Theodore Pak

How Next-Generation Sequencing and Multiscale Data Analysis Will Transform Infectious Disease Management

We should integrate next-generation sequencing data from pathogen specimens with phenotypes from electronic medical records to create quantitative, predictive models of infectious disease. Precision infection control and antimicrobial interventions can address urgent global problems, including healthcare-associated infections and multidrug resistance.

Whole-genome sequencing identifies emergence of a quinolone resistance mutation in a case of Stenotrophomonas maltophilia bacteremia.

ABSTRACT Whole-genome sequences for Stenotrophomonas maltophilia serial isolates from a bacteremic patient before and after development of levofloxacin resistance were assembled de novo and differed by one single-nucleotide variant in smeT, a repressor for multidrug efflux operon smeDEF. Along with sequenced isolates from five contemporaneous cases, they displayed considerable diversity compared against all published complete genomes. Whole-genome sequencing and complete assembly can conclusively identify resistance mechanisms emerging in S. maltophilia strains during clinical therapy.

Genomic confirmation of vancomycin-resistant Enterococcus transmission from deceased donor to liver transplant recipient

In a liver transplant recipient with vancomycin-resistant Enterococcus (VRE) surgical site and bloodstream infection, a combination of pulsed-field gel electrophoresis, multilocus sequence typing, and whole genome sequencing identified that donor and recipient VRE isolates were highly similar when compared to time-matched hospital isolates. Comparison of de novo assembled isolate genomes was highly suggestive of transplant transmission rather than hospital-acquired transmission and also identified subtle internal rearrangements between donor and recipient missed by other genomic approaches. Given the improved resolution, whole-genome assembly of pathogen genomes is likely to become an essential tool for investigation of potential organ transplant transmissions.

High-dimensional CyTOF analysis of dengue virus–infected human DCs reveals distinct viral signatures

Dengue virus (DENV) is the most prevalent mosquito-borne virus causing human disease. Of the 4 DENV serotypes, epidemiological data suggest that DENV-2 secondary infections are associated with more severe disease than DENV-4 infections. Mass cytometry by time-of-flight (CyTOF) was used to dissect immune changes induced by DENV-2 and DENV-4 in human DCs, the initial targets of primary infections that likely affect infection outcomes. Strikingly, DENV-4 replication peaked earlier and promoted stronger innate immune responses, with increased expression of DC activation and migration markers and increased cytokine production, compared with DENV-2. In addition, infected DCs produced higher levels of inflammatory cytokines compared with bystander DCs, which mainly produced IFN-induced cytokines. These high-dimensional analyses during DENV-2 and DENV-4 infections revealed distinct viral signatures marked by different replication strategies and antiviral innate immune induction in DCs, which may result in different viral fitness, transmission, and pathogenesis.

Genetic Basis of Emerging Vancomycin, Linezolid, and Daptomycin Heteroresistance in a Case of Persistent Enterococcus faecium Bacteremia

ABSTRACT Whole-genome sequencing was used to examine a persistent Enterococcus faecium bacteremia that acquired heteroresistance to three antibiotics in response to prolonged multidrug therapy. A comparison of the complete genomes before and after each change revealed the emergence of known resistance determinants for vancomycin and linezolid and suggested that a novel mutation in fabF, encoding a fatty acid synthase, was responsible for daptomycin nonsusceptibility. Plasmid recombination contributed to the progressive loss of vancomycin resistance after withdrawal of the drug.

Reply to Lesho

TO THE EDITOR—We welcome the letter by Lesho [1] and the points raised therein, and we generally agree that next-generation sequencing (NGS) is currently more likely to augment—not replace—existing methods for profiling multidrug-resistant (MDR) organisms in clinical microbiology laboratories. Although technical hurdles to widespread deployment still remain, we predict that these can be resolved within only a few years, according to current trends. First, quality control for methods that predict drug resistance from NGS data will ultimately be assessed by sensitivity and specificity comparisons to gold-standard methods; this has brought NGS profiling of MDR human immunodeficiency virus close to mainstream practice [2] and shows rising success for MDR Staphylococcus aureus and Mycobacterium tuberculosis, with the most recent method for S. aureus achieving >99% sensitivity and specificity [3]. These methods can set thresholds for base-call quality, read depth, and alignment quality metrics on the chosen sequencing platform. Longread sequencing (which obviates the limitations of short reads) has only become more widely available with the release of new, cheaper technologies such as Oxford Nanopore and the PacBio Sequel. Meanwhile, competition among secure cloud computing providers continues to lower the cost of bandwidth and storage capacity for sequencing data, and it stands to reason that microbiologists will profit handsomely from a technology ecosystem often designed to permit the scale of human sequencing data, as bacterial sequencing analyses are hundreds to thousands of times smaller than their human equivalents (typical completed assemblies, annotations, and variant call files are within the megabyte scale). Finally, we hope that efforts such as Mil-OSS (http://mil-oss.org/) and Forge.mil (http:// forge.mil) indicate growing adoption of open-source software and culture by the military, which will eventually be necessary to harness the full benefit of continuing developments in scientific software. Note Potential conflict of interest. Both authors: No reported conflicts. Both authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

Epigenomic landscape of the human pathogen Clostridium difficile

Abstract Clostridioides difficile is a leading cause of health care-associated infections. Although significant progress has been made in the understanding of its genome, the epigenome of C. difficile and its functional impact has not been systematically explored. Here, we performed the first comprehensive DNA methylome analysis of C. difficile using 36 human isolates and observed great epigenomic diversity. We discovered an orphan DNA methyltransferase with a well-defined specificity whose corresponding gene is highly conserved across our dataset and in all ~300 global C. difficile genomes examined. Inactivation of the methyltransferase gene negatively impacted sporulation, a key step in C. difficile disease transmission, consistently supported by multi-omics data, genetic experiments, and a mouse colonization model. Further experimental and transcriptomic analysis also suggested that epigenetic regulation is associated with cell length, biofilm formation, and host colonization. These findings open up a new epigenetic dimension to characterize medically relevant biological processes in this critical pathogen. This work also provides a set of methods for comparative epigenomics and integrative analysis, which we expect to be broadly applicable to bacterial epigenomics studies.Clostridium difficile is a leading cause of health care–associated infections. Although significant progress has been made in the understanding of its genome, the epigenome of C. difficile and its functional impact has not been explored. Here, we performed the first DNA methylome analysis of C. difficile using 36 human isolates and observed great epigenomic diversity. Strikingly, we discovered a DNA methyltransferase with a well-defined specificity, highly conserved across our dataset and in all the ∼300 global C. difficile genomes we further examined. Inactivation of the methyltransferase negatively impacted sporulation, a key step in C. difficile transmission, consistently supported by multi-omics data and genetic experiments and a mouse infection model. Transcriptomic analysis also suggested that epigenetic regulation is associated with host colonization and biofilm formation. The epigenomic landscape also allowed an integrative analysis of multiple defense systems with respect to their roles in host defense and in regulating gene flux in C. difficile. These findings open up a new epigenetic dimension to characterize medically relevant biological processes in this critical pathogen. This work also provides a set of methods for comparative epigenomics and integrative analysis, which we expect to be broadly applicable to bacterial epigenomics studies.

Comprehensive innate immune profiling of chikungunya virus infection in pediatric cases

Chikungunya virus (CHIKV) is a mosquito‐borne alphavirus that causes global epidemics of debilitating disease worldwide. To gain functional insight into the host cellular genes required for virus infection, we performed whole‐blood RNA‐seq, 37‐plex mass cytometry of peripheral blood mononuclear cells (PBMCs), and serum cytokine measurements of acute‐ and convalescent‐phase samples obtained from 42 children naturally infected with CHIKV. Semi‐supervised classification and clustering of single‐cell events into 57 sub‐communities of canonical leukocyte phenotypes revealed a monocyte‐driven response to acute infection, with the greatest expansions in “intermediate” CD14++CD16+ monocytes and an activated subpopulation of CD14+ monocytes. Increases in acute‐phase CHIKV envelope protein E2 expression were highest for monocytes and dendritic cells. Serum cytokine measurements confirmed significant acute‐phase upregulation of monocyte chemoattractants. Distinct transcriptomic signatures were associated with infection timepoint, as well as convalescent‐phase anti‐CHIKV antibody titer, acute‐phase viremia, and symptom severity. We present a multiscale network that summarizes all observed modulations across cellular and transcriptomic levels and their interactions with clinical outcomes, providing a uniquely global view of the biomolecular landscape of human CHIKV infection.

Estimating Local Costs Associated With Clostridium difficile Infection Using Machine Learning and Electronic Medical Records

BACKGROUND Reported per-patient costs of Clostridium difficile infection (CDI) vary by 2 orders of magnitude among different hospitals, implying that infection control officers need precise, local analyses to guide rational decision making between interventions. OBJECTIVE We sought to comprehensively estimate changes in length of stay (LOS) attributable to CDI at a single urban tertiary-care facility using only data automatically extractable from the electronic medical record (EMR). METHODS We performed a retrospective cohort study of 171,938 visits spanning a 7-year period. In total, 23,968 variables were extracted from EMR data recorded within 24 hours of admission to train elastic-net regularized logistic regression models for propensity score matching. To address time-dependent bias (reverse causation), we separately stratified comparisons by time of infection, and we fit multistate models. RESULTS The estimated difference in median LOS for propensity-matched cohorts varied from 3.1 days (95% CI, 2.2-3.9) to 10.1 days (95% CI, 7.3-12.2) depending on the case definition; however, dependency of the estimate on time to infection was observed. Stratification by time to first positive toxin assay, excluding probable community-acquired infections, showed a minimum excess LOS of 3.1 days (95% CI, 1.7-4.4). Under the same case definition, the multistate model averaged an excess LOS of 3.3 days (95% CI, 2.6-4.0). CONCLUSIONS In this study, 2 independent time-to-infection adjusted methods converged on similar excess LOS estimates. Changes in LOS can be extrapolated to marginal dollar costs by multiplying by average costs of an inpatient day. Infection control officers can leverage automatically extractable EMR data to estimate costs of CDI at their own institutions. Infect Control Hosp Epidemiol. 2017;38:1478-1486.