Tracking and surveillance SARS-CoV-2 variant-associated mutations from wastewater whole genome sequencing in Colorado

Background: Bioinformatic analysis of whole genome sequencing data has played a central role in surveillance throughout the SARS-CoV-2 pandemic by allowing scientists to identify novel variants and to track the proportions of variants across time and geographic location. Because SARS-CoV-2 is shed in fecal material that ends up in wastewater, whole genome sequencing of SARS-CoV-2 from wastewater samples can be an effective method for the surveillance of the virus. Methods: Here, we describe Colorado's wastewater whole genome sequencing and surveillance strategy for SARS-CoV-2. We have been collecting twice weekly samples from March of 2021 to present from 20 wastewater utility sites. After extraction of RNA from wastewater, we perform tiled PCR-amplicon whole genome sequencing for SARS-CoV-2. We also present the bioinformatics workflow that we developed for analyzing whole genome sequencing data from wastewater samples, which determines the variants likely present in a sample based on sets of mutations associated with variants of concern or interest. Variant-associated mutations are determined using the Spike mutations present in known variants of concern (VOC) or variants of interest (VOI) as defined in the outbreak.info database (https://outbreak.info/situation-reports) and are updated regularly. Briefly, this workflow performs quality control and filtering of raw reads, aligns reads to the reference genome, calls single nucleotide polymorphisms (SNPs) using FreeBayes and filters SNPs based on quality. We then use custom scripts to identify the presence and frequency of mutations associated with known VOCs and VOIs and to generate summary files for further analysis and visualization. We have made this workflow a publicly available resource for use on Google Cloud's Terra platform. Results: Using this surveillance and analysis strategy we have been able to determine which SARS-CoV-2 variants are likely present in wastewater samples based on the constellations of VOC/VOI-associated mutations that are present. For example, we observed the rapid transition from Alpha to Delta as the dominant variant circulating in Colorado (Fig. 1A). Furthermore, we can see the emergence of VOCs in CO prior to their emergence in clinical samples sequenced by either the state public health department or a commercial laboratory and reported to the state public health department. As shown in Fig. 1B, in most of the counties for which we have sampling, we identified the mutations associated with the Delta variant in wastewater prior to it being identified in clinical samples. These results also show geographic differences in which variants are present, when they are first detected in wastewater, and when they are first detected in clinical samples. Conclusions: These results demonstrate that SARS-CoV-2 wastewater sequencing can be used to monitor SARS-CoV-2 trends over geographic location and over time. Our ability to detect VOC/VOI-associated mutations prior to the detection of the viral lineages with which they are associated in clinical samples suggests that wastewater whole genome sequencing can be used as an early warning signal. More broadly, these results can inform public health strategies and can be applied to other pathogens of current or future public health concern.