**2. Results**

Seventeen laboratories submitted results, representing 17 countries from the Mediterranean and Black Sea regions (Figure 1).

### *2.1. Virus Genome Detection*

A total of 13 datasets were received from 13 labs (76.4% of response) for the generic detection of flaviviruses. The results of all labs are shown in Table 1. Four labs did not carry out this technique due to lack of specific equipment to perform conventional RT-PCR assays and were therefore unable to detect the flavivirus positive sample of the panel corresponding to Japanese encephalitis virus (JEV).



*Pathogens* **2020**, *9*, 1038

Out of the 13 datasets, only 6 (46.1%) were 100% correct, i.e., identified as positive the eight samples containing WNV L1, L2, USUV or JEV and as negative the two negative samples of the panel. However, it should be noted that for three of the positive samples, weak positive results (weak bands) were expected. The two samples (W2 and W8) with expected strong positive bands were correctly identified by all the labs (Table 1). Three false positive results were reported by two labs.

For the specific detection of WNV and USUV by RRT-PCR, we received 18 datasets from the 17 participating labs (100% of response), including one double dataset from lab #16 that used an alternative method [21] apart from the recommended one. The results of the labs using the recommended method are shown in Table 2. Overall, the results of the triplex RRT-PCR were very good, with the exception of one laboratory (#1) that reported incorrect results in the whole panel (Table 2). If we exclude the results of this lab from the global analysis of the EQA, out of 16 labs, 12 (75%) reported 100% concordant results, which means that they correctly identified WNV L1, L2 and USUV in all the positive samples (including co-infections).

However, a technical limitation was observed in the application of the triplex RRT-PCR in two labs (#8 and #12), because the Cy5 fluorescent channel necessary for the detection of USUV was lacking in the available real-time thermocyclers. Interestingly, one of these labs was able to overcome this limitation by applying an alternative conventional RT-PCR for the specific detection of USUV [22], obtaining 100% correct identification of USUV positive samples.

In general, highly satisfactory results were obtained for WNV L1 and L2 while more difficulties were observed for the correct identification of USUV positive samples (W3, W6 and W8) as shown in Table 2. In fact, out of the eight false negative results reported by four labs, seven corresponded to USUV and one to WNV L1.

Excluding results from lab #1, co-infections were successfully detected in 87.5% of cases for sample W3 (WNV L1+USUV) and in 81.2% of cases for sample W8 (WNV L2+USUV). The sample containing a related flavivirus (JEV) was correctly identified as negative by all the labs. Overall, the reported Ct values for the three viruses were in line with the reference values, except for labs #3 and #10 that reported lower Ct values for WNV L2 and labs #2 and #17 that reported higher Ct values than expected for WNV L2 and USUV in three samples. With regard to the negative samples, and excluding lab #1, only one laboratory reported a false positive result. Differences in qualitative results or Ct values were not attributable to a specific thermocycler.

Lab #16 correctly identified the WNV positive samples of the panel using a pair of alternative RT-PCR methods for specific detection of WNV L1 and L2 [21].



.



.



**Table 2.** *Cont*.

Blue: false negative results; Red: false positive results; NA: not analyzed due to lack of appropriate fluorophore (Cy5) † Percentage of samples where all the viruses have been correctly identified; ‡ Percentage of 100% correct results by sample (all the viruses of each sample have been correctly identified).
