Application of broad-spectrum, sequence-based pathogen identification in an urban population

Abstract

A broad spectrum detection platform that provides sequence level resolution of target regions would have a significant impact in public health, case management, and means of expanding our understanding of the etiology of diseases. A previously developed respiratory pathogen microarray (RPM v.1) demonstrated the capability of this platform for this purpose. This newly developed RPM v.1 was used to analyze 424 well-characterized nasal wash specimens from patients presenting with febrile respiratory illness in the Washington, D. C. metropolitan region. For each specimen, the RPM v.1 results were compared against composite reference assay (viral and bacterial culture and, where appropriate, RT-PCR/PCR) results. Across this panel, the RPM assay showed >or=98% overall agreement for all the organisms detected compared with reference methods. Additionally, the RPM v.1 results provide sequence information which allowed phylogenetic classification of circulating influenza A viruses in approximately 250 clinical specimens, and allowed monitoring the genetic variation as well as antigenic variability prediction. Multiple pathogens (2-4) were detected in 58 specimens (13.7%) with notably increased abundances of respiratory colonizers (esp. S. pneumoniae) during viral infection. This first-ever comparison of a broad-spectrum viral and bacterial identification technology of this type against a large battery of conventional "gold standard" assays confirms the utility of the approach for both medical surveillance and investigations of complex etiologies of illness caused by respiratory co-infections.

Figure 1. Rooted phylogenetic analysis of the hemagglutinin (HA3) genes of 15 representative influenza A/H3N2 isolates and reference strains.

(A) phylogenetic tree generated using sequences obtained from RPM v.1, (B) phylogenetic tree generated using de novo sequences from the same set of the isolates. Reference sequences were obtained from GenBank and indicated in underlined, bold font. Numbers above branches indicate bootstrap values from 1000 replicates. Note: ***-the A/Wyoming/3/03 (with IS)-like isolate. Bootstrap values above 50% were shown at branches. Scale near the bottom of each panel relates the length of a branch to the number of nucleic acid substitutions.