Molecular evidence of SARS-CoV-2 in New York before the first pandemic wave

Abstract

Numerous reports document the spread of SARS-CoV-2, but there is limited information on its introduction before the identification of a local case. This may lead to incorrect assumptions when modeling viral origins and transmission. Here, we utilize a sample pooling strategy to screen for previously undetected SARS-CoV-2 in de-identified, respiratory pathogen-negative nasopharyngeal specimens from 3,040 patients across the Mount Sinai Health System in New York. The patients had been previously evaluated for respiratory symptoms or influenza-like illness during the first 10 weeks of 2020. We identify SARS-CoV-2 RNA from specimens collected as early as 25 January 2020, and complete SARS-CoV-2 genome sequences from multiple pools of samples collected between late February and early March, documenting an increase prior to the later surge. Our results provide evidence of sporadic SARS-CoV-2 infections a full month before both the first officially documented case and emergence of New York as a COVID-19 epicenter in March 2020.

Fig. 1. Detection of SARS-CoV-2 nucleic acids in nasopharyngeal specimens collected in the first 10 weeks of 2020.

a Schematic representation of the study design. Nasopharyngeal swab specimens that tested negative for respiratory pathogens (RPN) were pooled. Each pool consisted of ten specimens from the same week from one of five hospital sites. Nucleic acid amplification testing (NAAT) was performed and RNA was processed for SARS-CoV-2 genome assembly. b Select events and responses to the evolving SARS-CoV-2 pandemic are annotated over the timeframe surveyed. Confirmed cases in NYC for the last 2 weeks are noted. Absolute counts of pools that tested positive or negative for RT-PCR targets (ORF1ab+E+ (magenta), ORF1ab+ (yellow), E+ (cyan), Negative (dark purple)) are depicted by week collected. c Distribution of pools with RT-PCR target results (ORF1ab+E+ (solid), ORF1ab+ (dotted), E+ (cross-hatched)) across the five different hospital sites in NYC (Hospital A, blue; B, red; C, green; D, purple; E, orange). d Distribution of SARS-CoV-2 sequences recovered by collection week and hospital site of RPN pools. Filled points reflect complete SARS-CoV-2 consensus genomes recovered and points with X’s reflect partial genomes recovered (e.g., incomplete genomes and those validated by SARS-CoV-2 reads). Colors denote hospital sites as indicated by the legend in (c).

Fig. 2. Phylogenetic relationships of previously undetected SARS-CoV-2 and other NY and global isolates.

a Multiple sequence alignment of >95% complete SARS-CoV-2 genome sequences obtained from RPN pools relative to Wuhan-Hu-1 (RefSeq: NC_045512). RPN pools are ordered by date and PANGO lineage as displayed in (a). The SARS-CoV-2 genome coordinates and gene annotations are shown above. Single- nucleotide variations (SNVs) are depicted with vertical lines in red (clade defining) or black (other). b Coverage for pools with detectable RT-PCR targets (ORF1ab+E+ (magenta), ORF1ab+ (yellow), E+ (cyan)) collected prior to the first confirmed case in NY (NY1) with detectable SARS-CoV-2 reads that could not be assembled to complete genomes (>Q30 reads are shown). Nextera XT comprises data from both whole-genome and targeted amplicon sequencing library preparations. c Maximum-likelihood (ML) phylogenetic inference shown as a time tree of seven SARS-CoV-2 genome sequences from this surveillance study in a global background of 2993. Tip circles indicate the position of the respiratory pathogen-negative (RPN) pools (red) described in this report, the first reported COVID-19 case in NYC (green) from 29 February, later NYC cases from MSHS (yellow) and other institutions (dark gray), and US (blue) early isolates prior to 1 March. Tips without circles correspond to the background global isolates. The yellow box delineates the position of the clade containing the majority of NYC sequences detected during the early spread. The PANGO lineage classification of the RPN pools is indicated on the right, and the NextStrain clades are shown as node labels. The specimen identifier is indicated for RPN pools detected earlier than NY1. The time tree was inferred under a strict clock model with a nucleotide substitution rate of 0.80 × 10⁻³.