Proof of concept of the potential of a machine learning algorithm to extract new information from conventional SARS-CoV-2 rRT-PCR results.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has been and remains one of the major challenges modern society has faced thus far. Over the past few months, large amounts of information have been collected that are only now beginning to be assimilated. In the present work, the existence of residual information in the massive numbers of rRT-PCRs that tested positive out of the almost half a million tests that were performed during the pandemic is investigated. This residual information is believed to be highly related to a pattern in the number of cycles that are necessary to detect positive samples as such. Thus, a database of more than 20,000 positive samples was collected, and two supervised classification algorithms (a support vector machine and a neural network) were trained to temporally locate each sample based solely and exclusively on the number of cycles determined in the rRT-PCR of each individual. Overall, this study suggests that there is valuable residual information in the rRT-PCR positive samples that can be used to identify patterns in the development of the SARS-CoV-2 pandemic. The successful application of supervised classification algorithms to detect these patterns demonstrates the potential of machine learning techniques to aid in understanding the spread of the virus and its variants.

Case report: BA.1 subvariant showing a BA.2-like pattern using a variant-specific PCR assay due to a single point mutation downstream the spike 69/70 deletion.

BACKGROUND: SARS-CoV-2 variant tracking is key to the genomic surveillance of the COVID-19 pandemic. While next-generation sequencing (NGS) is commonly used for variant determination, it is expensive and time-consuming. Variant-specific PCR (vsPCR) is a faster, cheaper method that detects specific mutations that are considered variant-defining. These tests usually rely on specific amplification when a mutation is present or a specific melting temperature peak after amplification. CASE PRESENTATION: A discrepant result between vsPCR and NGS was found in seventeen SARS-CoV-2 samples from Galicia, Spain. A cluster of BA.1 Omicron SARS-CoV-2 variant showed a BA.2-like melting temperature pattern due to a point mutation (C21772T) downstream the deletion of the spike amino acids 69/70. As the 69/70 deletion is widely used for differentiation between BA.1 and BA.2 by vsPCR, C21772T can cause BA.1 samples to be misinterpreted as BA.2. Over a thousand BA.1 sequences in the EpiCoV database contain this mutation. CONCLUSIONS: To our knowledge, this is the first case of a point mutation causing a vsPCR algorithm to misclassify BA.1 samples as BA.2. This is an example of how mutations in the probe target area of vsPCR tests based on melting curve analysis can lead to variant misclassification. NGS confirmation of vsPCR results is relevant for the accuracy of the epidemiological surveillance. In order to overcome the possible impact of novel mutations, diagnostic tools must be constantly updated.

Pooling for SARS-CoV-2 control in care institutions.

BACKGROUND: Workers and residents in Care Homes are considered at special risk for the acquisition of SARS-CoV-2 infection, due to the infectivity and high mortality rate in the case of residents, compared to other containment areas. The role of presymptomatic people in transmission has been shown to be important and the early detection of these people is critical for the control of new outbreaks. Pooling strategies have proven to preserve SARS-CoV-2 testing resources. The aims of the present study, based in our local experience, were (a) to describe SARS-CoV-2 prevalence in institutionalized people in Galicia (Spain) during the Coronavirus pandemic and (b) to evaluate the expected performance of a pooling strategy using RT-PCR for the next rounds of screening of institutionalized people. METHODS: A total of 25,386 Nasopharyngeal swab samples from the total of the residents and workers at Care Homes in Galicia (March to May 2020) were individually tested using RT-PCR. Prevalence and quantification cycle (Cq) value distribution of positives was calculated. Besides, 26 pools of 20 samples and 14 pools of 5 samples were tested using RT-PCR as well (1 positive/pool). Pooling proof of concept was performed in two populations with 1.7 and 2% prevalence. RESULTS: Distribution of SARS-CoV-2 infection at Care Homes was uneven (0-60%). As the virus circulation global rate was low in our area (3.32%), the number of people at risk of acquiring the infection continues to be very high. In this work, we have successfully demonstrated that pooling of different groups of samples at low prevalence clusters, can be done with a small average delay on Cq values (5 and 2.85 cycles for pools of 20 and 5 samples, respectively). CONCLUSIONS: A new screening system with guaranteed protection is required for small clusters, previously covered with individual testing. Our proposal for Care Homes, once prevalence zero is achieved, would include successive rounds of testing using a pooling solution for transmission control preserving testing resources. Scale-up of this method may be of utility to confront larger clusters to avoid the viral circulation and keeping them operative.