Tracking SARS-CoV-2 genomic variants in wastewater sequencing data with LolliPop

Area of Science:

• Environmental microbiology
• Computational biology
• Epidemiology

Background:

• Wastewater-based epidemiology (WBE) is a key tool for pathogen surveillance, offering advantages over clinical methods.
• WBE provides unbiased estimates and early detection of viral loads and variant outbreaks.
• Computational challenges exist in analyzing WBE data, particularly for variant deconvolution.

Purpose of the Study:

• To develop a computational method for estimating the relative abundances of genomic variants in mixed wastewater samples.
• To address the variant deconvolution problem using next-generation sequencing data from wastewater.

Main Methods:

• Introduction of LolliPop, a novel computational method for variant deconvolution.
• Application of temporal regularization (fused ridge penalty) tailored for wastewater time series sequencing data.
• Utilizing bootstrap and developing analytical standard errors for confidence intervals.

Main Results:

• LolliPop effectively estimates variant abundances from wastewater sequencing data.
• Temporal regularization enhances robustness to high levels of missing data common in wastewater samples.
• Analytical standard errors provide similar confidence intervals to bootstrap at lower computational cost.

Conclusions:

• LolliPop is a robust computational tool for variant deconvolution in wastewater surveillance.
• The method improves the reliability and efficiency of analyzing wastewater sequencing data.
• LolliPop can be applied to real-world surveillance data, as demonstrated with Swiss wastewater data.