MetaNutriUnify Collection

This work is linked to the iTARGET project (https://qualiment.fr/des-projets-pour-anticiper-les-besoins-de-recherche-des-entreprises-agroalimentaires-2022/hot-topics/), aiming at performing in silico and in vitro targeting of healthy gut bacteria with fiber degrading metabolic potential. In this context, we developed MetaNutriUnify, the first collection of curated and harmonized metagenomic data with unified nutritional data from public study cohorts, with a particular attention on fiber.

Overview of the public studies included in MetaNutriUnify MetaNutriUnify is composed of 21 harmonized and curated public projects with available shotgun metagenomes from human adults stool samples and nutritional and anthropometric data. It consists of 949 individuals from 15 countries, totalizing 1656 metagenomes, from which we generated microbial species and associated functional modules abundance tables. We also unified nutritional data, including diet type, study type (observational/interventional), time points for stool sampling, diet intervention and associated information, macro and micronutrients when available, and reported available anthropometric data (gender, country, age, weight, height, BMI).

The list of the 21 included studies is:

Bioproject PRJEB8249 (2015, SWE, 21 subjects, PMID 26244932) Bioproject PRJNA278393 (2015, TZA & ITA, 33 subjects, PMID 25981789) Bioproject PRJNA328899 (2016, MNG & CHN, 110 subjects, PMID 27708392) Bioproject PRJNA305507 (2017, USA, 33 subjects, PMID 28797298) Bioproject PRJEB28687 (2018, USA & THA, 50 subjects, PMID 30388453) Bioproject PRJEB32794 (2019, IRL, 37 subjects, PMID 31558359) Bioproject PRJNA472785 (2019, USA, 12 subjects, PMID 31235964) Bioproject PRJNA386503 (2019, USA, 4 subjects, PMID 30810441) Bioproject PRJNA397112 (2019, IND, 88 subjects, PMID 30698687) Bioproject PRJEB33500 (2020, ITA, 82 subjects, PMID 32075887) Bioproject PRJNA647720 (2021, USA, 20 subjects, PMID 33727392) Bioproject PRJNA755720 (2021, ESP, 20 subjects, PMID 34444797) Bioproject PRJNA892265 (2022, ESP, 20 subjects, PMID 36364873) Bioproject PRJEB42906 (2022, USA, 50 subjects, PMID 35312171) Bioproject PRJEB45944 (2022, NLD, 149 subjects, PMID 35115599) Bioproject PRJEB48663 (2022, FRA, 39 subjects, PMID 35311446) Bioproject PRJNA762543 (2022, SGP, 62 subjects, PMID 35549618) Bioproject PRJEB48605 (2023, DEU, 68 subjects, PMID 35760036) Bioproject PRJNA939268 (2023, SGP, 10 subjects, PMID 36997838) Bioproject PRJEB26842 (2023, GBR, 29 subjects, PMID 37587110) Bioproject PRJNA906167 (2023, ESP, 12 subjects, PMID 37457982)

Method summary Metagenomic data and associated metadata were recovered from the European Nucleotide Archive, while nutritional and anthropometric data were collected from various online resources (main publication, supplementary files, GitHub or BioProject information). QC validation was performed using fastp (version 0.23.4) and host related reads were filtered out with bowtie using the human reference genome (Homo sapiens T2T-CHM13v2.0). Resulting high quality reads were mapped onto the 10.4 million gut IGC2 catalogue of the human microbiome and onto the 8.4 million human oral microbial catalogue using the METEOR software clustered into Metagenomic Species Pangenomes (MSP species) that were previously taxonomically and functionally annotated.

MetaNutriUnify characteristics The provided data consists of:

    Metagenomic Species Pangenomes (MSP) species abundance table and related GTDB taxonomy 
        (GTBD-tk version r220) (final_msp.7z and species_taxonomy_20241119.tab)
    KEGG, GMM and GBM Functional modules abundance table and related modules definition 
        (KEGG version 107, GMM modules and GBM modules) (final_modules.tab and all_modules_definition_GMM_GBM_KEGG_107_20241119.tab)
    Manually curated and unified data collected from bioprojects (final_metadata.tab):

            Metadata from the metagenomes obtained from ENA, as well as nutritional and anthropometric data.
            Additional data on each sample, such as the evaluation of cross contamination within each bioproject using the CroCodeEL tool, together with the number of high quality reads, the MSP species richness and if the number of reads was below 1M. 
                We proposed a “to_exclude” variable in the deposited MetaNutriUnify file, derived from these data. 
                If one of the following conditions was met: low_read = YES, is.contaminated = YES or MSP_richness < 20, we propose to exclude the sample from downstream analysis.

            Nutritional data, carefully extracted from each study and reporting information on diet type, energy, macro- and micronutrients when available. 
                No frequency data were included, because of the great variability between studies. We reported variables and their modalities as they were described in the different studies, 
                and only modified units of nutritional data when appropriate. We encourage users to modulate the modalities for some variables, such as time point description, and to refer to the original studies for any enquiries.

            Anthropometric data, including country, gender, age, weight, height and BMI.


    General information on the cohorts, such as reference, study design, country, number of subjects and metagenomes, diet type and reported nutritional records (cohorts.xlsx)
    Legends for cohorts and final_metadata files (cohorts.xlsx)

Codes for data recovery and harmonization are available here.

Funding The research leading to these results has received funding under grant (iTARGET; doi.org/10.17180/h5gd-gk88) from Carnot Qualiment© supported by Agence Nationale de la Recherche. Additional funding was from the MetaGenoPolis grant ANR-11-DPBS-0001.