MASCOT-DataStreams (MASCOT-DS) is a new BEAST2 software package that extends the structured coalescent model MASCOT to enable joint inference of prevalence over time and transmission (migration) rates between locations using multiple epidemiological data streams. The tool addresses the challenge that common surveillance sources—reported case counts, wastewater viral concentrations, seroprevalence surveys, and pathogen genomic sequences—are each imperfect and capture different aspects of underlying transmission processes. By integrating these complementary sources within a unified phylodynamic framework, MASCOT-DS aims to improve accuracy and certainty of transmission dynamics estimates.
MASCOT-DS can incorporate any combination of the following inputs:
Each data type contributes distinct information: serology helps scale the outbreak magnitude, genomic data inform between-location transmission, and case counts and wastewater primarily capture outbreak shape.
MASCOT-DS is implemented as a BEAST2 package that extends the MASCOT structured coalescent. The preprint reports that the software jointly infers prevalence trajectories and between-location migration rates from the supplied combination of data streams. The authors provide scripts and instructions for reproducing results in public repositories: https://github.com/Pweidemueller/MASCOT-DS_materials, https://github.com/Pweidemueller/Mascot_datastreams, and related GitHub links cited in the source.
The authors evaluated MASCOT-DS using simulated outbreaks in structured populations. According to the preprint, these simulations show that MASCOT-DS accurately recovers true prevalence trajectories and estimates migration rates between locations. Details of simulation design, parameter settings, and quantitative performance metrics were reported in the source material and in the supplementary code repositories.
MASCOT-DS was applied to a multi-stream dataset from the SARS-CoV-2 Epsilon wave (winter 2020–21) in three San Francisco Bay Area counties. The analysis combined genomic sequences, case counts, wastewater concentrations, and seroprevalence data to reconstruct county-level prevalence dynamics and to quantify transmission both within the region and into it. The integrated analysis produced county-level prevalence reconstructions and migration estimates that leveraged information across data types.
The authors performed systematic analyses in which individual data streams were removed to assess their unique contributions. Key findings reported in the preprint include:
These observations indicate that combining streams yields more certain estimates than any single data source alone.
MASCOT-DS provides a framework for formally evaluating and combining diverse surveillance streams in a phylodynamic model. The approach demonstrates that integrating complementary data increases certainty in inferred transmission dynamics and can reveal which surveillance modalities are most informative for particular inference goals (for example, migration estimation versus prevalence scaling). The preprint suggests this framework can be used to assess the added value of surveillance strategies and to guide data collection priorities.
The source is a preprint posted on medRxiv and includes standard declarations: the authors report no competing interests and state that relevant ethical approvals and participant consent were obtained. The preprint provides code and materials in public repositories to enable reproduction. Specific methodological details, quantitative performance results, and full implementation settings are provided in the paper and linked repositories; if additional specifics are required, the reader should consult those materials directly.