DRiDO Microbiome Study

Author: Lev Litichevskiy
Last updated: October 28, 2024

This repository contains code and data that can be used to reproduce figures from the Dietary Restriction in Diversity Outbred mice (DRiDO) microbiome manuscript. The starting point for this repository is summarized tables of taxonomic and functional classification results (not fastq files).

1. Citation

L Litichevskiy, M Considine, J Gill, V Shandar, … A Di Francesco, GA Churchill, M Li, CA Thaiss. Interactions between the gut microbiome, dietary restriction, and aging in genetically diverse mice. https://www.biorxiv.org/content/10.1101/2023.11.28.568137

2. Quick start

Data

• Taxonomy: data/kraken_matrix_agg_by_stool_ID_n1303x2997.txt with data/kraken_taxonomy_n1303.txt
  • Absolute counts
• Pathways: data/pathabundance_tpm_agg_by_stool_ID_n422x2997.txt
  • TPM (transcripts-per-million) abundances
  • That is, normalized for gene length and sequencing depth

Tutorial

This tutorial demonstrates how to import taxonomic data and perform several basic analyses.

3. Repository organization

• analysis contains .Rmd notebooks used for generating figures
• plots contains figures, i.e. the output of analysis
• scripts contains a mix of .R and .Rmd files used for data processing and running linear models
• results contains the output of scripts
• data contains the inputs to scripts and analysis, including metadata

See here for more details about the overall workflow.

See here for which script was used to produce every figure panel in the manuscript.

Note that there are multiple layers of metadata: sequencing metadata, library metadata, and stool metadata are stored separately. Multiple sequencing IDs (seq.ID) can correspond to the same library ID (lib.ID), and multiple library IDs can correspond to the same stool sample (stool.ID). Mice contributed one or more stool samples.

4. System requirements

This code was run on macOS Big Sur using R v4.2.2. All R packages are available from CRAN or Bioconductor — except for ASReml, which requires a license. ASReml was used for estimating heritability and running linear mixed models. Identical results can be produced using the lme4qtl package (see run_lme4qtl.R for an example).

All analyses except for mediation and QTL mapping were run on a laptop. Mediation analysis was performed on a cluster using Snakemake (Snakefile_mediation, run_mediation_one_diet_one_pheno.R), and QTL mapping was performed on a cluster using R/qtl2 (run_genetic_mapping_rqtl2.R).

5. QTL mapping

Karl Broman’s R/qtl2 was written specifically to handle multi-parent QTL mapping crosses such as DO mice.

QTL mapping was performed as described in Zhang et al., Genetics, 2022 (“Genetic linkage analysis”).

Input files not included in this repo:

1. cc_variants.sqlite: Imputed variants from DO founders. Download here.

2. prob.8state.allele.qtl2_200131.Rdata: 8 state allele probabilities for DRiDO mice. Download here.