Impacts of the gut microbiome on infant cognition and behavior

Repository Summary. Dataset and code related to a study of how the infant gut microbiome impacts infant brain cognition and behavior. 

Project Summary. Data is related to a project which investigated the relationship of the early-life gut microbiome to brain function and development, using infant fecal samples collected by the COMBINE study (Cork, Ireland). These samples were further used to establish preclinical models, including humanized-microbiota mice and bioreactors, to study the taxonomic and functional impact of the microbiome on infant behavior and metabolic health. 

Data Collection. This dataset accompanies an article which will be available as a preprint on bioRxiv under the title "Infant gut microbiomes contribute to metabolic states that impact brain function". The Methods section in this article fully describes the biological nature of these samples and how the samples were processed and analyzed. Raw sequencing data (whole-genome shotgun sequencing) is available on NCBI, see PRJNA1413407.

Repository Content.

• General note on folder organization. Each folder has a name that indicates the type of method used and/or the data used/included. Folder contains code, input, and the outputs related to a specific computational experiment of visualization. Some folders are standalone, while others read data from other folders or export data into other folders.

• Which data sets are shared in this repository? Several folders contain tabular or summary data, including ones generated or derived from sequencing data: 
  • "demographic-table-create" has baseline clinical/demographic data and cognitive assessment scores used in the related study. 
  • "seq-meta-data-tidy" has sample meta-data, sequencing meta-data, and summary results for mouse behavioral assessment, mouse brain immunohistochemistry/imaging, and mouse biomarkers. The mouse data can be used directly to generate related figures included in the accompanying manuscript. 
  • "repo_data" is the main repository of data including taxonomic tables, functional table, and other genome-scale summaries. 
  • "metabolomics-data-tidy" has untargeted metabolomics data related to humanized-microbiota mice and mini-bioreactor array experiments. 
  • "FBA-data-tidy" has results from microbial community-scale metbaolic modeling generated using MICOM on publicly available genome-scale metabolic models.

• What type of analyses are shared in this repostiroy? 
  • Upstream bioinformatics workflows in this repositroy include: 
    • Analysis of whole-genome shotgun sequencing with biobakery tools including KneadData, MetaPhlAn, and HUMAnN. 
    • Assembly and annotation of metagenome-associated genomes (MAGs) with MetaWRAP and bakta. 
    • Genomic mining of MAGs with GapMind, or gutSMSH and BiG-MAP.
    • Microbial community-scale meatbolic modeling with MICOM.
  • Downstream statistical tests and analysis include:
    • Differential abundance analysis with LinDA, MaAsLin2, or linear mixed effects models (LMM).
    • Alpha-diversity and beta-diversity calculations and analysis with univariate tests, LMMs, and PERMANOVA.
    • Correlation analysis using Spearman's correlation or LMMs. 
    • Gene/Feature set enrichment analysis. 

• What do you need for performing the analyses? All downstream anlaytical scripts are in bash, python, and R either standalone or in a Jupyter Notebook (*.ipynb). Upstream analysis was performed mostly on Nextflow workflows on a high-performing computing (HPC) cluster. 
  • The "downstream-conda-environments" describes four conda environments used for downstream analysis including the following. Each folder has a YAML file and file that lists all package versions. 
    • coffee, this is a single environment for all Python-based analyses.
    • r-linda, an environment to run LinDA in R, for differentiabl species abundance analysis
    • r-maaslin2, an environment to run MaAsLin2 in R, for functional abundance analysis
    • r-vegan, an environment to run Vegan in R, for microbial diversity analysis
  • The "bioinforamtic workflows" contains the nextflow files for upstream analyses.