Sharing Environmental Microbiome Data

Collaborating Institutions: 

Duke University and University of Iowa (U of I) SRP Centers

Abstract:

Studies involving microorganisms that break down pollutants are often conducted with single microbial cultures or simplified bacterial communities in the laboratory. While these experiments are useful to uncover mechanistic insights, they do not capture the complex microbial interactions that exist in nature. Researchers at the Duke University and University of Iowa SRP Centers collaborated to establish a common computational framework and infrastructure, and to standardize sampling approaches to better share, integrate, and analyze large microbiome sequencing datasets collected from natural sediments across SRP centers. Both Duke and the University of Iowa are working to engineer microbes to clean up contaminants in the environment. The team hoped that by sharing data, they could shed light on how complex populations of microbes interact within an environment to provide useful information to improve bioremediation strategies.

The Use Case first sought to establish a reproducible and sharable microbiome bioinformatics pipeline focusing on their existing 16S rRNA high throughput sequencing data from PAH- and PCB-contaminated sediments, respectively, which could then be applied to other data types.

Research Question: 

Can we establish a common computational framework, infrastructure, and standardized sampling approaches to enable data sharing, integration, and analysis of environmental microbiome data to facilitate precise bioremediation?  

Data Sources and Data Sets: 

Data Repositories: 

NCBI’s Sequence Read Archive (SRA); Metadata accompanying environmental microbiome raw sequence data submitted to SRA will be improved in near future. 

Metadata:

Vocabularies Utilized: 

Did not use any ontologies. They intended to use the Earth Microbiome Project Ontology (EMPO) however it was limited with respect to describing contaminated environments. 

Approach:  

The team sought to establish a reproducible and sharable microbiome bioinformatics pipeline focusing on their existing 16S rRNA high throughput sequencing data. All of the team’s experimental protocols and analysis code were uploaded to a GitHub repository and included version control to ensure methods were up to date as techniques and tools evolved. They evaluated computational procedures looking for ways maximize reproducibility and FAIRness of data management practices.

The team also developed a software container, which is a standalone, transferable, and executable assembly of software, to run their microbiome analyses. Their computing environment is in a container on the Singularity Hub and has all the software packages with source coded needed to run analyses, such as more than 60 R packages pre-installed, which greatly reduces the amount of time required for someone to develop and install the analysis pipeline.

Outcomes:  

The team developed a Docker container (i.e., standalone, transferable and executable assemblies of software) to house their microbiome analysis pipeline (specifically for 16S rRNA) so anyone on EUC team could run their analyses reproducibly from any location. Also developed protocols for standardizing environmental sampling from soil, sequencing (e.g., DNA extraction and library preparation) and analysis, and visualization of environmental microbiome data which are stored on GitHub providing version control ensuring that even as protocols improve and change over time, the original methodology remains available. Both the container, protocols, and guidance documents will be made available in the future. 

Integrated Datasets, Portals/Dashboards, Tools, Code: 

Their Singularity containers are available online, instructions for running the containers are on GitHub, and their data is stored in Sequence Read Archive (SRA; e.g., BioProject Number PRJNA382682).

Collaborating Institutions:

Integrating and Creating Broad Access for Transcriptomic, Proteomic, Microbiome and Physicochemical Datasets of Phytoremediator and Phytostabilizer Plants: University of Arizona (UA) and University of California (UC) San Diego SRP Centers

Data Interoperability for Investigating Biogeochemical Controls on Metal Mixture Toxicity Using Stable Isotopes and Gene Expression: UA SRP Center and researchers at the Colorado School of Mines (CSM)

Abstract:

Integrating and Creating Broad Access for Transcriptomic, Proteomic, Microbiome and Physicochemical Datasets of Phytoremediator and Phytostabilizer Plants: Soil and water with high levels of toxic metals, including cadmium, lead, mercury, and arsenic, can be harmful to human and environmental health. Traditional approaches to decontaminate heavy metals include excavating and removing soils, which can be costly and impractical. Using plants to take up metals and stabilize them is a cost-effective alternative, but there are many complex interactions and genes and pathways involved that are not well characterized. The University of Arizona (UA) SRP Center team worked with researchers from the University of California (UC) San Diego SRP Center to understand interactions between metals, microbes, and plants that help some plants tolerate contaminants and stabilize metals in contaminated soils. Specifically, they sought to investigate genomic, transcriptomic, microbiome, and physicochemical properties to identifying genes and pathways that enable plants to grow and stabilize metals in semi-arid environments.

Data Interoperability for Investigating Biogeochemical Controls on Metal Mixture Toxicity Using Stable Isotopes and Gene Expression: In a closely related Use Case, the UA SRP Center team collaborated with researchers from an SRP-funded individual research project at the Colorado School of Mines (CSM) to explore how remediation of mining waste affects diversity in terrestrial and aquatic systems. Researchers at CSM examined stream impacts of mining and the processes involved in recovery following clean-up, however complex interactions among metal mixtures are not well captured by current predictive toxicity models used by regulatory agencies. The group conducted toxicology experiments with aquatic organisms to uncover the molecular mechanisms involved in interactions between metals to improve model predictions. The UA team has tackled the issue of mining waste on the terrestrial side, looking at plants that can take up and stabilizing metals. They have tested a phytoremediation strategy called compost-assisted phytostabilization that promotes plant and root growth that locks metals underground.

Research Questions:

Can we combine data to better understand interactions between metals, microbes, and plants that help some plants tolerate contaminants and stabilize metals in contaminated soils? Can combining toxicity data in terrestrial and aquatic systems reveal new insights in predicting recovery after cleanup? Can we create a data analysis portal to enable robust analyses that shed light on the complex relationships between environmental factors, microbial communities, and remediation success?

Data Sources and Data Sets:

Data Repositories:

Sequence data for both EUCs is being deposited in NCBI ‘s SRA using minimum information standards (i.e., Minimum Information about any (x) sequence -MixS), while environmental data will be deposited in the Knowledge Network for Biocomplexity (KNB) data repository.

Metadata:

Vocabularies Utilized: 

Environment Ontology (EnvO) for metal concentrations, Plant Ontology for plant structures, BioCollections Ontology for observations and measurements, and Population and Community Ontology for diversity metrics.

Approach:

The research team focused on making data more standardized and interoperable, mapping terms to existing ontologies and contributing to ontologies by adding more terms and details as needed. The team is still in the process of analyzing and annotating transcriptomic datasets to then integrate that data into a single graph database hosted on the UC San Diego Superfund Portal. 

Outcomes:

The team enhanced existing repositories and made progress in making data harmonized and interoperable.

Publications:

Ramírez-Andreotta MD, Walls R, Youens-Clark K, Blumberg K, Isaacs KE, Kaufmann D, Maier RM. 2021. Alleviating Environmental Health Disparities Through Community Science and Data Integration. Front. Sustain. Food Syst 5.

Integrated Datasets, Portals/Dashboards, Tools, Code: 

CSM data on water chemistry and benthic invertebrate diversity along with preprocessing scripts are available on GitHub.