days later (patient 232) (Additional (median sequencing depth of 1.19 × 10 reads [IQR 6 6 File 6: Table S4). 1.02 × 10 –1.46 × 10 ]) (Additional File 9: Table  S7). AMR gene targets could be identified within 10 min of real-time nanopore sequencing, suggesting a potential Association between LRTI positivity, hospital exposure, turnaround time of less than 6  h for a single sample and AMR genes in the respiratory microbiome (Fig. 3C). Assessment of the lower respiratory resistome using both DNA and RNA mNGS revealed a diversity of AMR genes in both LRTI-positive and negative patients (Fig.  2A, Discussion Additional File 6: Table  S4). AMR gene burden did not Antimicrobial resistance has emerged as one of the differ based on LRTI status (p = 0.28) (Fig.  2B). Sub- most pressing issues facing human health, and effective jects with hospital-onset (≥ 48  h after admission) LRTI treatment of complicated infections increasingly neces- had a greater burden of AMR genes in their respiratory sitates early and accurate assessment of microbial drug microbiome compared to those with community-onset resistance. Our study builds on prior respiratory mNGS LRTI (p = 0.00030) or those without LRTI (p = 0.0024) studies [6, 8–12] by demonstrating that RNA and DNA (Fig. 2C). mNGS can enable culture-independent prediction of AMR in critically ill patients with bacterial LRTI, with variable performance across pathogens and antimicro- Targeted enrichment and rapid detection of AMR genes bials. While false-negative susceptibility predictions for using Cas9 and nanopore sequencing gram-negative pathogens were not observed, we found We validated the utility of a recently described pro- a VME of 30% for gram-positive pathogens, suggesting grammable CRISPR/Cas9-based method called FLASH mNGS has a role for complementing, rather than replac- (Finding Low Abundance Sequences by Hybridization) ing, current standard of care culture-based approaches. to enrich for low abundance AMR genes [12] by study- Prior work has demonstrated the utility of mNGS in ing 10 patients from the primary analysis (Additional cases of culture-negative LRTI, which represent more File 8: Table  S6). The FLASH + DNA mNGS library than half of all pneumonia cases [4, 6, 8, 10]. Our results prep protocol added 2.5  h to the standard 3-h NEBNext suggest that mNGS may also have potential for predict- DNA-seq workflow [29] and could enrich detection of ing AST in culture-negative LRTI, where detection of AMR genes associated with the culture-confirmed path - an AMR gene could inform the need for treatment of an ogen > 2500 × compared to DNA-seq alone (Fig.  3A). In occult resistant organism. Further, our findings support four (40%) patients, FLASH enabled detection of AMR recent observations [10] that mNGS may have utility for (See figure on next page.) Fig. 3 A FLASH (Finding Low Abundance Sequences by Hybridization) CRISPR/Cas9 targeted Illumina sequencing enriched the detection of culture-confirmed bacterial LRTI pathogen AMR alleles by 46 × to > 2500 × versus DNA-seq alone. B Workflow diagram for FLASH targeted enrichment coupled with nanopore sequencing. Time estimates provided for a single sample. C Real-time detection of AMR genes by FLASH targeted nanopore sequencing was achieved within 10 min following mNGS library preparation. Data from two representative Staphylococcus aureus LRTI cases are highlighted. Case 212 (left panel) highlights a case where detection of BlaZ and MsrA/ErmA genes correlated with phenotypically determined penicillin and macrolide/lincosamide resistance, respectively. Case 288 (right panel) highlights a case where detection of MecA, BlaZ, and MsrA correlated with phenotypically confirmed methicillin, penicillin, and macrolide resistance, respectively S erpa et al. Genome Medicine (2022) 14:74 Page 9 of 12 - FLASH + FLASH -1 -2 -3 289 314 407 409 268 298 212 288 350 382 300 10000 25 BlaZ MecA Erm BlaZ MsrA 0 0 10 30 90 10 30 90 time (min) time (min) Fig. 3 (See legend on previous page.) BlaZ reads AMR gene dpM BlaZ ErmA BlaZ MsrA MecA BlaZ ErmC BlaZ Erm reads BlaZ AMR gene reads ErmA MecA OXA-50 ACT-MIR SRT-SST SRT-SST Serpa et al. Genome Medicine (2022) 14:74 Page 10 of 12 early identification of future secondary respiratory infec - LRTI and that it systematically assessed performance tions. For instance, patient 232, who was admitted for for multiple classes of antimicrobials against a clinical severe Klebsiella pneumoniae LRTI, developed MRSA reference standard. Further, we provide the first culture- ventilator-associated pneumonia (VAP) eight days later independent assessment of healthcare exposure and while undergoing treatment with aztreonam, an anti- resistance gene burden in the respiratory tract, and a biotic lacking MRSA coverage. mNGS analysis of a TA novel demonstration of Cas9 targeted enrichment cou- specimen obtained 7 days before VAP onset revealed the pled with rapid nanopore sequencing. Our study also mecA gene, providing an early indication of the subse- has several limitations, including sample size, spectrum quent AMR VAP pathogen. of antimicrobial classes assessed, spectrum of patho- Public health surveillance is essential for pandemic gens assessed, and the need for independent validation preparedness, understanding trends in AMR, and pre- of findings. Future work in a larger, prospective cohort venting outbreaks of resistant organisms. Our findings with a greater diversity of bacterial pathogens and resist- demonstrate the feasibility of mNGS for epidemiologi- ance mechanisms can address these limitations. In addi- cal surveillance of AMR and highlight an association tion, a randomized clinical trial will be needed to assess between hospital exposure and AMR gene burden in the the potential impact of mNGS on time to appropriate lower respiratory tract microbiome. Importantly, our antimicrobial treatment, antimicrobial stewardship, and analysis included culture-negative LRTI cases and adds LRTI outcomes. Lastly, genotype to phenotype predic- to prior literature demonstrating a similar association in tion remains imperfect, even for cultured isolates [7, 35]. cases of culture-confirmed bacterial pneumonia [30]. As with other molecular testing modalities for AMR, Nanopore sequencing has proven useful for microbial this should be recognized when considering the utility detection from respiratory samples with high pathogen and clinical applicability of mNGS for AMR phenotypic abundance [8, 10, 31]; however, basecalling accuracy and prediction. sensitivity challenges have historically limited its capacity for detecting underrepresented sequences in metagen- Conclusions omic datasets [32]. Targeted enrichment can overcome In summary, we characterize the utility of mNGS for pre- this through AMR signal amplification, and consistent dicting AMR in bacterial LRTI and demonstrate proof of with this, we found high concordance for AMR gene concept for both epidemiological AMR surveillance and detection between nanopore and Illumina samples that rapid resistance gene detection using Cas9 and nanopore underwent FLASH Cas9 targeted enrichment. Our sequencing. results suggest that this method could also potentially augment AMR gene detection and resistance prediction Supplementary Information if coupled with established nanopore mNGS workflows The online version contains supplementary material available at https:// doi. for detecting respiratory pathogens [8–10]. org/ 10. 1186/ s13073- 022- 01072-4. Rapid detection of AMR is essential in critically ill Additional file 1: Table S1. Clinical and demographic features of cohort. patients with severe bacterial infections given that time to appropriate antimicrobials correlates with mortality [33]. Additional file 2: Table S2. mNGS detection of respiratory patho- gens compared to a reference standard of clinical microbiological FLASH adds 2.5  h to standard DNA library preparation testing. workflow and, when coupled with real-time detection of Additional file 3: Dataset 1. AMR gene database utilized for analyses. AMR genes from mNGS libraries, could potentially allow Additional file 4: Dataset 2. AMR genes and guide RNAs for FLASH for sample to answer in under 6 h, a significant time sav - targeted mNGS. AMR genes and guide RNAs for FLASH targeted mNGS. ings compared to the ≥ 24 h required for Illumina proto- A) AMR genes targeted by FLASH. B) AMR alleles and guide RNA targets. Legend: Bla = beta lactam; MLS = macrolide, Tmp/Sul = trimethoprim/ cols [34]. In our cohort, clinical AST required an average sulfamethoxazole. of 74 h, suggesting that Cas9-targeted nanopore sequenc- Additional file 5: Table S3. Time to return of clinical antimicrobial suscep - ing may be a promising future approach for more rapidly tibility testing results for culture-confirmed bacterial pathogens. identifying patients with resistant infections. We found Additional file 6: Table S4. AMR genes detected by mNGS. that FLASH also enriched for AMR genes unrelated to Additional file 7: Table S5. Performance of RNA-seq, DNA-seq and the culture-confirmed pathogen, presumably derived combined RNA-seq + DNA-seq for AMR prediction compared to clinical from the lung microbiome. Improved methods for anno- antimicrobial susceptibility testing. tating the species-specificity of detected AMR genes may Additional file 8: Table S6. Pathogen AMR genes detected by help address this issue, which otherwise could lead to FLASH-mNGS. increased ME due to false-positive results. Additional file 9: Table S7. AMR genes detected by FLASH coupled with Nanopore mNGS. Strengths of this study include that it is the largest to assess the performance of mNGS AMR prediction in S erpa et al. Genome Medicine (2022) 14:74 Page 11 of 12 spectrum antibiotic treatment in patients with community acquired Acknowledgements pneumonia: a prospective randomised study. Thorax. 2005;60:672–8. Not applicable. 4. Jain S, Self WH, Wunderink RG, et al. Community-acquired pneu- monia requiring hospitalization among U.S. adults. N Engl J Med. Authors’ contributions 2015;373:415–27. PHS, XD: collection and assembly of data, data analysis and interpretation, 5. Leffler DA, Lamont JT. Clostridium difficile infection. N Engl J Med. manuscript editing. MA, KM, SC, MF, NN: collection and assembly of data, 2015;372:1539–48. manuscript editing. RG, TD: collection and assembly of data. EC, AL.: methods 6. Langelier C, Kalantar KL, Moazed F, et al. Integrating host response and development and manuscript editing. AM, KK: data analysis and interpretation. unbiased microbe detection for lower respiratory tract infection diagno- SD, SM: manuscript editing and data interpretation. CCa, CCh, JD: conception sis in critically ill adults. Proc Natl Acad Sci USA. 2018;115(52):E12353–62 and design, manuscript editing, funding. CL: conception and design, data (201809700). analysis and interpretation, manuscript writing, funding. All authors read and 7. Mahfouz N, Ferreira I, Beisken S, von Haeseler A, Posch AE. Large-scale approved the final manuscript. assessment of antimicrobial resistance marker databases for genetic phenotype prediction: a systematic review. J Antimicrob Chemother. Funding 2020;75:3099–108. NIH R35 HL140026 (CSC), NHLBI K23HL138461-01A1 (CL), Chan Zuckerberg 8. Charalampous T, Kay GL, Richardson H, et al. Nanopore metagenomics Biohub (CL, JLD). The funders of the study had no role in study design, data enables rapid clinical diagnosis of bacterial lower respiratory infection. collection, data analysis, data interpretation, or writing of the report. Nat Biotechnol. 2019;37:783–92. 9. Yang L, Haidar G, Zia H, et al. Metagenomic identification of severe pneu- Availability of data and materials monia pathogens in mechanically-ventilated patients: a feasibility and Raw microbial sequences are available via NCBI BioProject PRJNA450137 clinical validity study. Respir Res. 2019;20:265. [36], https:// www. ncbi. nlm. nih. gov/ biopr oject/? term = PRJNA450137, and 10. Charalampous T, Alcolea-Medina A, Snell LB, et al. Evaluating the poten- BioProject PRJNA635133 [37], https:// www. ncbi. nlm. nih. gov/ biopr oject/? tial for respiratory metagenomics to improve treatment of secondary term = PRJNA635133. FLASHit [26] and SURPIrt [28] have been deposited on infection and detection of nosocomial transmission on expanded COVID- Github and are available for download at: https:// github. com/ czbio hub/ flash 19 intensive care units. Genome Med. 2021;13:182. and https:// github. com/ chiul ab/ SURPI- plus- dist, respectively. 11. Chao L, Li J, Zhang Y, Pu H, Yan X. Application of next generation sequencing-based rapid detection platform for microbiological diagnosis and drug resistance prediction in acute lower respiratory infection. Ann Declarations Transl Med. 2020;8:1644. 12. Quan J, Langelier C, Kuchta A, et al. FLASH: a next-generation CRISPR Ethics approval and consent to participate diagnostic for multiplexed detection of antimicrobial resistance Patients were enrolled via a previously described observational cohort study sequences. Nucleic Acids Res 2019; published online May 22. https:// doi. [6, 38] which was approved by the UCSF Institutional Review Board (IRB), org/ 10. 1093/ nar/ gkz418. which granted a waiver of initial consent for respiratory sampling. Informed 13. United States Centers for Disease Control and Prevention. CDC/NHSN consent was subsequently obtained from patients or their surrogates for surveillance definitions for specific types of infections. 2021; published continued study participation, following the previously detailed IRB protocol online Jan. https:// www. cdc. gov/ nhsn/ pdfs/ pscma nual/ pcsma nual_ #17–24056 [38]. This research conformed to the principles of the Helsinki curre nt. pdf. Declaration. 14. Kalantar KL, Carvalho T, de Bourcy CFA, et al. IDseq—An open source cloud-based pipeline and analysis service for metagenomic pathogen Consent for publication detection and monitoring. Gigascience 2020; 9. https:// doi. org/ 10. 1093/ Consent to publish patient-related information was obtained as part of gigas cience/ giaa1 11. informed consent for study participation as described above. 15. Dobin A, Davis CA, Schlesinger F, et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics. 2013;29:15–21. Competing interests 16. Ruby JG, Bellare P, Derisi JL. PRICE: software for the targeted assembly The authors declare that they have no competing interests. of components of (Meta) genomic sequence data. G3 (Bethesda). 2013;3:865–80. Author details 1 17. Tsitsiklis A, Osborne CM, Kamm J, et al. Lower respiratory tract infections Division of Infectious Diseases, Department of Medicine, University of Cali- 2 in children requiring mechanical ventilation: a multicentre prospective fornia, San Francisco, San Francisco, CA, USA. Chan Zuckerberg Biohub, San 3 surveillance study incorporating airway metagenomics. Lancet Microbe. Francisco, CA, USA. Department of Laboratory Medicine, University of Califor- 4 2022;3:e284–93. nia, San Francisco, CA, USA. Department of Microbiology and Immunology, 5 18. Jain S, Self WH, Wunderink RG, et al. Community-acquired pneu- University of California, San Francisco, CA, USA. Division of Pulmonary, Critical monia requiring hospitalization among U.S. adults. N Engl J Med. Care, Allergy and Sleep Medicine, Department of Medicine, University of Cali- 6 2015;373:415–27. fornia, San Francisco, CA, USA. Chan Zuckerberg Initiative, San Francisco, CA, 7 19. Infection in Organ Transplantation - Fishman - 2017 - American Journal of USA. Department of Biochemistry and Biophysics, University of California, San Transplantation - Wiley Online Library. https:// doi. org/ 10. 1111/ ajt. 14208 Francisco, CA, USA. (Accessed 21 Jul 2021). 20. Magill SS, Edwards JR, Bamberg W, et al. Multistate point-preva- Received: 6 February 2022 Accepted: 15 June 2022 lence survey of health care-associated infections. N Engl J Med. 2014;370:1198–208. 21. Kalil AC, Metersky ML, Klompas M, et al. Management of adults with hospital-acquired and ventilator-associated pneumonia: 2016 Clinical Practice Guidelines by the Infectious Diseases Society of America and the References American Thoracic Society. Clin Infect Dis. 2016;63:e61-111. 1. O’Neill J. Tackling drug-resistant infections globally: final report and rec- 22. Mandell LA, Wunderink RG, Anzueto A, et al. Infectious Diseases Society ommendations. London, UK May 16, 2016 2016. 2016; published online of America/American Thoracic Society consensus guidelines on the May 16. http:// amr- review. org/ Publi catio ns. management of community-acquired pneumonia in adults. Clin Infect 2. Zaas AK, Garner BH, Tsalik EL, Burke T, Woods CW, Ginsburg GS. The cur- Dis. 2007;44(Suppl 2):S27-72. rent epidemiology and clinical decisions surrounding acute respiratory 23. Inouye M, Dashnow H, Raven L-A, et al. SRST2: Rapid genomic surveil- infections. Trends Mol Med. 2014;20:579–88. lance for public health and hospital microbiology labs. Genome Med. 3. van der Eerden MM, Vlaspolder F, de Graaff CS, et al. Comparison 2014;6:90. between pathogen directed antibiotic treatment and empirical broad Serpa et al. Genome Medicine (2022) 14:74 Page 12 of 12 24. Jia B, Raphenya AR, Alcock B, et al. CARD 2017: expansion and model- centric curation of the comprehensive antibiotic resistance database. Nucleic Acids Res. 2017;45:D566–73. 25. CLSI. CLSI M100-ED31:2021 Performance Standards for Antimicrobial Susceptibility Testing, 31st Edition. 2021; published online March. http:// em100. edapt ivedo cs. net/ dashb oard. aspx. 26. Quan J, Langelier CR, Batson J, Crawford ED. FLASHit. Github. https:// github. com/ czbio hub/ flash 2020. 27. Deng X, Achari A, Federman S, et al. Metagenomic sequencing with spiked primer enrichment for viral diagnostics and genomic surveillance. Nature Microbiology 2020; published online Jan 13. https:// doi. org/ 10. 1038/ s41564- 019- 0637-9. 28. Chiu CY, Stryke D. SURPI+, a rapid computational pipeline for compre- hensive identification of pathogens from clinical metagenomic sequence data. Github. https:// github. com/ chiul ab/ SURPI- plus- dist 2019. 29. New England Biolabs. Improved library preparation with the NEBNext UltraTM II DNA Library Prep Kit for Illumina. https:// www. neb- online. de/ wp- conte nt/ uploa ds/ 2020/ 03/ Impro ved- libra ry- prepa ration- with- the- NEBNe xt- Ultra- II- DNA- Libra ry- Prep- Kit- for- Illum ina- E7645. pdf (Accessed 9 Dec 2021). 30. Kalil AC, Metersky ML, Klompas M, et al. Management of adults with hospital-acquired and ventilator-associated pneumonia: 2016 Clinical Practice Guidelines by the Infectious Diseases Society of America and the American Thoracic Society. Clin Infect Dis. 2016;63:e61-111. 31. Greninger AL, Naccache SN, Federman S, et al. Rapid metagenomic identification of viral pathogens in clinical samples by real-time nanopore sequencing analysis. Genome Med. 2015;7:99. https:// doi. org/ 10. 1186/ s13073- 015- 0220-9. 32. Rang FJ, Kloosterman WP, de Ridder J. From squiggle to basepair: compu- tational approaches for improving nanopore sequencing read accuracy. Genome Biol. 2018;19:90. https:// doi. org/ 10. 1186/ s13059- 018- 1462-9. 33. Lee C-C, Lee C-H, Yang C-Y, Hsieh C-C, Tang H-J, Ko W-C. Beneficial effects of early empirical administration of appropriate antimicrobials on survival and defervescence in adults with community-onset bacteremia. Crit Care. 2019;23:363. 34. Illumina, Inc. Run time estimates for each sequencing step on the Illumina sequencing platforms. 2020; published online May 15. https:// emea. suppo rt. illum ina. com/ bulle tins/ 2017/ 02/ run- time- estim ates- for- each- seque ncing- step- on- illum ina- seque nci. html (Accessed 9 Dec 2021). 35. Yee R, Dien Bard J, Simner PJ. The genotype-to-phenotype dilemma: how should laboratories approach discordant susceptibility results? J Clin Microbiol. 2021;59:e00138-e220. 36. Langelier CR, Kalantar KL. Combined host and microbe NGS for lower res- piratory tract infection diagnosis in critically ill adults microbial sequence reads. BioProject PRJNA450137, Sequence Read Archive. 2018. https:// www. ncbi. nlm. nih. gov/ biopr oject/? term=% 20PRJ NA450 137. 37. Langelier CR, Hayakawa Serpa P. CAS9-mNGS detects antimicrobial resist- ant pathogens. BioProject PRJNA635133, Sequence Read Archive. 2020. https:// www. ncbi. nlm. nih. gov/ biopr oject/? term= PRJNA 635133. 38. Sarma A, Christenson SA, Byrne A, et al. Tracheal aspirate RNA sequencing identifies distinct immunological features of COVID-19 ARDS. Nat Com- mun. 2021;12:5152. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations. Re Read ady y to to submit y submit your our re researc search h ? Choose BMC and benefit fr ? Choose BMC and benefit from om: : fast, convenient online submission thorough peer review by experienced researchers in your field rapid publication on acceptance support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Genome Medicine Springer Journals

Metagenomic prediction of antimicrobial resistance in critically ill patients with lower respiratory tract infections

Download PDF
12 pages

Loading next page...
 
/lp/springer-journals/metagenomic-prediction-of-antimicrobial-resistance-in-critically-ill-ZWoAg0k3n8

References (23)

Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2022
eISSN
1756-994X
DOI
10.1186/s13073-022-01072-4
Publisher site
See Article on Publisher Site

Abstract

Background: Antimicrobial resistance (AMR) is rising at an alarming rate and complicating the management of infec- tious diseases including lower respiratory tract infections (LRTI). Metagenomic next-generation sequencing (mNGS) is a recently established method for culture-independent LRTI diagnosis, but its utility for predicting AMR has remained unclear. We aimed to assess the performance of mNGS for AMR prediction in bacterial LRTI and demonstrate proof of concept for epidemiological AMR surveillance and rapid AMR gene detection using Cas9 enrichment and nanopore sequencing. Methods: We studied 88 patients with acute respiratory failure between 07/2013 and 9/2018, enrolled through a previous observational study of LRTI. Inclusion criteria were age ≥ 18, need for mechanical ventilation, and respira- tory specimen collection within 72 h of intubation. Exclusion criteria were decline of study participation, unclear LRTI status, or no matched RNA and DNA mNGS data from a respiratory specimen. Patients with LRTI were identi- fied by clinical adjudication. mNGS was performed on lower respiratory tract specimens. The primary outcome was mNGS performance for predicting phenotypic antimicrobial susceptibility and was assessed in patients with LRTI from culture-confirmed bacterial pathogens with clinical antimicrobial susceptibility testing (n = 27 patients, n = 32 pathogens). Secondary outcomes included the association between hospital exposure and AMR gene burden in the respiratory microbiome (n = 88 patients), and AMR gene detection using Cas9 targeted enrichment and nanopore sequencing (n = 10 patients). Results: Compared to clinical antimicrobial susceptibility testing, the performance of respiratory mNGS for predict- ing AMR varied by pathogen, antimicrobial, and nucleic acid type sequenced. For gram-positive bacteria, a combi- nation of RNA + DNA mNGS achieved a sensitivity of 70% (95% confidence interval (CI) 47–87%) and specificity of 95% (CI 85–99%). For gram-negative bacteria, sensitivity was 100% (CI 87–100%) and specificity 64% (CI 48–78%). Patients with hospital-onset LRTI had a greater AMR gene burden in their respiratory microbiome versus those with Paula Hayakawa Serpa and Xianding Deng contributed equally to this work. *Correspondence: chaz.langelier@ucsf.edu Chan Zuckerberg Biohub, San Francisco, CA, USA Full list of author information is available at the end of the article © The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. The Creative Commons Public Domain Dedication waiver (http:// creat iveco mmons. org/ publi cdoma in/ zero/1. 0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Serpa et al. Genome Medicine (2022) 14:74 Page 2 of 12 community-onset LRTI (p = 0.00030), or those without LRTI (p = 0.0024). We found that Cas9 targeted sequencing could enrich for low abundance AMR genes by > 2500-fold and enabled their rapid detection using a nanopore platform. Conclusions: mNGS has utility for the detection and surveillance of resistant bacterial LRTI pathogens. Background AMR, facilitate epidemiological AMR surveillance, and Antimicrobial resistance (AMR) presents a clear threat rapidly detect clinically relevant resistance genes using to human health and is responsible for increasing rates CRISPR/Cas9 targeted enrichment  coupled with real- of treatment failure in patients with lower respiratory time nanopore sequencing. tract infections (LRTI), the leading cause of infectious disease-related mortality [1]. Implementing effective Methods and targeted therapies in patients with LRTI neces- Study design sitates not only accurate detection of a broad range We studied 70 mechanically ventilated patients with of pathogens, but also requires assessment of their LRTI and 18 with non-infectious respiratory illnesses resistance to antimicrobials. In many cases, assess- (Fig. 1, Additional File 1: Table S1) who were admitted to ment of AMR is not possible due to the need to first the University of California San Francisco (UCSF) Medi- isolate a bacterial pathogen in culture prior to antimi- cal Center between 07/2013 and 9/2018. Subjects with crobial susceptibility testing (AST), a process that can LRTI were identified by two-physician adjudication using require several days and have low yield in the setting of the United States Centers for Disease Control/National prior antibiotic use [2, 3]. In the absence of a definitive Healthcare Safety Network (CDC/NHSN) surveillance microbiologic diagnosis, LRTI treatment is by neces- case definition [13], a reference list of established respira - sity empiric, which leads to broad-spectrum antibiotic tory pathogens [6], and retrospective electronic medical overuse and selects for resistant pathogens [4, 5]. record review, blinded to mNGS results. Study inclu- Metagenomic next-generation sequencing (mNGS) sion criteria were age ≥ 18, need for mechanical ventila- holds promise for overcoming the limitations of tra- tion, and lower respiratory specimen (tracheal aspirate ditional respiratory diagnostics by affording culture- (TA) or mini-bronchoalveolar lavage (mBAL)) collected independent detection of pathogens and simultaneous within 72 h of intubation. Patients were excluded if they profiling of host gene expression signatures of infec - declined study participation, had unclear LRTI status, or tion [6]. In principle, mNGS can also be used to predict did not have matched RNA and DNA mNGS data avail- pathogen AMR by detecting bacterial resistance genes. able from a respiratory specimen (Fig. 1). While the performance of cultured bacterial isolate Primary analyses were performed for 27 patients, sec- whole genome sequencing has been extensively charac- ondary analyses for all subjects. The primary analysis terized [7], studies assessing the performance of direct focused on patients with bacterial LRTI due to culture- respiratory specimen mNGS for predicting AMR have confirmed pathogens that had been clinically tested remained more limited [8–12]. for susceptibility to antimicrobials (n = 27 patients , This is in part due to the low abundance of pathogen n = 32 pathogens) (Fig.  1, Table  1, Additional File 2: AMR genes in respiratory and other clinical body flu - Table  S2). Of these, 18 patients had respiratory sam- ids, which challenges their detection using conventional ples sequenced for a prior mNGS study by our group mNGS methods [12]. Recent work has demonstrated [6]. For secondary analyses, 43 additional patients with the potential for CRISPR/Cas9 targeted enrichment clinically adjudicated LRTI and 18 patients with no using FLASH (Finding Low Abundance Sequences by evidence of LRTI were assessed. In total, the second- Hybridization) to overcome this challenge by enhanc- ary outcome analysis of hospital exposure and AMR ing detection of low abundance AMR genes in clinical gene burden in the respiratory microbiome assessed 70 samples. Independent validation of FLASH in a clinical patients with LRTI and 18 patients with no evidence of cohort, however, has been needed. LRTI. Assessment of Cas9 targeted Illumina and nano- Here, we address these gaps by studying a cohort of pore sequencing for detecting AMR genes included 10 critically ill patients to assess the potential of both DNA patients from the primary analysis with culture-con- and RNA mNGS to predict LRTI bacterial pathogen firmed bacterial LRTI. S erpa et al. Genome Medicine (2022) 14:74 Page 3 of 12 Fig. 1 Study overview and analysis workflow. A Enrollment flow diagram for the critically ill adult cohort with acute respiratory illnesses that was studied. B Metagenomic next-generation sequencing (mNGS) approach and analysis workflow. The primary analysis assessed the performance of metagenomic next-generation sequencing (mNGS) antimicrobial resistance (AMR) prediction in 27 subjects with LRTI due to 32 culture-confirmed bacterial pathogens. Secondary analyses included mNGS epidemiological assessment of hospital exposure and AMR gene burden in the airway microbiome, and proof of concept assessment of CRISPR/Cas9 targeted mNGS using Illumina and real-time nanopore sequencing Procedures filtering with PRICESeqfilter [16], and additional filter - Nucleic acid extraction and Illumina metagenomic ing to remove non-microbial sequences. The identities of sequencing the remaining microbial reads were determined by query- RNA extraction from mBAL or TA and Illumina ing the NCBI nucleotide (NT) and non-redundant pro- metagenomic sequencing were carried out as described tein (NR) databases using GSNAP-L and RAPSEARCH2, previously [6, 14]. respectively [14]. Microbial alignments detected by RNA- seq and DNA-seq were aggregated to the genus level Pathogen detection bioinformatics and the sequencing reads comprising each genus were Detection of respiratory microbes leveraged the ID-Seq then evaluated for taxonomic assignment at the species pipeline [14] that incorporates the STAR [15] aligner to level based on species relative abundance. A recently subtract the human genome (NCBI GRC h38), quality developed rules-based model (RBM) [6] was employed Serpa et al. Genome Medicine (2022) 14:74 Page 4 of 12 Table 1 Performance of mNGS for genotypic prediction of antimicrobial susceptibility compared to a reference standard of clinical microbiologic testing. Sensitivity, specificity, and accuracy of DNA + RNA mNGS compared to a reference standard of clinical antimicrobial susceptibility testing based on Clinical & Laboratory Standards Institute (CLSI) minimum inhibitory concentration (MIC) breakpoints. A Gram-positive pathogens. B Gram-negative pathogens. AMR gene(s) detected by mNGS indicated. With respect to genotype-phenotype predictions, squares filled red indicate true positives, squares filled blue indicate true-negatives, squares with purple text = false negatives, squares with orange text are false positives *mutations in PBP1a/2x, Sens Sensitivity, Spec Specificity, TN True negative, FN False negative; n/a phenotypic susceptibility to antibiotic not tested in the clinical laboratory. 95% confidence interval (CI) listed below each sensitivity and specificity value S erpa et al. Genome Medicine (2022) 14:74 Page 5 of 12 to differentiate putative pathogens from commensal excluded due to the unclear clinical significance of the microbiota. isolated microbes. This left a primary analysis cohort of The RBM leverages previous findings demonstrating 27 patients and 32 bacterial pathogens (Fig. 1). that microbial communities in patients with LRTI are We assessed susceptibility to the most common anti- typically characterized by one or more dominant patho- biotics used for complicated infections from bacterial gens present in high abundance [6, 14]. More specifically, pathogens identified in the cohort: S. aureus, S. pneumo - the RBM ranks microbial genera present in a sample by niae, E. faecium, Enterobacteriaceae, P. aeruginosa, and descending abundance (number of taxonomic align- S. maltophila. Initial AMR gene class assignment (beta ments). The greatest difference between any two sequen - lactam, aminoglycoside, macrolide/lincosamide/ strep- tial taxa is then identified to capture genera present at togramin, glycopeptide, trimethoprim/ sulfamethoxa- disproportionately high abundance compared to the rest zole) was made using ontology in ARG-ANNOT [23] and of the lung microbiota [6, 17]. a more refined AMR phenotype assignment was made All genera with an abundance greater than this largest based on CARD [24] resistome ontological relationships. gap threshold are then evaluated at the species level, by In addition to sensitivity and specificity, we assessed very identifying the most abundant species within each genus. major error (VME; predicted susceptible but phenotypi- If the species is present within a previously curated ref- cally resistant) and major error (ME; predicted resistant erence index of established respiratory pathogens [6, 17] but phenotypically susceptible) rates. derived from landmark epidemiologic surveillance stud- ies [18–22], it is selected as a putative pathogen by the Clinically tested antimicrobials used in mNGS AMR prediction RBM. A detailed description of the principles and clinical benchmarking validation of the RBM has been previously published [6, Resistance predictions were made for antibiotics rou- 17]. tinely tested in the clinically microbiology laboratory for Staphylococcus aureus, Streptococcus pneumoniae, Ente- rococcus faecium, Pseudomonas aeruginosa, Stenotropho- Detection of AMR genes monas maltophila, and Enterobacteriaceae. For S. aureus AMR genes present in RNA-seq or DNA-seq data were these included penicillin, methicillin, clindamycin or identified using SRST2 coupled with an expanded ver - erythromycin, trimethoprim/ sulfamethoxazole (TMP/ sion of the ARG-ANNOT database [23] (Additional File SMZ), and vancomycin; for S. pneumoniae: penicillin, 3), and genes with ≥ 5% allele coverage were included in ceftriaxone, and vancomycin; for E. faecium: ampicillin analyses. Because Streptococcus pneumoniae is a leading and vancomycin; for Enterobacteriaceae: ampicillin + sul- cause of bacterial LRTI [4], we also screened for point bactam, cefazolin, ceftriaxone, gentamicin, piperacillin- mutations in pbp genes associated with Streptococcus tazobactam, TMP-SMX, ertapenem, and meropenem; for beta lactam resistance using the CARD resistance gene P. aeruginosa: ampicillin + sulbactam, ceftazidime, gen- identifier tool and the ‘loose’ setting [24]. Average read tamicin, piperacillin-tazobactam, and meropenem; and depth across each allele, normalized by gene length and for S. maltophila: ceftazidime and TMP/SMZ. For some total reads (depth per million reads sequenced, dpM), isolates, clinical susceptibility testing for certain antimi- was calculated for each sample. crobials was not performed by the clinical laboratory, and thus was unavailable for our analysis. Assessing performance of genotypic antimicrobial susceptibility prediction FLASH Cas9 targeted mNGS for AMR gene detection As a reference standard, we used clinical AST results FLASH Cas9 targeted Illumina mNGS for AMR gene performed in the UCSF Clinical Microbiology Labora- detection was carried out as described in the original tory during each patient’s admission. To calculate sen- proof of concept study [12]. Briefly, FLASHit software sitivity and specificity, which was done both by microbe [26] was first used to design guide RNAs targeting clini - and by drug (Table 1), we compared mNGS-based resist- cally relevant AMR genes derived from the CARD and ance predictions against phenotypic AST determined by ResFinder databases, merging exact duplicates [12]. In the Clinical & Laboratory Standards Institute minimum total, 2226 guide RNAs targeting 381 beta lactam and 111 inhibitory concentration breakpoints [25]. We studied MLS resistance genes, in addition to the 127 diverse AMR samples from subjects with culture-confirmed bacte - genes from the original FLASH pilot study, were utilized rial pathogens for which AST was performed. One iso- for Cas9 targeted enrichment. Guide RNAs targeted mul- late that only underwent chromogenic beta lactamase tiple sites on each AMR gene, which in total represented screening was excluded. Two subjects (252, 297) with 2226 target sequences (Additional File 4). DNA templates highly polymicrobial cultures of ≥ 4 organisms were also for producing CRISPR RNAs (crRNAs) for each AMR Serpa et al. Genome Medicine (2022) 14:74 Page 6 of 12 gene target were synthesized, pooled, transcribed, and outcome s include d the a ss o c i ation b e twe en ho s - purified according to described methods [12]. pital exposure and burden of A MR genes in the Ten nanograms of DNA was 5′ dephosphorylated respiratory microbiome and AMR gene detec - using rAPid alkaline phosphatase that was subsequently tion using Cas9 targeted enrichment and real- deactivated with sodium orthovanadate. The dephos - time nanop ore s e quenc ing . phorylated DNA was added to a master mix containing the CRISPR/Cas9 ribonucleoprotein complex and incu- Statistical analysis bated at 37  °C for 2  h. The Cas9 was deactivated with Statistical significance was defined as P less than 0.05, proteinase K and removed with SPRI bead purification. using two-tailed tests of hypotheses. Nonparamet- Samples were dA-tailed and then carried forward for ric continuous variables were analyzed by Wilcoxon Illumina Sequencing according to the NEBNext Ultra rank-sum. II library prep kit (New England Biolabs, Ipswich, MA) protocol according to previously described detailed methods [12]. AMR gene identification was carried out Results using ARG-ANNOT [23] as for the primary analyses, Cohort features and genes that were detected at a dpM of > 0.1 were Seventy subjects with LRTI and 18 with no evi- assessed for enrichment compared to DNA-seq alone. dence of LRTI were identified based on inclusion and exclusion criteria (Fig.  1, Additional File 1: Nanopore sequencing Table  S1). Primar y analyses were performed for 27 FLASH-enriched DNA libraries were quantified and patients , secondar y analyses for all subjects . Clini- 200–800  ng of DNA input was used for Nanopore 1D cal AST results were returned a median of 74  h library preparation (protocol SQK-LSK109, Oxford following sample collection (95% confidence inter- Nanopore, UK). Individual sample libraries were loaded val (CI) 49–115  h, (Additional File 5: Table  S3)). into a single flow cell of a GridION instrument, and Twenty-seven subjects with culture-confirmed sequencing reads were base called in real-time mode in bacterial LRTI, representing 32 pathogens with MINKNOW. The SURPIrt pipeline running in -a mode clinical AST data performed on ≥ 2 drugs, were was utilized to identify AMR genes every 100,000– identified and assessed in the primary analysis 200,000 reads as previously described [27, 28]. (Additional File 2: Table  S2). For secondary anal- yses, 43 additional patients with clinically adju- Mitigation of background contaminants dicated LRTI and 18 patients with no evidence of To minimize inaccurate taxonomic assignments due LRTI were assessed. to environmental contaminants, we processed nega- tive water controls with each group of samples that Metagenomic sequencing, pathogen, and AMR gene underwent nucleic acid extraction, and included these, detection as well as positive control clinical samples, with each A mean of 4.3 × 10 (interquartile range (IQR) 7 7 sequencing run. We directly subtracted alignments to 1.9–4.4 × 10 ) DNA-seq reads and 6.9 × 10 (IQR those taxa in water control samples detected by both 4.8–8.3 × 10 ) RNA-seq reads were generated from res- RNA-seq and DNA-seq analyses from the raw reads per piratory samples. In the primary AMR analysis group, we million (rpm) values in all samples [6]. To account for used a previously validated [6] metagenomic rules-based selective amplification bias of contaminants in water model (RBM) to identify bacterial respiratory pathogens controls resulting from PCR amplification of metagen - that were disproportionately abundant as compared to omic libraries to a fixed standard concentration across the rest of the lung microbiome. The RBM identified 26 all samples, prior to direct subtraction, we scaled taxa of 32 (81%) of the culture-confirmed bacterial patho - rpms in the water controls to the median percent gens from the primary analysis. Four (67%) of the missed microbial reads present across all samples as previously pathogens were present in the context of polymicro- described [6]. To address environmental contaminants bial cultures, and one (17%) was identified as a different in AMR gene analyses, resistance alleles detected in streptococcal species (Additional File 2: Table S2). A total water controls at a depth > 1 were excluded. of 138 and 234 acquired AMR genes were identified by RNA-seq and DNA-seq, respectively (Additional File 6: Study outcomes Table S4). With respect to AMR gene classes, beta lactam The pr imar y out c ome w a s p er f or manc e of mN GS resistance genes were most common (81/372 total genes, for predicting phenotypic AST. Secondary 35%). S erpa et al. Genome Medicine (2022) 14:74 Page 7 of 12 Fig. 2 A AMR genes detected in the lower respiratory microbiome of critically ill patients. Composite results of DNA and RNA mNGS. AMR genes are listed in rows and are grouped by antimicrobial class. Each column represents a patient respiratory sample and is grouped by LRTI status. B AMR gene burden in the respiratory tract, measured by averaging sequencing depth across the AMR allele per million reads sequenced (dpM) in the respiratory microbiome did not differ between LRTI-positive patients and those with non-infectious acute respiratory illnesses. C The burden of AMR genes detected in the lower respiratory tract microbiome was greater in patients with hospital-onset LRTI versus those with either community-onset LRTI or no evidence of LRTI. Legend: depth = average sequencing depth across each AMR gene allele normalized per million reads sequenced. Legend: Bla = beta lactam; AGly = aminoglycoside; Fos = Fosfomycin; Flq = fluoroquinolone; Gly = glycopeptide; Mac/ Lin/Str = macrolide, lincosamide, streptogramin; Phe = phenicol; Tet = tetracycline; Tmp-Sul = trimethoprim/sulfamethoxazole; depth = average sequencing depth across each AMR gene allele normalized per million reads sequenced. The horizontal bars in panels B and C indicate mean values Comparison of mNGS versus phenotypic antimicrobial 47–87%), specificity of 95% (CI 85–99%), and an accuracy susceptibility testing of 87% (CI 78–94%) (Table  1). This equated to a VME We assessed the performance of mNGS for predicting rate of 30% and a ME rate of 5%. For gram-negative path- resistance to clinical guideline-recommended antimicro- ogens, a combination of DNA-seq and RNA-seq yielded bials used for complicated gram-negative (n = 8 drugs) a sensitivity of 100% (CI 87–100%), specificity of 64% (CI and gram-positive (n = 6 drugs) infections. AMR genes 48–78%), and accuracy of 78% (CI 67–87%) (Table  1). unrelated to the culture-confirmed bacterial pathogen This equated to a VME rate of 0% and a ME rate of 36%. were identified through the resistome ontology annota - We also assessed the performance of RNA-seq and tions in CARD [24] and excluded from this analysis. Sen- DNA-seq performed independently (Table  1, Additional sitivity and specificity compared to a reference standard File 7: Table S5). RNA-seq performed with a sensitivity of of culture-based AST varied by pathogen, drug, patient, 52% (CI 31–73%), specificity of 100% (CI 94–100%), and and nucleic acid type sequenced (Table 1, Additional File accuracy of 86% (CI 76–93%) for gram-positive patho- 7: Table S5). For gram-positive pathogens, a combination gens, and a sensitivity of 100% (CI 89–100%), specificity of DNA-seq and RNA-seq yielded a sensitivity of 70% (CI of 64% (CI 48–78%), and accuracy of 79% (CI 68–88%) Serpa et al. Genome Medicine (2022) 14:74 Page 8 of 12 for gram-negative pathogens. DNA-seq performed with genes that were associated with the culture-confirmed a sensitivity of 39% (CI 20–61%), specificity of 95% (CI pathogen and resistance phenotype, but missed by DNA- 85–99%), and accuracy of 78% (CI 67–87%) for gram- seq alone, including mecA in two patients with methicil- positive pathogens, and a sensitivity of 58% (CI 39–75%), lin-resistant Staphylococcus aureus LRTI. FLASH also specificity of 67% (CI 50–80%), and accuracy of 63% (CI resulted in the detection of AMR genes unrelated to the 51–74%) for gram-negative pathogens. culture-confirmed pathogens in five (50%) of patients In two of seven cases with genotype to phenotype false- (Additional File 8: Table S6). positive (ME) predictions, mNGS identified AMR genes We subsequently assessed the potential for rapid unrelated to the culture-confirmed microbe but related AMR gene detection using FLASH combined with to resistant pathogens that would be cultured several an Oxford nanopore sequencing platform, which days later in the context of ventilator-associated pneu- affords real-time data generation (Fig.  3B). All AMR monia (VAP). These included SST-1 from a patient who genes identified by Illumina (median sequenc - 8 7 8 developed Serratia marcescens VAP (patient 213) 4  days ing depth of 1.20 × 10 [IQR 5.41 × 10 –1.36 × 10 ]) later, and mecA from a patient who developed Staphylo- were also identified by FLASH-nanopore sequencing coccus aureus VAP 7 days later (patient 232) (Additional (median sequencing depth of 1.19 × 10 reads [IQR 6 6 File 6: Table S4). 1.02 × 10 –1.46 × 10 ]) (Additional File 9: Table  S7). AMR gene targets could be identified within 10 min of real-time nanopore sequencing, suggesting a potential Association between LRTI positivity, hospital exposure, turnaround time of less than 6  h for a single sample and AMR genes in the respiratory microbiome (Fig. 3C). Assessment of the lower respiratory resistome using both DNA and RNA mNGS revealed a diversity of AMR genes in both LRTI-positive and negative patients (Fig.  2A, Discussion Additional File 6: Table  S4). AMR gene burden did not Antimicrobial resistance has emerged as one of the differ based on LRTI status (p = 0.28) (Fig.  2B). Sub- most pressing issues facing human health, and effective jects with hospital-onset (≥ 48  h after admission) LRTI treatment of complicated infections increasingly neces- had a greater burden of AMR genes in their respiratory sitates early and accurate assessment of microbial drug microbiome compared to those with community-onset resistance. Our study builds on prior respiratory mNGS LRTI (p = 0.00030) or those without LRTI (p = 0.0024) studies [6, 8–12] by demonstrating that RNA and DNA (Fig. 2C). mNGS can enable culture-independent prediction of AMR in critically ill patients with bacterial LRTI, with variable performance across pathogens and antimicro- Targeted enrichment and rapid detection of AMR genes bials. While false-negative susceptibility predictions for using Cas9 and nanopore sequencing gram-negative pathogens were not observed, we found We validated the utility of a recently described pro- a VME of 30% for gram-positive pathogens, suggesting grammable CRISPR/Cas9-based method called FLASH mNGS has a role for complementing, rather than replac- (Finding Low Abundance Sequences by Hybridization) ing, current standard of care culture-based approaches. to enrich for low abundance AMR genes [12] by study- Prior work has demonstrated the utility of mNGS in ing 10 patients from the primary analysis (Additional cases of culture-negative LRTI, which represent more File 8: Table  S6). The FLASH + DNA mNGS library than half of all pneumonia cases [4, 6, 8, 10]. Our results prep protocol added 2.5  h to the standard 3-h NEBNext suggest that mNGS may also have potential for predict- DNA-seq workflow [29] and could enrich detection of ing AST in culture-negative LRTI, where detection of AMR genes associated with the culture-confirmed path - an AMR gene could inform the need for treatment of an ogen > 2500 × compared to DNA-seq alone (Fig.  3A). In occult resistant organism. Further, our findings support four (40%) patients, FLASH enabled detection of AMR recent observations [10] that mNGS may have utility for (See figure on next page.) Fig. 3 A FLASH (Finding Low Abundance Sequences by Hybridization) CRISPR/Cas9 targeted Illumina sequencing enriched the detection of culture-confirmed bacterial LRTI pathogen AMR alleles by 46 × to > 2500 × versus DNA-seq alone. B Workflow diagram for FLASH targeted enrichment coupled with nanopore sequencing. Time estimates provided for a single sample. C Real-time detection of AMR genes by FLASH targeted nanopore sequencing was achieved within 10 min following mNGS library preparation. Data from two representative Staphylococcus aureus LRTI cases are highlighted. Case 212 (left panel) highlights a case where detection of BlaZ and MsrA/ErmA genes correlated with phenotypically determined penicillin and macrolide/lincosamide resistance, respectively. Case 288 (right panel) highlights a case where detection of MecA, BlaZ, and MsrA correlated with phenotypically confirmed methicillin, penicillin, and macrolide resistance, respectively S erpa et al. Genome Medicine (2022) 14:74 Page 9 of 12 - FLASH + FLASH -1 -2 -3 289 314 407 409 268 298 212 288 350 382 300 10000 25 BlaZ MecA Erm BlaZ MsrA 0 0 10 30 90 10 30 90 time (min) time (min) Fig. 3 (See legend on previous page.) BlaZ reads AMR gene dpM BlaZ ErmA BlaZ MsrA MecA BlaZ ErmC BlaZ Erm reads BlaZ AMR gene reads ErmA MecA OXA-50 ACT-MIR SRT-SST SRT-SST Serpa et al. Genome Medicine (2022) 14:74 Page 10 of 12 early identification of future secondary respiratory infec - LRTI and that it systematically assessed performance tions. For instance, patient 232, who was admitted for for multiple classes of antimicrobials against a clinical severe Klebsiella pneumoniae LRTI, developed MRSA reference standard. Further, we provide the first culture- ventilator-associated pneumonia (VAP) eight days later independent assessment of healthcare exposure and while undergoing treatment with aztreonam, an anti- resistance gene burden in the respiratory tract, and a biotic lacking MRSA coverage. mNGS analysis of a TA novel demonstration of Cas9 targeted enrichment cou- specimen obtained 7 days before VAP onset revealed the pled with rapid nanopore sequencing. Our study also mecA gene, providing an early indication of the subse- has several limitations, including sample size, spectrum quent AMR VAP pathogen. of antimicrobial classes assessed, spectrum of patho- Public health surveillance is essential for pandemic gens assessed, and the need for independent validation preparedness, understanding trends in AMR, and pre- of findings. Future work in a larger, prospective cohort venting outbreaks of resistant organisms. Our findings with a greater diversity of bacterial pathogens and resist- demonstrate the feasibility of mNGS for epidemiologi- ance mechanisms can address these limitations. In addi- cal surveillance of AMR and highlight an association tion, a randomized clinical trial will be needed to assess between hospital exposure and AMR gene burden in the the potential impact of mNGS on time to appropriate lower respiratory tract microbiome. Importantly, our antimicrobial treatment, antimicrobial stewardship, and analysis included culture-negative LRTI cases and adds LRTI outcomes. Lastly, genotype to phenotype predic- to prior literature demonstrating a similar association in tion remains imperfect, even for cultured isolates [7, 35]. cases of culture-confirmed bacterial pneumonia [30]. As with other molecular testing modalities for AMR, Nanopore sequencing has proven useful for microbial this should be recognized when considering the utility detection from respiratory samples with high pathogen and clinical applicability of mNGS for AMR phenotypic abundance [8, 10, 31]; however, basecalling accuracy and prediction. sensitivity challenges have historically limited its capacity for detecting underrepresented sequences in metagen- Conclusions omic datasets [32]. Targeted enrichment can overcome In summary, we characterize the utility of mNGS for pre- this through AMR signal amplification, and consistent dicting AMR in bacterial LRTI and demonstrate proof of with this, we found high concordance for AMR gene concept for both epidemiological AMR surveillance and detection between nanopore and Illumina samples that rapid resistance gene detection using Cas9 and nanopore underwent FLASH Cas9 targeted enrichment. Our sequencing. results suggest that this method could also potentially augment AMR gene detection and resistance prediction Supplementary Information if coupled with established nanopore mNGS workflows The online version contains supplementary material available at https:// doi. for detecting respiratory pathogens [8–10]. org/ 10. 1186/ s13073- 022- 01072-4. Rapid detection of AMR is essential in critically ill Additional file 1: Table S1. Clinical and demographic features of cohort. patients with severe bacterial infections given that time to appropriate antimicrobials correlates with mortality [33]. Additional file 2: Table S2. mNGS detection of respiratory patho- gens compared to a reference standard of clinical microbiological FLASH adds 2.5  h to standard DNA library preparation testing. workflow and, when coupled with real-time detection of Additional file 3: Dataset 1. AMR gene database utilized for analyses. AMR genes from mNGS libraries, could potentially allow Additional file 4: Dataset 2. AMR genes and guide RNAs for FLASH for sample to answer in under 6 h, a significant time sav - targeted mNGS. AMR genes and guide RNAs for FLASH targeted mNGS. ings compared to the ≥ 24 h required for Illumina proto- A) AMR genes targeted by FLASH. B) AMR alleles and guide RNA targets. Legend: Bla = beta lactam; MLS = macrolide, Tmp/Sul = trimethoprim/ cols [34]. In our cohort, clinical AST required an average sulfamethoxazole. of 74 h, suggesting that Cas9-targeted nanopore sequenc- Additional file 5: Table S3. Time to return of clinical antimicrobial suscep - ing may be a promising future approach for more rapidly tibility testing results for culture-confirmed bacterial pathogens. identifying patients with resistant infections. We found Additional file 6: Table S4. AMR genes detected by mNGS. that FLASH also enriched for AMR genes unrelated to Additional file 7: Table S5. Performance of RNA-seq, DNA-seq and the culture-confirmed pathogen, presumably derived combined RNA-seq + DNA-seq for AMR prediction compared to clinical from the lung microbiome. Improved methods for anno- antimicrobial susceptibility testing. tating the species-specificity of detected AMR genes may Additional file 8: Table S6. Pathogen AMR genes detected by help address this issue, which otherwise could lead to FLASH-mNGS. increased ME due to false-positive results. Additional file 9: Table S7. AMR genes detected by FLASH coupled with Nanopore mNGS. Strengths of this study include that it is the largest to assess the performance of mNGS AMR prediction in S erpa et al. Genome Medicine (2022) 14:74 Page 11 of 12 spectrum antibiotic treatment in patients with community acquired Acknowledgements pneumonia: a prospective randomised study. Thorax. 2005;60:672–8. Not applicable. 4. Jain S, Self WH, Wunderink RG, et al. Community-acquired pneu- monia requiring hospitalization among U.S. adults. N Engl J Med. Authors’ contributions 2015;373:415–27. PHS, XD: collection and assembly of data, data analysis and interpretation, 5. Leffler DA, Lamont JT. Clostridium difficile infection. N Engl J Med. manuscript editing. MA, KM, SC, MF, NN: collection and assembly of data, 2015;372:1539–48. manuscript editing. RG, TD: collection and assembly of data. EC, AL.: methods 6. Langelier C, Kalantar KL, Moazed F, et al. Integrating host response and development and manuscript editing. AM, KK: data analysis and interpretation. unbiased microbe detection for lower respiratory tract infection diagno- SD, SM: manuscript editing and data interpretation. CCa, CCh, JD: conception sis in critically ill adults. Proc Natl Acad Sci USA. 2018;115(52):E12353–62 and design, manuscript editing, funding. CL: conception and design, data (201809700). analysis and interpretation, manuscript writing, funding. All authors read and 7. Mahfouz N, Ferreira I, Beisken S, von Haeseler A, Posch AE. Large-scale approved the final manuscript. assessment of antimicrobial resistance marker databases for genetic phenotype prediction: a systematic review. J Antimicrob Chemother. Funding 2020;75:3099–108. NIH R35 HL140026 (CSC), NHLBI K23HL138461-01A1 (CL), Chan Zuckerberg 8. Charalampous T, Kay GL, Richardson H, et al. Nanopore metagenomics Biohub (CL, JLD). The funders of the study had no role in study design, data enables rapid clinical diagnosis of bacterial lower respiratory infection. collection, data analysis, data interpretation, or writing of the report. Nat Biotechnol. 2019;37:783–92. 9. Yang L, Haidar G, Zia H, et al. Metagenomic identification of severe pneu- Availability of data and materials monia pathogens in mechanically-ventilated patients: a feasibility and Raw microbial sequences are available via NCBI BioProject PRJNA450137 clinical validity study. Respir Res. 2019;20:265. [36], https:// www. ncbi. nlm. nih. gov/ biopr oject/? term = PRJNA450137, and 10. Charalampous T, Alcolea-Medina A, Snell LB, et al. Evaluating the poten- BioProject PRJNA635133 [37], https:// www. ncbi. nlm. nih. gov/ biopr oject/? tial for respiratory metagenomics to improve treatment of secondary term = PRJNA635133. FLASHit [26] and SURPIrt [28] have been deposited on infection and detection of nosocomial transmission on expanded COVID- Github and are available for download at: https:// github. com/ czbio hub/ flash 19 intensive care units. Genome Med. 2021;13:182. and https:// github. com/ chiul ab/ SURPI- plus- dist, respectively. 11. Chao L, Li J, Zhang Y, Pu H, Yan X. Application of next generation sequencing-based rapid detection platform for microbiological diagnosis and drug resistance prediction in acute lower respiratory infection. Ann Declarations Transl Med. 2020;8:1644. 12. Quan J, Langelier C, Kuchta A, et al. FLASH: a next-generation CRISPR Ethics approval and consent to participate diagnostic for multiplexed detection of antimicrobial resistance Patients were enrolled via a previously described observational cohort study sequences. Nucleic Acids Res 2019; published online May 22. https:// doi. [6, 38] which was approved by the UCSF Institutional Review Board (IRB), org/ 10. 1093/ nar/ gkz418. which granted a waiver of initial consent for respiratory sampling. Informed 13. United States Centers for Disease Control and Prevention. CDC/NHSN consent was subsequently obtained from patients or their surrogates for surveillance definitions for specific types of infections. 2021; published continued study participation, following the previously detailed IRB protocol online Jan. https:// www. cdc. gov/ nhsn/ pdfs/ pscma nual/ pcsma nual_ #17–24056 [38]. This research conformed to the principles of the Helsinki curre nt. pdf. Declaration. 14. Kalantar KL, Carvalho T, de Bourcy CFA, et al. IDseq—An open source cloud-based pipeline and analysis service for metagenomic pathogen Consent for publication detection and monitoring. Gigascience 2020; 9. https:// doi. org/ 10. 1093/ Consent to publish patient-related information was obtained as part of gigas cience/ giaa1 11. informed consent for study participation as described above. 15. Dobin A, Davis CA, Schlesinger F, et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics. 2013;29:15–21. Competing interests 16. Ruby JG, Bellare P, Derisi JL. PRICE: software for the targeted assembly The authors declare that they have no competing interests. of components of (Meta) genomic sequence data. G3 (Bethesda). 2013;3:865–80. Author details 1 17. Tsitsiklis A, Osborne CM, Kamm J, et al. Lower respiratory tract infections Division of Infectious Diseases, Department of Medicine, University of Cali- 2 in children requiring mechanical ventilation: a multicentre prospective fornia, San Francisco, San Francisco, CA, USA. Chan Zuckerberg Biohub, San 3 surveillance study incorporating airway metagenomics. Lancet Microbe. Francisco, CA, USA. Department of Laboratory Medicine, University of Califor- 4 2022;3:e284–93. nia, San Francisco, CA, USA. Department of Microbiology and Immunology, 5 18. Jain S, Self WH, Wunderink RG, et al. Community-acquired pneu- University of California, San Francisco, CA, USA. Division of Pulmonary, Critical monia requiring hospitalization among U.S. adults. N Engl J Med. Care, Allergy and Sleep Medicine, Department of Medicine, University of Cali- 6 2015;373:415–27. fornia, San Francisco, CA, USA. Chan Zuckerberg Initiative, San Francisco, CA, 7 19. Infection in Organ Transplantation - Fishman - 2017 - American Journal of USA. Department of Biochemistry and Biophysics, University of California, San Transplantation - Wiley Online Library. https:// doi. org/ 10. 1111/ ajt. 14208 Francisco, CA, USA. (Accessed 21 Jul 2021). 20. Magill SS, Edwards JR, Bamberg W, et al. Multistate point-preva- Received: 6 February 2022 Accepted: 15 June 2022 lence survey of health care-associated infections. N Engl J Med. 2014;370:1198–208. 21. Kalil AC, Metersky ML, Klompas M, et al. Management of adults with hospital-acquired and ventilator-associated pneumonia: 2016 Clinical Practice Guidelines by the Infectious Diseases Society of America and the References American Thoracic Society. Clin Infect Dis. 2016;63:e61-111. 1. O’Neill J. Tackling drug-resistant infections globally: final report and rec- 22. Mandell LA, Wunderink RG, Anzueto A, et al. Infectious Diseases Society ommendations. London, UK May 16, 2016 2016. 2016; published online of America/American Thoracic Society consensus guidelines on the May 16. http:// amr- review. org/ Publi catio ns. management of community-acquired pneumonia in adults. Clin Infect 2. Zaas AK, Garner BH, Tsalik EL, Burke T, Woods CW, Ginsburg GS. The cur- Dis. 2007;44(Suppl 2):S27-72. rent epidemiology and clinical decisions surrounding acute respiratory 23. Inouye M, Dashnow H, Raven L-A, et al. SRST2: Rapid genomic surveil- infections. Trends Mol Med. 2014;20:579–88. lance for public health and hospital microbiology labs. Genome Med. 3. van der Eerden MM, Vlaspolder F, de Graaff CS, et al. Comparison 2014;6:90. between pathogen directed antibiotic treatment and empirical broad Serpa et al. Genome Medicine (2022) 14:74 Page 12 of 12 24. Jia B, Raphenya AR, Alcock B, et al. CARD 2017: expansion and model- centric curation of the comprehensive antibiotic resistance database. Nucleic Acids Res. 2017;45:D566–73. 25. CLSI. CLSI M100-ED31:2021 Performance Standards for Antimicrobial Susceptibility Testing, 31st Edition. 2021; published online March. http:// em100. edapt ivedo cs. net/ dashb oard. aspx. 26. Quan J, Langelier CR, Batson J, Crawford ED. FLASHit. Github. https:// github. com/ czbio hub/ flash 2020. 27. Deng X, Achari A, Federman S, et al. Metagenomic sequencing with spiked primer enrichment for viral diagnostics and genomic surveillance. Nature Microbiology 2020; published online Jan 13. https:// doi. org/ 10. 1038/ s41564- 019- 0637-9. 28. Chiu CY, Stryke D. SURPI+, a rapid computational pipeline for compre- hensive identification of pathogens from clinical metagenomic sequence data. Github. https:// github. com/ chiul ab/ SURPI- plus- dist 2019. 29. New England Biolabs. Improved library preparation with the NEBNext UltraTM II DNA Library Prep Kit for Illumina. https:// www. neb- online. de/ wp- conte nt/ uploa ds/ 2020/ 03/ Impro ved- libra ry- prepa ration- with- the- NEBNe xt- Ultra- II- DNA- Libra ry- Prep- Kit- for- Illum ina- E7645. pdf (Accessed 9 Dec 2021). 30. Kalil AC, Metersky ML, Klompas M, et al. Management of adults with hospital-acquired and ventilator-associated pneumonia: 2016 Clinical Practice Guidelines by the Infectious Diseases Society of America and the American Thoracic Society. Clin Infect Dis. 2016;63:e61-111. 31. Greninger AL, Naccache SN, Federman S, et al. Rapid metagenomic identification of viral pathogens in clinical samples by real-time nanopore sequencing analysis. Genome Med. 2015;7:99. https:// doi. org/ 10. 1186/ s13073- 015- 0220-9. 32. Rang FJ, Kloosterman WP, de Ridder J. From squiggle to basepair: compu- tational approaches for improving nanopore sequencing read accuracy. Genome Biol. 2018;19:90. https:// doi. org/ 10. 1186/ s13059- 018- 1462-9. 33. Lee C-C, Lee C-H, Yang C-Y, Hsieh C-C, Tang H-J, Ko W-C. Beneficial effects of early empirical administration of appropriate antimicrobials on survival and defervescence in adults with community-onset bacteremia. Crit Care. 2019;23:363. 34. Illumina, Inc. Run time estimates for each sequencing step on the Illumina sequencing platforms. 2020; published online May 15. https:// emea. suppo rt. illum ina. com/ bulle tins/ 2017/ 02/ run- time- estim ates- for- each- seque ncing- step- on- illum ina- seque nci. html (Accessed 9 Dec 2021). 35. Yee R, Dien Bard J, Simner PJ. The genotype-to-phenotype dilemma: how should laboratories approach discordant susceptibility results? J Clin Microbiol. 2021;59:e00138-e220. 36. Langelier CR, Kalantar KL. Combined host and microbe NGS for lower res- piratory tract infection diagnosis in critically ill adults microbial sequence reads. BioProject PRJNA450137, Sequence Read Archive. 2018. https:// www. ncbi. nlm. nih. gov/ biopr oject/? term=% 20PRJ NA450 137. 37. Langelier CR, Hayakawa Serpa P. CAS9-mNGS detects antimicrobial resist- ant pathogens. BioProject PRJNA635133, Sequence Read Archive. 2020. https:// www. ncbi. nlm. nih. gov/ biopr oject/? term= PRJNA 635133. 38. Sarma A, Christenson SA, Byrne A, et al. Tracheal aspirate RNA sequencing identifies distinct immunological features of COVID-19 ARDS. Nat Com- mun. 2021;12:5152. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in pub- lished maps and institutional affiliations. Re Read ady y to to submit y submit your our re researc search h ? Choose BMC and benefit fr ? Choose BMC and benefit from om: : fast, convenient online submission thorough peer review by experienced researchers in your field rapid publication on acceptance support for research data, including large and complex data types • gold Open Access which fosters wider collaboration and increased citations maximum visibility for your research: over 100M website views per year At BMC, research is always in progress. Learn more biomedcentral.com/submissions

Journal

Genome Medicine – Springer Journals

Published: Jul 12, 2022

There are no references for this article.