A package for the meta-analysis of shotgun datasets
The link to manuscript is here. Only individuals with access can use this link.
This repository will contain meta-analysis contents and concepts for publicly available shotgun metagenomics datasets. Part of the difficulty in gathering all is that metadata is almost always omitted, making incomplete datasets virtually useless.
Data were collected with the scope of finding Parkinson disease datasets with publicly available FASTQ and per-sample metadata.
Data were uniformly processed with curatedMetagenomicNextflow and stored in Google
Cloud bucket. Parquet files were generated and hosted on huggingface.
Data retrieval is made available via parkinsonsMetagenomicData, a convenience
R packages that expands functionalities of curatedMetagenomicData.
First, to use gene families, one must download the data locally
curl https://huggingface.co/datasets/waldronlab/metagenomics_mac/resolve/main/genefamilies_unstratified_uuid.parquet -o ~/Downloads/genefamilies_unstratified.parquet
quarto render 02-download-data.qmd -P pMD_data_type:relative_abundance
quarto render 02-download-data.qmd -P pMD_data_type:pathabundance_unstratified
# this last step only works in machines with large amounts of memory (> 64G)
quarto render 02-download-data.qmd -P genefamilies_unstratified
quarto render 03-Main-Analysis.qmd -P pMD_data_type:relative_abundance
quarto render 03-Main-Analysis.qmd -P pMD_data_type:pathabundance_unstratified
quarto render healthy-constipated-analysis.qmd -P pMD_data_type:relative_abundance
quarto render healthy-constipated-analysis.qmd -P pMD_data_type:pathabundance_unstratified