# Tracing mumps transmission in Washington state

#### Louise H. Moncla*<sup>1</sup>, Allison Black*<sup>1,2</sup>, Misty Lang<sup>3</sup>, Nicholas R. Graff<sup>3</sup>, Ailyn C. Perez-Osorio<sup>3</sup>, Nicola F. Müller<sup>1</sup>,  Dirk Haselow<sup>4</sup>, Chas DeBolt<sup>3</sup>,Trevor Bedford<sup>1,2</sup>

<sup>1</sup>Fred Hutchinson Cancer Research Center, Seattle, Washington, United States
<sup>1</sup>University of Washington, Seattle, Washington, United States
<sup>3</sup>Office of Communicable Disease Epidemiology, Washington State Department of Health, United States
<sup>4</sup>Arkansas Department of Health, Little Rock, Arkansas, United States

<sup>*</sup>these authors contributed equally to this work 

## Abstract 
Mumps has resurged among vaccinated individuals in recent years, sparking an outbreak in 2016 and 2017 that affected nearly 12,000. Burgeoning evidence demonstrates that mumps vaccine-induced immunity wanes over time, triggering post-vaccine era outbreaks to which young adults are particularly susceptible. During the 2016/2017 outbreak, Washington state experienced one of the highest incidence rates of mumps infection, despite high overall vaccination rates. Curiously, Washington cases were enriched among school-aged children and members of a particular close-knit ethnic community, suggesting that waning immunity alone may not explain transmission. To elucidate how the mumps outbreak in Washington originated and spread, we sequenced 110 near-complete genomes of mumps viruses collected throughout the outbreak, and trace mumps transmission into and within the state. Contrary to contact tracing data, we uncover that mumps was introduced into Washington at least 13 times, primarily from Arkansas, sparking overlapping transmission chains. We do not find evidence for antigenic evolution, or that vaccination status or age were the primary determinants of transmission. Instead, the outbreak was overwhelmingly sustained by transmission within a close-knit ethnic community. Community members were twice as likely to transmit mumps as non-community members, and transmission to non-community members lead to short, terminal transmission chains. These data suggest that transmission was inefficient outside of close-contact settings. Our data underscore the ability of genomic data to clarify epidemiologic factors driving transmission, and pinpoint contact networks as critical determinants of mumps transmission in Washington. Our results suggest that in contact patterns should be considered along with waning immunity when formulating outbreak response and 3rd dose vaccine recommendations.  

## Overview
This repo contains all of the code used to perform the analyses for this project tracing mumps transmission during an outbreak in Washington. Sequence data generated for this project can be found [here](https://github.com/blab/mumps-seq/tree/master/data), along with the [protocols and pipelines](https://github.com/blab/mumps-seq) for amplifying and sequencing mumps genomes from buccal swabs. We have deposited all of the raw sequencing reads (with human reads removed) onto the Short Read Archive under Project number [PRJNA641715](https://www.ncbi.nlm.nih.gov/sra/?term=PRJNA641715). Consensus genomes are available in Genbank under accession numbers [MT859507-MT859672](https://www.ncbi.nlm.nih.gov/nuccore/?term=MT859507%3AMT859672%5Baccn%5D). 

Alignments and xml files used for beast analyses are available in the `phylogeography` and `structured-coalescent-analyses` directories. We currently are not able to publish full metadata that links the sequence strain name to age, vaccination status, and Marshallese status. We have therefore removed these metadata fields from the uploaded alignments and xml files and replaced them with generic placeholder text, like `vaccination_status`. If we receive permission from the Department of Health to release this metadata, we will update these files.

Similarly, we have a public metadata file with the date of sample collection, Ct value, specimen type, and vaccination status available [here](https://github.com/blab/mumps-wa-phylodynamics/blob/master/sample-metadata-public-2020-10-06.txt). If we are able to release further metadata, we will update this file with the appropriate fields.

Many of the plotting scripts in this repo require [baltic](https://github.com/evogytis/baltic). I have also included a version of baltic that will work with all of these jupyter notebooks within this repository.