Skip to main content
Premium Trial:

Request an Annual Quote

NIST Analysis Discovers Sample Processing Bias in Metagenomic Wastewater Surveillance


BALTIMORE – A systematic analysis of publicly available wastewater metagenomic data has revealed that variations among sample processing workflows can skew the results of antibiotic resistance gene (ARG) surveillance.

Led by researchers from the National Institute of Standards and Technology (NIST), the study underscores the need to standardize sampling protocols to minimize biases when using wastewater for ARG monitoring.

The impetus for the analysis was "to see how differential workflow biases might alter the antimicrobial resistance surveillance conclusions," said Ishi Keenum, a postdoctoral researcher in NIST’s Complex Microbial Systems Group, who presented the finding at this year’s American Society for Microbiology Conference on Rapid Applied Microbial Next-Generation Sequencing and Bioinformatic Pipelines (ASM NGS) on Tuesday. "Once you know where you are getting the most bias in your workflow, then you can really zoom in on that workflow to fix it," she added.

According to Keenum, there are several advantages to using shotgun metagenomic sequencing for analyzing the ARG landscape in community wastewater samples, which has increasingly become a common public health surveillance tool. One, she noted, is that unlike PCR-based methods, which can only target a limited number of genes, shotgun metagenomics allows researchers to simultaneously detect ARGs, mobile genetic elements, and taxonomy. In addition, the approach can tackle non-culturable organisms, plus the sequencing data can be used for future reanalysis.

To identify potential biases in wastewater ARG surveillance, Keenum and her team examined 1,000 global wastewater metagenomes from publicly available datasets as part of a systematic literature review. The researchers extracted metadata on these samples from the National Center for Biotechnology Information (NCBI) and other relevant publications while keeping track of the different DNA extraction methods used in their protocols.

Ideally, differences seen in the microbiome and resistome signatures should be primarily driven by sample location and timing and less affected by sample processing, the researchers said. However, their meta-analysis revealed that while sampling location was overall the primary driver for the varied microbiome and resistome profiles, DNA extraction protocols also notably influenced the surveillance results.

In her presentation, Keenum showcased two geographic regions, one in Sweden and one in the US, where multiple wastewater ARG surveillance studies had been conducted using different sampling workflows. Despite the relative proximity of some of the sampling sites within each region, different DNA isolation schemes seemingly resulted in "significantly different" microbiome annotations, Keenum said, as indicated by the divergent Firmicutes to Bacteroidetes ratios — a hallmark of microbiota composition — of the samples.

To combat such biases, Keenum proposed that spiking in reference materials during metagenomic sample processing "is the way to go." By analyzing the recovery rates of different types of organisms in the reference material, researchers can better gauge the sampling bias introduced to the study, she said. Previous studies that already have DNA extracted, Keenum said, can include synthetic genes as a reference control.

"What I think is really important is that when we don't include standards to measure recovery, we actually don't know what we're missing," she said.