Each assay’s analytical sensitivity was determined by testing duplicate DNA samples, RNA transcripts or whole organism for different analytes and repeated at least three times to determine the assay’s reproducibility. This LOD was determined by the lowest dilution which gave a positive EHA reading greater than 0.4. The LOD of the DNA assay for genomic DNA and whole organisms is shown in Table 1.

The LOD for genomic DNA was at 1×10¹ copies/mL for BA pX01 and FT, 1×10² copies/mL for BA pX02 and 1×10⁰ copies/mL for YP. The LOD for VZV whole organism was 1×10^-2 TCID₅₀/mL. The LOD for recombinant plasmid DNA controls was evaluated in the same way and is shown in Table 1. The LOD for both BA pX01 and BA pX02 recombinant controls was 1×10²copies/mL. The analytical sensitivity for FT, YP and VM were 1×10³, 2.5×10² and 5×10² copies/mL, respectively.

The LOD of the RNA assay with or without extraction for RNA transcript control is shown in Table 2.

The LOD without extraction for RNA transcript control was at 1×10⁶ copies/mL for Ebola and RVF, 1×10⁵ copies/mL for Lassa and 1×10⁴ copies/mL for HSN. The LOD with extraction for RNA transcript control was at 1×10⁵ copies/mL for Ebola, Lassa and HSN, and 1×10⁶ copies/mL for RVF.

The LOD for dengue whole organism (4 serotypes) is shown in Table 2. The LOD for dengue whole organisms was approximately 1×10^-4 dilution for dengue 1 and 2, 1×10⁴ LD₅₀/mL for dengue 3 and 1×10² LD₅₀/mL for dengue 4.

The LOD without extraction for recombinant plasmid DNA controls was evaluated in the same way and is shown in Table 2. The LOD without extraction for recombinant plasmid DNA controls was at 1×10³ copies/mL for all 4 targets. This demonstrated outstanding sensitivity (~1.5 input copies/reaction) for the primers and probes of the RNA assay in PCR-EHA and suggests that our RNA controls did not represent the LOD we would expect to see in true clinical testing. We hypothesize that our RNA transcripts are unstable and degradation is likely during the dilution and extraction process.

The analytical specificity of the assays was determined using a number of closely related organisms and other common respiratory pathogens that might be found in clinical samples. These samples were tested for cross-reactivity to the primers and probes. Table 3 described the organisms, the dilutions used and the results.

No cross-reactivity between primers, probes and these organisms was detected in either assay. The internal control probe in the RNA assay can identify human parainfluenza virus 2 (HPIV-2) because the RNA assay used HPIV-2 as an internal control. HPIV-2 produced negative result for all other primers and probes in this assay except the internal control which should be positive.

Two hundred and sixty four clinical samples (196 NP swabs, 45 skin swabs, 15 serum, 7 sputum, 1 tracheal) were tested with the DNA assay. The results were shown in Table 4A. In this testing there were three indeterminants for FT and when reprobed as per standard protocol for indeterminants (OD of 0.300-0.399) they were found to be negative. There was one false positive (low positive) for FT (OD=0.838) which when retested was also found to be negative. The clinical specificity of the BioT DNA assay was 99.6% (263/264) (95% CI 98-100). Of the 549 clinical samples that were tested by the RNA assay, there was one false positive (OD=0.626) for HSN and when retested as per standard protocol for false positive (OD>0.400) it was found to be negative (OD=0.040). The overall clinical specificity of the BioT RNA assay was 99.8% (548/549) (95% CI 99-100).

The results were shown in Table 4A. Surrogate clinical NP samples with FT (20/20), BA pX02 (20/20), YP (20/20), and skin specimens with VZV (20/20) and VM (20/20) were detected and had sensitivities of 100% (95% CI 83-100). BA pX01 was detected in 19/20 NP samples (sensitivity 95% (95% CI 75-100).

There was one sample positive for VZV in a clinical specimen spiked with VM. Using the standard assay protocol to re-probe any positive (~30 minute procedure) it was found to be a true negative. This was most likely a technical error. However, even counting it as a false positive would not have much of an impact on the specificity for VZV which was 99.8% (95%CI 99-100). One sample tested very low positive for VM (OD=0.434) in the BA pX01 spiked samples. Re-probing of the PCR product continued to show the sample to be positive. However, retesting the original clinical sample demonstrated it to be negative. Therefore, this represented a true false positive (1/400 samples). Specificity for VM was 99.8% (95% CI 99-100). No other false positives where detected using the FT, BA pX01, BA pX02, and YP probes (specificities 100% (95% CI 99-100). Finally, all 36 negative control M4 samples were negative for all organisms.

A total of 104 serum specimens were used to test the surrogate clinical sensitivity and specificity of the RNA assay. The results are shown in Table 4B. For dengue 2, 3, and 4, 20/20 serum samples spiked were correctly detected resulting in sensitivities of 100% (95%CI 83-100), and dengue 1 testing showed 19/20 samples were correct detected and the sensitivity was 95% (95% CI 75-100). No false positive results were found using any of the probes (including Ebola, HSN, RVF, and Lassa) and the BioT RNA assay showed 100% specificity (95% CI 96-100). The 24 negative control samples were negative for all targets making the overall specificity of the assay 100% (95% CI 99-100).

The primers and probes used in these 2 assays were designed from highly conserved regions of specific genes for each organism listed in Table 5A and Table 5B.

The primers and probe for the DNA assay specifically targeted the Tul4 gene for F. tularensis, the HA gene for V. major, the ORF29 gene for the Varicella zoster virus and the Virulence Antigen gene for Y. pestis. 2 pairs of primers for B. anthracis were designed from the conserved regions of the Protective Antigen (PA) and Cap gene located respectively on the plasmids pX01 and pX02 because both plasmids are required for pathogenicity. The primers and probes for the RNA assay specifically targeted the L gene for Ebola, the GP2 gene for RVF, the L gene for the Lassa, the S segment for Hanta and the 3’UTR for dengue.

Varicella zoster virus (VZV) and all four serotypes of dengue virus were obtained from ATCC. Genomic DNA of FT and of BA (BA pX01-Sterne strain) were obtained from the Medical College of Wisconsin. Genomic DNA of YP was obtained from the Blood Center of Wisconsin. The genomic DNA of BA (BA pX02) was obtained from the University of Chicago.

To create positive controls, PCR products generated from whole organisms were purified and cloned into plasmids using the TOPO TA cloning system (Invitrogen, Carlsbad, CA), following the manufacturer’s instructions. All positive controls were sequenced to verify the amplicon sequences and the possibilities of any mismatches generated during cloning. The positive controls were further amplified to generate large quantities that were purified using Qiagen Plasmid Midi Kit (Qiagen, Valencia, CA). The synthetic RNA standards were generated from linearized target sequence plasmids by using T7 in vitro transcription system (Promega, Madison, WI).

The quantity of DNA and RNA transcripts were measured by the absorbance at 260 nm and the purity of the DNA or RNA by the ratio at A260/280. To determine the approximate copy number of plasmid controls and synthetic RNA controls, the absorbance value of the positive control (A₂₆₀) was used in combination with Avogadro’s Number (6.022 × 10²³) as described by Khanna et al. (2005) [41]. The RNA transcripts were diluted in RNase and DNase free water with RNase Inhibitor (Applied Biosystems, Foster City, CA), frozen at -80˚C in 30 μL aliquots. Serial dilutions of the plasmids and RNA transcripts were used to establish the analytical sensitivity and specificity of BioT DNA and RNA multiplex assays.

The analytical sensitivity for the DNA assay was determined using serial dilutions of both DNA extracted from whole organisms and recombinant DNA controls. Ten-fold serial dilutions of genomic DNA (BA, YP, and FT), previously extracted from whole organisms and plasmid DNA controls were made (10⁴ to 10¹ or 10^-1 copies/mL) in M4 medium (Remel, Lenexa, KS), extracted using the High Pure Viral Nucleic Acid Kit according to manufacturer’s instructions (Roche Applied Science, Indianapolis, IN) and then tested. VZV whole organism was serially diluted (10³ and 10^-3 CFU/mL) in M4 and extracted using the same method.

The analytical sensitivity for the RNA assay was determined using serial dilutions of both RNA transcripts and recombinant DNA controls for HSN, RVF, Lassa, and Ebola virus. The analytical sensitivity of the assay for dengue was determined using serial dilutions of all four serotypes of dengue virus. Ten-fold serial dilutions of RNA transcript and recombinant DNA control were made (10⁷ to 10³ copies/mL for RNA transcripts, 10⁶ to 10² copies/mL for recombinant DNA control) using two matrices: M4 medium and RNase/DNase-free water. The RNA and plasmid controls diluted in water were used as a template in RT-PCR to determine the limit of detection (LOD) without extraction. The RNA controls diluted in M4 were extracted to determine the LOD with extraction. Dengue (4 serotypes) whole virus was serially diluted at 10^-2 to 10^-5 dilutions for dengue serotypes 1 and 2 (the virus stock for these serotypes came from ATCC without quantitation ), 10⁵ to 10² LD₅₀/mL for dengue serotype 3, and 10³ to 10⁰ LD₅₀/mL for dengue serotype 4. Diluted whole viruses were extracted using high pure viral nucleic acid kit (Roche) according to manufacturer’s instructions and tested.

Specificity of the DNA and RNA assays were evaluated by respectively testing 17 and 15 common respiratory pathogens that could be present in clinical specimens for potential cross-reactivity. ATCC bacterial and viral strains tested by the DNA assay were diluted to 1×10³ CFU/mL and 1×10² -10⁴ TCID₅₀ /mL. ATCC bacterial or viral strains tested by the RNA assay were diluted to 1×10³ CFU/mL, 1×10⁴ CFU/mL, 1×10³TCID₅₀ /mL or 1×10⁴ cells/mL respectively. The bacterial and viral strains were then tested using the same method as previously described. The plasmid DNA controls and whole virus, which are the targets of both BioT assays, were diluted and tested at 1 to 2 log higher than the LOD using the same method.

For the DNA assay, 264 clinical samples (196 NP swabs, 45 skin swabs, 15 serum, 7 sputum, 1 tracheal), were collected from 264 subjects at Children’s Hospital of Wisconsin and Froedtert Memorial Hospital during a 17-month period from October 2005 through March 2007 and were tested. Of the 264 subjects, 130 presented with respiratory symptoms and 134 presented without respiratory symptoms. Sixty three of the samples were children aged 0-6 years old, 20 were children aged 7-18 years old, 154 were aged 19-65 years old, and 27 were older than 66. Of the 264 clinical samples tested, 139 were females and 125 were males.

For the RNA assay, 549 clinical samples (417 NP swabs, 95 skin swabs, 21 blood, 12 sputum, 3 tracheal, 1 nasal swabs) were tested. The samples were collected from 418 subjects as described above. Of these samples, 302 were from the subjects with respiratory symptoms and 247 were from subjects that had no respiratory symptoms. Two hundred and thirty nine of the samples were from children age 0-6 years old, 106 were from children age 6-19 years old, 182 were from subjects age 19-65 years old and 22 were from subjects older than 66. Of the 549 clinical samples tested, 258 were from females and 291 were from males. Any sample with OD value lower than 0.300 was considered negative. Any sample with OD value between 0.300-0.399 was considered indeterminant and was reprobed to confirm the results. Any sample with OD value over 0.400 was considered positive and was retested.

To evaluate the performance of the DNA assay we used surrogate positive clinical samples. A total of 120 clinical NP swab and skin swab specimens that previously tested negative with the same assay were used for spiking experiments. Twenty clinical specimens were spiked with plasmid DNA positive control and six M4 transport medium specimens were set up as negative specimens for each of the six organisms mentioned (36 total). The recombinant positive controls of FT and BA pX01 were spiked into NP swab specimens at the concentration 1×10⁴ copies/mL. The recombinant positive controls of YP and BA pX02 were spiked into NP swab specimens at the concentration 1×10⁵ copies/mL. The recombinant positive control of VM was spiked into skin swab samples at the concentration 1×10⁴ copies/mL. The VZV whole organism was spiked into skin swab specimens at 1×10⁰ TCID₅₀/mL.

To evaluate the performance of the RNA assay for positive clinical samples, pooled serum samples that previously tested negative with this assay was aliquoted into 400μL samples and used for the spiking experiment. Twenty serum samples were spiked with whole organism of each of the serotypes of dengue virus (80 total) and 6 serum samples were run as negative specimens with each serotype of dengue virus (24 total). dengue 1 and dengue 2 were spiked at 1×10^-2 dilution, the concentration of which was 2 logs higher than the LOD of dengue 1 and 1 log higher than the LOD of dengue 2. Dengue 3 was spiked at 5×10⁵ LD₅₀/mL, the concentration of which was 1.5 logs higher than the LOD of dengue 3. Dengue 4 was spiked at 1×10⁴ LD₅₀/mL, the concentration of which was 2 logs higher than the LOD of dengue 4. All of these spiked specimens were tested blindly.