*2.1. SARS-CoV-2 RBD and N ELISA Validation*

RBD and N ELISA assays had a limit of detection of 1:100. Specificity of the SARS-CoV-2 ELISA assays was determined using 183 human serum samples collected before the pandemic [15]. This included healthy volunteers, as well as patients who tested positive for other viral or inflammatory conditions, including other human coronaviruses (See Supplementary Table S1. Legend for details).

With an endpoint titer cutoff of 300, the assays had 89.6% (RBD) or 94.5% (N) specificity. With an endpoint titer cutoff of 900, the assays had 97.8% (RBD) or 98.9% (N) specificity (Figure 1a,b). Sensitivity of the assays was determined using 134 human serum samples from patients known to have COVID-19 between 16 March 2020 and 12 May 2020 [15]. In some cases, multiple samples from an individual patient were available on different days post symptom onset. Samples collected on or after Day 14 post symptom onset were used for sensitivity analysis. In this cohort, there was 100% (RBD) and 93.5% (N) sensitivity at an endpoint titer cutoff of 300, and 94.8% (RBD) and 83.1% (N) sensitivity at an endpoint titer cutoff of 900 (Figure 1c,d). An endpoint titer cutoff of 900 was chosen for use in the RBD assay and an endpoint titer cutoff of 300 was chosen for use in the N assay, with a specificity at 97.8% for RBD and 94.5% for N, and a sensitivity at 94.8% for RBD and 93.5% for N. RBD and N ELISAs were further validated using WHO international standards (Supplementary Figure S1).

**Figure 1.** Specificity and sensitivity analysis of RBD and N ELISAs. Each circle represents one sample for RBD (**a**,**c**) and each square represents one sample for N (**b**,**d**) in each assay. Samples were grouped by positivity for a specific infection or assay (**a**,**b**) and see Supplemental Table S1 for details of "Others". Samples from patients with COVID-19 disease are grouped by the day post-self-reported-symptom onset (**c**,**d**). Dashed line indicates a titer at 900 and dotted line indicates a titer at 300. Numbers next to each line represents the specificity or sensitivity of the assay at the given cutoff value. Samples collected on or after Day 14 post symptom onset were used for sensitivity analysis.

#### *2.2. Seroprevalence of COVID-19 in Allegheny County, Western PA*

A total of 199 human blood samples were collected in the Fall of 2020 and 194 samples in February of 2021. A total of 88.5% of the study subjects were from Allegheny County and 9.9% were from other counties in PA (Supplementary Table S2). Women, seniors, and African Americans were more represented in the study cohort compared to that in the population of Allegheny County (Supplementary Table S2).

The endpoint titers for both RBD and N were determined for each sample and using predefined cutoffs, the seroprevalence of SARS-CoV-2 in the Fall cohort was 5.5% by RBD ELISA, 4.5% by N ELISA, and 2.5% for both. In the February cohort, the seroprevalence increased to 24.7% by RBD ELISA, 14.9% by N ELISA, and 12.9% for both (Table 1).

The 11 samples positive for RBD in Fall 2020 included two documented recovered cases of COVID-19 (one male aged 50–59 and one female aged 80+) and two volunteers who were enrolled in the Phase III clinical trial of the Moderna vaccine (Figure 2a).

The remaining seven samples positive for RBD in Fall 2020 were all from young adults aged 19–29, several of which had reported a history of either contact with COVID-19 confirmed cases, or mild COVID-19 related symptoms, but were never tested. This suggested a higher local prevalence of COVID-19 in young adults in Fall 2020 (Figure 2a).

49


**Table 1.** Analysis and comparison of seroprevalence and antibody neutralization between Fall 2020 and February 2021.

**Figure 2.** RBD and N endpoint titers of all samples. Each circle represents one sample for RBD (**a**,**c**) and each square represents one sample for N (**b**,**d**) at the two time points of the study. Dashed line indicates a titer at 900 as cutoff for RBD positive and dotted line indicates a titer at 300 as cutoff for N positive. Samples are grouped by age.

In the February cohort, the increased seroprevalence suggested increases in either acquisition of natural disease during the winter peak or vaccination following EUA for BNT162b2 by Pfizer and BioNTech on 11 December 2020 and mRNA-1273 by Moderna on 18 December 2020 (Table 1). Both BNT162b2 and mRNA-1273 are S protein mRNA-based vaccines, so vaccination would be expected to elicit RBD antibodies but not N antibodies. In contrast, infection by SARS-CoV-2 would trigger immune responses against both RBD and N. Therefore, based on their RBD and N titers, all 59 individuals who tested positive for RBD were separated into three groups, infected (*n* = 30, RBD+ and N+), vaccinated (*n* = 19, RBD+ and N−), or unclear (*n* = 10 RBD+ and N−) (Figure 3). These classification groups were informed by self-reporting or chart review. The group designated as unclear

had RBD titers that were either at or one dilution above the endpoint titer cutoff value and had no corroborating data from chart review or self-reporting.

**Figure 3.** Titer comparison between RBD ELISA, N ELISA, and Neutralization assay. RBD endpoint titer, N endpoint titer, and FRNT50 titer of all RBD positive samples in groups of infected (**a**), vaccinated (**b**), and unclear (**c**). Left y axis is Log10 endpoint titers for RBD and N ELISAs. Right y axis is Log2FRNT50 titer. Dotted line indicates a titer at 300 which was the cutoff titer for N positive. Dash-dotted line indicates a titer of 40 as cutoff for positive neutralization.

All RBD-positive samples were tested in an FRNT50 assay. The neutralization assay had a limit of detection of 1:20, and samples were considered positive if the titer was ≥40. Among the 30 individuals classified as infected, 24 (80%) were positive by FRNT50; the 6 that were unable to neutralize had an RBD titer ≤2700 (Figure 3a). In comparison, among the 19 individuals classified as vaccinated, 12 (63.2%) were positive by FRNT50; the 7 that were unable to neutralize had an RBD titer ≤8100 (Figure 3b). All unclear samples failed to neutralize the virus, even though they had a positive RBD titer (Figure 3c). Comparison between FRNT50 and ELISA titers revealed a significant correlation for RBD but not N (Figure 4a,b). Notably, samples with RBD titers at 8100 and positive N titers were more often able to neutralize SARS-CoV-2, than samples with RBD titers at 8100 but negative N titers, suggesting that despite having the same RBD titer, there might be a qualitative difference in spike antibodies generated during infection versus vaccination.

**Figure 4.** Correlation between ELISA titer and Neutralization titer. Samples from Figure 3 with titers above the detection threshold (100 for RBD and N, 20 for FRNT50) of each assay were selected for the correlation analysis between neutralization titer and RBD titer (**a**) or N titer (**b**). Spearman's Rank Correlation Coefficient r and Probability (p) value (two-tailed) are shown.
