Reference material development for detection of human respiratory syncytial virus using digital PCR

Establishment of RT-dPCR for quantification and typing of RSV

According to previous reports, the most commonly tested target genes were N, M, and L. Four sets of primers and probes (Table 1) were tested and compared by one-step and two-step RT-dPCR using the in vitro transcribed RNA of RSV-A and RSV-B as templates (Fig. 1). The first two sets were universal for RSV and the last two sets were specific for type A and B RSV. As the three genes were constructed into the same molecule, theoretically their quantification results should be 1:1:1. Due to the difference in efficiency of amplification and reverse transcription, the latter being more critical especially for RNA, the four assay sets yielded different results. Both RSV-A and RSV-B obtained higher results by the one-step RT-dPCR than by the two-step RT-dPCR, except for assay 3 of RSV-A. For one-step RT-dPCR, the results of assay 4 were about 2.5% higher than assay 1, about 5% higher than assay 2, and about 40–60% higher than assay 3. Therefore, assay 4 was selected for the subsequent study for both type A and B RSV. The primer and probe concentration and annealing temperature were then further optimized, and the duplex RT-dPCR assay for simultaneous detection of RSV-A and RSV-B was established.

Validation of the RT-dPCR assay

The limit of blank (LoB), limit of detection (LoD), and limit of quantification (LoQ) for the RT-dPCR assays were determined to be 2, 5, 23 copies per reaction for RSV-A and 4, 8, and 20 copies per reaction for RSV-B according to EP 17A [15]. Sixty blank samples were used to determine the LoB (Table S2). Three or four levels of low-concentration samples with 20 samples for each level were used to determine the LoD (Table S3) and LoQ (Table S4).

The dynamic range of the assay was evaluated using serially diluted RNA samples. Ideal linearity was achieved over a 5-log dynamic range, from 101 to 105 copies per reaction, with R2 of 0.9994 for RSV-A and 0.9991 for RSV-B (Fig. 2). The repeatability of the assay represented by the relative standard deviation (RSD) of copy number concentration ranged from 1.5% to 23.7% for RSV-A and 2.6% to 16.8% for RSV-B (Table S5). The specificity of the duplex assay was evaluated by cross-checking of RSV-A and RSV-B RNA, and no false-positive results were obtained (Fig S1). The established duplex RT-dPCR assay was confirmed to be sensitive for quantitative detection and subgroup-specific identification of RSV.

Fig. 2figure 2

Linearity of RT-dPCR assays for RSV-A (A) and RSV-B (B), and homogeneity of reference materials for RSV-A (C) and RSV-B (D). In (C) and (D), dots represent the means of each unit, solid lines indicate the reference value of the RMs, and dashed lines indicate the expanded uncertainty of the RMs

Homogeneity and stability of the RMs

Under the same conditions, ten units of each RM were assessed by RT-dPCR for homogeneity (Fig. 2, Table S6). The data did not indicate outlying unit means or any obvious trend in the filling sequence. The overall %CV of the data was 7.9% for RSV-A and 5.9% for RSV-B. We evaluated the between-unit standard deviation (sbb) for each RM and found that sbb was comparable with method repeatability standard deviation (sr). Therefore, the RMs were considered sufficiently homogeneous, and sbb was used as the estimate of standard uncertainty related to possible between-unit heterogeneity (ubb).

STS study showed that there was no change in copy number concentration of RMs after storage at 4 °C and 25 °C for 7 days. LTS study showed that the two RMs were stable for as long as 13 months under −70 °C (Fig. S2 and Table S7). The LTS study data were plotted against storage time, and regression lines of RNA concentration (yi) versus time (xi) were calculated. The slope of the regression lines (β1) and the standard deviation of the slope [s(β1)] were then tested for statistical significance. The uncertainty contribution of LTS study (ults) was calculated as the product of the chosen shelf life and s(β1).

Characterization of the RMs

The certified value of the RM is represented as the copy number concentration of RNA determined by the established RT-dPCR method. The result of measurements are shown in Table 2. Shapiro–Wilk test showed that the data followed normal distribution. The Dixon and Grubbs tests showed no outliers. The certified values were calculated from the arithmetic mean of the five samples. The estimated relative standard uncertainty related to characterization (\(_)}\)) was calculated as

$$_)}=\sqrt^}_)}+^}_)}^}_)}}$$

where \(_)}\) is the RSD of the measurement, \(_)}\) is the uncertainty of the partition volume of dPCR (0.008) [18], and \(_)}\) is the calibration of the pipette (0.008).

Table 2 Certified values and expanded uncertainties (U, k = 2) of RMs

The source of uncertainty for RMs mainly included characterization, homogeneity, and LTS study of the RM (Table S8). Finally, the certified value and expanded uncertainty (U, k = 2) of the two RMs were determined to be (6.1 ± 1.4) × 104 copies/μL for RSV-A and (5.3 ± 1.2) × 104 copies/μL for RSV-B.

留言 (0)

沒有登入
gif