Lessons learned in virulence factor identification and data management from a hackathon on microbial virulence

Brett E Pickett,Allissa Dillman,William Klimke,Vadim Zalunin,Kumardeep Chaudhary,Defne Surujon,Vyacheslav Brover,S Elizabeth Norred,Moamen M Elmassry,Sherry Bhalla,Xin Huang,Eric Holloway,Jonathan Parkinson,Ben Busby,Ryan Connor,Samantha Sevilla,Udana Torian,Michael Feldgarden,Tamiru Berhanu-Denka,Ousmane Cissé ,Mónica García Solache ,Michael Chambers ,Catarina Mendes

doi:10.12688/f1000research.26452.1

Abstract

Virulence is a complex mix of microbial traits and host susceptibility that could ultimately lead to disease. The increased prevalence of multidrug resistant infections complicates treatment options, augmenting the need for developing robust computational methods and pipelines that enable researchers and clinicians to rapidly identify the underlying mechanism(s) of virulence in any given sample/isolate. Consequently, the National Center for Biotechnology and Information at the National Institutes of Health hosted an in-person hackathon in Bethesda, Maryland during July 2019 to assist with developing cloud-based methods to reduce reliance on local computational infrastructure. Groups of attendees were assigned tasks that are relevant to identifying relevant tools, constructing pipelines capable of identifying microbial virulence factors, and managing the associated data and metadata. Specifically, the assigned tasks consisted of the following: data indexing, metabolic functions, virulence factors, antimicrobial resistance, mobile elements in enterococci, and metatranscriptomics. The cloud-based framework established by this hackathon can be augmented and built upon by the research community to aid in the rapid identification of microbial virulence factors.

Highlights

A variety of publications have reported the diverse mechanisms of microbial virulence (Leshem et al 2020; Tarsillo and Priefer 2020; Nogueira et al 2019; Saeki et al 2020; Lange et al 2019; Saiardi et al 2018; Russo and Marr 2019; Feßler et al 2018; Geisinger and Isberg 2017; Schroeder et al 2017; Diard and Hardt 2017)
Metabolic functions BioProjects within the National Center for Biotechnology and Information (NCBI)-hosted Sequence Read Archive were selected for analysis based upon the presence of shotgun sequencing metagenomic samples from the gut microbiome and paired labels identifying the sample with a health status
This work describes our efforts to demonstrate the usefulness of using existing cloud-based platforms to perform analytical tasks related to identifying markers of microbial virulence, and to use these platforms to store the relevant data and metadata

Summary

Introduction

A variety of publications have reported the diverse mechanisms of microbial virulence (Leshem et al 2020; Tarsillo and Priefer 2020; Nogueira et al 2019; Saeki et al 2020; Lange et al 2019; Saiardi et al 2018; Russo and Marr 2019; Feßler et al 2018; Geisinger and Isberg 2017; Schroeder et al 2017; Diard and Hardt 2017). In the vast majority of cases, one or more features/biomarkers can be identified and compared against databases of known virulence factors. Examples of such features can include genetic (e.g. pathogenicity islands, antimicrobial resistance (AMR) genes, mobile elements) (Kaushik et al 2018; Vestergaard et al 2019; Madec et al 2017) and metabolomic (e.g. carbohydrates, lipids, small molecules) (Jia et al 2019; Lloyd-Price et al 2019) features. (July 2019), a hackathon was convened to develop methods to detect various aspects of microbial virulence. The pipelines and methods that were generated during this hackathon were made publicly available to rapidly disseminate these tools and to enable the rapid identification of various microbial virulence factors

Objectives

Methods

Results

Discussion

Conclusion