สถาบันวิจัยวิทยาศาสตร์สาธารณสุข

National Institute of Health of Thailand

Consistency of Identification of Pseudomonads and Related Organisms

Authors : Micah I. Krichevsky*, Paul De Vos**, Surang Dejsirilert***, Patrick Gillevet****, Deborah Henry*****, Jorge Lalucat******, Edward Moore*******, Mark Segal********, Jane Tang*********, Sue Whitehead**********, Yuguang Zhou ***********, Hans Yu************

 

Affiliations:    *Bionomics International, Wheaton, MD 20902
**University of Gent, B-9000 Gent, Belgium
***National Institute of Health, Nonthaburi 11000 Thailand
****George Mason University, Manassas, Virginia, 20110
*****University of British Columbia, Vancouver, British Columbia V5Z 4H4 Canada
******Universitat de les Illes Balears, E-07071 Palma de Mallorca, Spain
*******TheMacaulayResearch Institute,AB15 8QH Aberdeen,Scotland,UK
********Environmental Protection Agency, Washington, DC 20460
*********American Type Culture Collection, Manassas, Virginia, 20110
**********Children's & Women's Health Centre of B.C., Vancouver, B.C. V6H 3V4 Canada
***********Chinese Academy of Sciences, Beijing 100080 P.R. China
************Health Canada, Tunney's Pasture, Ontario K1A 0L2, Canada
 
 Source:       Present at the ASM Annual Meeting 2004
 
Language:     English
 
Abstract:
 
Those wanting identification of unknown bacteria need to know the reproducibility of identifications, especially to determine possible pathogenicity. One way to evaluate reproducibility is to distribute identical cultures to independent laboratories for analysis. Sponsored by Health Canada, 8 laboratories studied 300 isolates of PseudomonasBurkholderia and related genera (of medical and ecological interest) to support regulatory determination of potential hazard. The laboratories vary in mission, methodologies, and location. A distributing laboratory sent randomly encoded cultures to each laboratory. Using their normal methods, each laboratory submitted best estimates of the identities of the isolates with supporting data. We created databases of genera, species, subspecific identifications; identification failures; and the primary data. We report 1) the proportion of laboratories with the same identification by isolate and 2) for 20 isolates sent in duplicate, the proportion of agreement by laboratory and by isolate. The modal best estimates comprised the species names for calculation of the average consistency. Consistency of reported identification varied widely with taxon. The 46 species averages, with at least 2 laboratories agreeing, spanned 88% (P. aeruginosa) to 50% (16 species) and 49% to 25% (30 species). The intra-lab consistency for the duplicate isolates spanned 75% to 45%. Consistency/isolate was 100% to 25%. However, 80% of modal names assigned to the duplicates agreed. The data demonstrate two overlapping classes of isolates, those consistently identified and those not, largely regardless of the methods used. Identifications based on multiple methods usually were more consistent than single methods. We found no “gold standard” for validating the identifications. The boundaries of the taxa are unknown (unknowable?). From the data and the rules of nomenclature, we conclude there is no absolute measure of the “correctness” of an identification. Only consensus and usage dictate correctness. The underlying taxonomy is continually in flux which further confounds identification reporting.