2063. Comparison of Pulsed-Field Gel Electrophoresis and Whole Genome Sequencing in Clostridium difficile Typing
Session: Poster Abstract Session: Clostridium difficile: Epidemiology
Saturday, October 29, 2016
Room: Poster Hall
Posters
  • Poster 201610 final.pdf (349.2 kB)
  • Background:

    Pulsed-field gel electrophoresis (PFGE) is a common typing method for Clostridium difficile (C. difficile), but whole genome sequencing (WGS) may become a new gold standard. We compared the two methods by examining matched PFGE and WGS data for a set of C. difficile isolates.

    Methods:

    467 C. difficile isolates, collected in a cohort study in 2006-2007, underwent PFGE at the time of study and WGS in 2015. Strain relatedness by PFGE was identified as indistinguishable, closely related, possibly related, or different using Tenover’s criteria. Numbers of single nucleotide variants (SNV) across all possible pairs were compared with PFGE categories. To assess ability to recognize clonality, isolates were considered the same type by WGS if they had 0-2 SNV, and by PFGE if indistinguishable, closely related, or possibly related. Using these definitions, adjusted Wallace coefficient was calculated to assess congruence between the methods. Simpson’s index was calculated to assess discriminatory power.

    Results:

    The average number of SNV between pairs increased across increasingly dissimilar PFGE categories, but was large even for “indistinguishable” isolates (Table 1). Smaller medians suggest that a few outliers might be at play. Three isolates accounted for 97 “indistinguishable” pairs with most SNV. One also accounted for most pairs “different” by PFGE with 0 SNV. These isolates’ PFGE groups were verified to be correct, and the discrepancies will require further investigation. Overall, 85.6% of pairs had concordant results. 14.3% of pairs were clonal by PFGE, but not WGS (Figure 1). The Wallace coefficient showed that isolates clonal by WGS had a 94.2% chance of sharing a PFGE group, while the converse was true for 14.2%. Compared with PFGE, WGS had a higher Simpson’s index (0.973 vs. 0.845).

    Conclusion:

    C. difficile strain relatedness assessment correlated between PFGE and WGS, but WGS achieved higher levels of resolution. PFGE classified a significant portion of genomically distinct isolates as related.

    Table 1. Numbers of SNV among PFGE categories

    Indistinguishable

    Closely related

    Possibly related

    Different

    SNV ± SD

    489.6 ± 3138.0

    1803.4 ± 4760.0

    3895.0 ± 6090.1

    23846.7 ± 18163.3

    Median (IQR)

    8 (5-12)

    29 (22-58)

    32 (15-10850)

    23648 (12514-25044)

    Title: Relationship between PFGE classification and SNV by WGS

    Ling Yuan Kong, MD1, David Eyre, BM BCh DPhil MCRP FRCPath2, A. Sarah Walker, PhD3, Jacques Corbeil, PhD4, Mark Wilcox, MD5, Anne-Marie Bourgault, MD1,6, Andre Dascal, MD7, Matthew Oughton, MD7, Sophie Michaud, MD, MPH8, Baldwin Toye, MD9, Eric Frost, PhD8, Louise Poirier, MD, FRCPC10, Paul Brassard, MD1, Nathalie Turgeon, MD11, Rodica Gilca, MD PhD12 and Vivian Loo, MD, MSc, FSHEA1, (1)McGill University Health Centre, Montreal, QC, Canada, (2)Nuffield Department of Medicine, University of Oxford, Oxford, United Kingdom, (3)NIHR Oxford Biomedical Research Centre, Oxford, United Kingdom, (4)Centre de recherche CHUQ-CHUL, Sainte-Foy, QC, Canada, (5)University of Leeds, Leeds, United Kingdom, (6)Centre Hospitalier de l'Université de Montréal, Montreal, QC, Canada, (7)Jewish General Hospital, Montreal, QC, Canada, (8)Centre Hospitalier Universitaire de Sherbrooke, Sherbrooke, QC, Canada, (9)Ottawa Hospital, University of Ottawa, Ottawa, ON, Canada, (10)Hôpital Maisonneuve-Rosemont, Montréal, QC, Canada, (11)Centre Hospitalier Universitaire de Québec - Hôtel-Dieu de Québec, QC, Canada, (12)Quebec Institute of Public Health, Quebec, QC, Canada

    Disclosures:

    L. Y. Kong, None

    D. Eyre, None

    A. S. Walker, None

    J. Corbeil, None

    M. Wilcox, None

    A. M. Bourgault, None

    A. Dascal, None

    M. Oughton, None

    S. Michaud, None

    B. Toye, None

    E. Frost, None

    L. Poirier, None

    P. Brassard, None

    N. Turgeon, None

    R. Gilca, None

    V. Loo, None

    Findings in the abstracts are embargoed until 12:01 a.m. CDT, Wednesday Oct. 26th with the exception of research findings presented at the IDWeek press conferences.