Smartphone screen testing, a novel pre-diagnostic method to identify SARS-CoV-2 infectious individuals

The COVID-19 pandemic will likely take years to control globally, and constant epidemic surveillance will be required to limit the spread of SARS-CoV-2, especially considering the emergence of new variants that could hamper the effect of vaccination efforts. We developed a simple and robust – Phone Screen Testing (PoST) – method to detect SARS-CoV-2-positive individuals by RT-PCR testing of smartphone screen swab samples. We show that 81.3–100% of individuals with high-viral-load SARS-CoV-2 nasopharyngeal-positive samples also test positive for PoST, suggesting this method is effective in identifying COVID-19 contagious individuals. Furthermore, we successfully identified polymorphisms associated with SARS-CoV-2 Alpha, Beta, and Gamma variants, in SARS-CoV-2-positive PoST samples. Overall, we report that PoST is a new non-invasive, cost-effective, and easy-to-implement smartphone-based smart alternative for SARS-CoV-2 testing, which could help to contain COVID-19 outbreaks and identification of variants of concern in the years to come.

3 together with the Beta (B.1.351, Tegally et al., 2020) and Gamma (P.1,Faria et al.,n.d.), which 25 share the antigenic drift E484K, and the signature mutations K417N and N501Y in the region of 26 the spike protein that is recognised by neutralising antibodies (Collier et al., 2021;Greaney et al., 27 2021b;Wang et al., 2021;Wibmer et al., 2021). 28 Early in the pandemic, the RNA sequence of the SARS-CoV-2 virus was made available (Wu et 29 al., 2020), enabling the testing of infected patients by Reverse Transcription (RT)-PCR (Arnaout et 30 al., 2020;Corman et al., 2020). Regular and broad testing of SARS-CoV-2 seems essential to 31 contain the propagation of SARS-CoV-2, as many infected individuals express no symptoms, 32 inadvertently spreading the infection (Ferretti et al., 2020;Kronbichler et al., 2020;Petersen and 33 Phillips, 2020;Pollock and Lancaster, 2020;Sayampanathan et al., 2021). Therefore, successful 34 epidemiological surveillance strategies required to monitor the spread of SARS-CoV-2 and the 35 outbreak of new strains should include large-scale screening methods that enable periodic and 36 continuous testing of the general population. 37 However, testing capacity has been limited, hampering attempts to control the spread of 38 SARS-CoV-2. One obstacle is that reliable nasopharyngeal sampling and RT-PCR testing is highly 39 invasive and requires both specialised staff and appropriate conditions for the manipulation and 40 transport of the samples to comply with clinical standards and protocols expected by regulatory 41 bodies (CDC, 2021b;UK-Government, 2021). Lateral flow device antigen tests are cheaper, 42 accurate when detecting individuals with high viral load, and an epidemiologically effective 43 option to identify SARS-CoV-2 contagious people (Dinnes et al., 2021;Pavelka et al., 2021;44 Wagenhäuser et al., 2021). Yet, correct testing also requires nasopharyngeal sampling. 45 Therefore, regular large-scale testing is difficult because accurate tests are either too invasive, 46 expensive or logistically complicated to implement, which make them unviable for the task. 47 To provide a simple alternative to identify SARS-CoV-2 infected individuals, we designed and 48 validated a method by which SARS-CoV-2 RNA can be RT-PCR detected from samples taken from 49 7 PoST and nasopharyngeal positive samples corresponded to  symptomatic individuals at the time of testing (n=43/122, Source file 2). This validation study 126 was also performed in the same double-blind conditions as the pilot study described above. 127 The overall specificity of PoST during this validation was 97.6% (Figure supplement 1B); 14 128 negative SARS-CoV-2 nasopharyngeal samples were identified positively by PoST ( Figure  129 supplement 1B, Source file 2). Of these cases, we could contact 4 individuals, of which three had 130 clear COVID-19 symptoms, one of which tested positive when the nasopharyngeal sample was 131 repeated (Source file 2). This suggests that some of the discrepancies may be due to 132 nasopharyngeal false-negative test results. 133 Early in the pandemic, the notion that SARS-CoV-2 could spread via surfaces made regular 134 disinfection of surfaces a common practice (Goldman, 2020). We therefore evaluated whether 135 cleaning smartphone screens before PoST sampling could affect the test results. Of all the SARS-136 CoV-2 PoST and nasopharyngeal positive samples where the screen had been cleaned within 24 137 hrs (n=23), 22% (5/23) were taken from phones that had been sanitised the same day (<6 hrs), 138 and 48% (8/23) less than 2 hrs before the PoST sample was taken (Source file 3). This suggests 139 that sanitising or cleaning the smartphone before the PoST sampling may not affect the 140 detection of SARS-CoV-2 traces on the phone screen. 141 This validation study confirms that the PoST method shows a high sensitivity when identifying 142 SARS-CoV-2 infected individuals with a low RT-PCR Ct value, regardless of their symptoms, 143 providing a new valuable tool to tackle the COVID-19 pandemic. 144

SARS-CoV-2 variant identification in PoST samples 145
The current challenge in COVID-19 diagnostics is to find ways to distinguish the specific SARS-146 CoV-2 variant present in the tested samples to aid in containing outbreaks of variants with 147 higher virulence or infectivity. To aid this task, we performed a screen to address whether SARS-  Even though these samples did not test positive for K417T, it is still likely that the virus in these 157 samples corresponds to either Beta or Gamma, as the combination of variations that tested 158 positive have only been found together in these two SARS-CoV-2 variants. One sample only 159 tested positive for SNP K417T, which is only found in Gamma and Beta SARS-CoV-2 variants. Our results imply that the capacity to identify SARS-CoV-2 RNA by the primer/probe set we 164 used for PoST is not affected by the mutations described for the variant of concern we identified. 165 The capacity of identifying SARS-CoV-2 variants from PoST samples, differentiates this assay from 166 lateral flow device antigen testing which can only identify the presence of the virus and not 167 discriminate specific variants. 168 169

Discussion 170
A recent study suggests that socioeconomic status and delay in testing for SARS-CoV-2 was a 171 significant factor in the high mortality rate observed during 2020 in Santiago, Chile (Mena et al., 172 2021). Therefore, finding new methods to enhance epidemiological surveillance strategies are 173 necessary to limit the struggle generated by the COVID-19 pandemic.

9
Our study suggests that the PoST method could be an effective non-invasive way to rapidly 175 identify COVID-19 positive contagious individuals who actively spread the virus, regardless of 176 their symptoms. This method does not require specialised operators or conditions for sampling. 177 Furthermore, because PoST is an environmental test, the protocols and reagents required to 178 manipulate and process the samples for RT-PCR have been optimised to make PoST substantially 179 more cost-effective, circumventing the need for clinical grade reagents and standards. All which 180 makes PoST a good alternative for large-scale population testing. to contain the spreading of variants of concern that are more pathogenic or infective, to contain 201 and isolate their dissemination. We successfully identified one PoST sample that presented all 202 three polymorphisms present in Beta (B.1.351) and Gamma (P.1) variants. The rest of the tests 203 did not unequivocally show which SARS-CoV-2 variant was present in the sample but was 204 enough to conclude the nature of the variant present in the sample. Even though optimisation 205 will be required to enhance SARS-CoV-2 variant identification in PoST samples, our results are 206 encouraging, and could make PoST a viable option to screen for variants of concern. In this 207 scenario, and although the output of results would take longer, PCR and sequencing will likely 208 enhance the performance of variant detection in PoST samples. 209 Regarding nasopharyngeal RT-PCR results with a high Ct value (i.e., lower-viral load), these 210 are probably in transition, and either starting or ending a COVID-19 infection. Therefore, they 211 are less likely to shed SARS-CoV-2 virions that can be detected by PoST. This transition is the 212 reason why periodic regular testing is required to identify those infected individuals once they 213 enter a contagious phase. 214 To a lesser extent and against the overall trend we observe for PoST detection, PoST positive 215 results were identified in this high Ct value nasopharyngeal RT-PCR results group. Given the 216 unstable nature of the SARS-CoV-2 virus on surfaces (Goldman, 2020), and the pervasive activity 217 of RNases which are ubiquitously present in the environment (Probst et al., 2006), it is unlikely 218 that these positive PoST results are due to the detection of long-lasting virus or RNA on the 219 phone screen surface. Hence, we could exclude the possibility that these PoST positive results 220 are the consequence of detecting RNA from when individuals were passing through a moment of 221 higher COVID-19 infection. Therefore, one could speculate that these results are the 222 consequence of suboptimal nasopharyngeal sampling, which would explain the low amount of 223 virus in the sample and hence, a high Ct value. Alternatively, it is plausible that these samples 224 belong to individuals with a low viral load that are still actively contagious, which is against the 225 observed trend (Bullard et al., 2020;Jefferson T et al., 2020;Sonnleitner et al., 2021), but cannot 226 be excluded as a possibility. It will be interesting to perform a follow-up study on these 227 individuals as they could be part of a group with a higher capacity to shed SARS-CoV-2 along 228 their infectious cycle, potentially explaining the COVID-19 superspreading capacity observed in 229 some people (Lewis, 2021). 230 The distribution profile of the nasopharyngeal RT-PCR Ct test results in our study was 231 different in the pilot and validation cohorts. The pilot Ct results presented an apparent bimodal 232 distribution with low and high Ct value populations ( Figure 1B). On the other hand, in the 233 validation study, the distribution was unimodal and most samples tended towards low RT-PCR Ct 234 results ( Figure 1D). We speculate that this stark difference could be due to an inherent  One of the main advantages of PoST is its low cost compared to clinical nasopharyngeal tests. 249 Excluding staff and premises, the net cost of a nasopharyngeal test in Chile ranges between $20-250 25 USD which includes: sample tube, transport media, swab, RNA extraction kit and RT-PCR kit. 251 Because PoST does not use clinical-grade consumables and reagents, its net cost for the same 252 items ranges between $2-3USD when it is the result of a single sample tested RT-PCR reaction. 253 This net cost can go below $1USD when PoST samples are pooled and groups of 5 and a single 254 RT-PCR reaction is performed, which we have shown to produce reliable results for PoST samples 255 (data not shown). 256 Besides PoST not using clinical-grade consumables and reagents, one key optimisation is that 257 PoST does not require an RNA extraction step, which reduces the cost by approximately $7USD. 258 Furthermore, instead of purchasing an RT-PCR kit specifically designed for SARS-CoV-2 detection, 259 we assembled a combination of off-the-shelf probe/primers to detect the SARS-CoV-2 gene 'N' 260 together with a generic RT-PCR kit. All together this optimisation enables a 10-fold price 261 reduction of the net cost of PoST compared to a regular nasopharyngeal RT-PCT test. 262

b. High sensitivity and specificity 263
The PoST method has a similar sensitivity and specificity, compared to antigen lateral flow 264 devices, which are extensively used for routine testing (Dinnes et al., 2021;Jääskeläinen et al., 265 2021;Wagenhäuser et al., 2021). Clinical grade diagnostic RT-PCR kits include the detection of 266 three SARS-CoV-2 genes, plus a human positive control, either multiplexed or individually 267 detected. The specificity reached by PoST in this study was achieved only when detecting the 'N' 268 SARS-CoV-2 gene. This could explain why PoST can miss identifying some positive 269 nasopharyngeal RT-PCR samples in the low Ct range. A study adding the detection of two or 270 three SARS-CoV-2 genes in the PoST protocol to assess if a higher sensitivity is achieved would 271 enable us to calculate whether this trade-off is enough to justify increasing the net cost of the 272 PoST assay. Especially considering that the sensitivity described in this study is already high 273 13 enough to detect positive COVID-19 individuals to affect limiting the transmission of the SARS-274 CoV-2 virus (Kennedy-Shaffer et al., 2021;Larremore et al., 2021;Mina et al., 2020). 275

c. Sampling speed, result turnaround of PoST. 276
One other advantage of the PoST method is that the sampling process takes around a minute 277 at most, and does not require a particular setup besides having a sterile swab and sample tube. CoV-2 samples are found, and 6.5 hours if positive individual samples are to be identified from 289 pools. These times consider two technicians processing the samples to feed one 96 well plate 290 real-time PCR machine. Therefore, under these conditions, results for 940 tests could be 291 provided within the same day of sampling, which is ideal to isolate contagious cases, and 292 effectively curb the spreading of COVID-19. 293

d. PoST as an alternative self-testing method 294
Due to its high sensitivity, specificity, and rapid result turnaround time, lateral flow device 295 antigen testing has become widely used to screen for COVID-19 cases operated by trained staff 296 (Pavelka et al., 2021), and self-administered (Riley et al., 2021). Because this method uses highly 297 invasive nasopharyngeal swabbing, trained operators are the preferred option to deliver 298 14 accurate and reliable results from these tests. From this point of view, the PoST method offers a 299 valuable alternative for accurate self-tests results as it is much easier and more reliable to 300 effectively swab the screen of a smartphone than performing a self-administered 301 nasopharyngeal test. This, together with the fact that SARS-CoV-2 variants can be identified in 302 PoST samples, gives this method further advantages compared to lateral flow antigen testing. 303

e. Penetration of smartphone use in the population 304
One aspect that is important to mention is that although it is estimated that there are around 305 3.8 billion smartphones in the world and their penetration is very high among the young and 306 adult population, their global distribution is not equitable (Turner, 2021). While penetration is 307 almost total among adults in developed countries, in countries such as India or Bangladesh it 308 does not exceed 33% of the population (Berenguer et al., 2016). On the other hand, among 309 senior citizens, who are precisely a vulnerable population, even in developed countries some do 310 not use a telephone or have older devices. 311 312 Overall, this study was aimed to validate PoST as method to identify infected COVID-19 cases 313 by using the smartphone as a proxy from which traces of the SARS-CoV-2 virus of the owner can 314 be detected. We propose that this highly-sensitive, non-invasive and cost-effective method 315 could well be used for mass testing and help to contain the spreading of other airborne 316 contagious diseases and outbreaks when tackling future epidemics. This could be particularly 317 useful as an early warning system for rapid detection of respiratory pathogens in public health 318 efforts to contain local outbreaks to prevent further escalating to other areas. 319 320

Materials and Methods 321
Smartphone screen swab sampling, sample processing and RT-PCR. 322 Informed consent and consent to publish was obtained from all the individuals that participated 323 in this study before performing the sampling process. Dacron swabs were briefly dipped in 324 Weise medium (Merck, 1.09468.0100) and then used to swab the bottom half of mobile phone 325 screens by a member of our team as shown in Figure 1A. Swabs were then introduced in sterile 326 conical tubes containing Weise medium and briefly hand spun. Samples were processed for RT-327 PCR within 8 hours. 328 Aliquots of swab samples were incubated at 70ºC for 10min as previously described (Miranda et 329 al., 2020). Samples were left to cool at room temperature and 3.3µl aliquots were used for RT-330 PCR using Promega GoTaq Probe 1-Step RT-qPCR system (A6121) according to the manufacturer 331 instructions and supplemented with SARS-CoV-2 N2 probes and primers (IDT#10006606) on an 332 Illumina Eco Real-Time PCR System. We considered a test as positive if the PCR amplification 333 obtained followed the expected standard sigmoidal kinetics of amplification. 334

Clinical sampling and RT-PCR. 335
For each patient, nasopharyngeal swabs were collected using standard technique, as 336