https://doi.org/10.5061/dryad.dbrv15f9z

This dataset contains:

• Pre-formatted selected Taxonomic profiles from stool metagenomic sequencing obtaine from a combination of:
  • 8 previously-published and publicly-available studies: Asnicar, F. et al. (2017) [1], Backhed, F. et al. (2015) [2], Kostic, A. D. et al. (2015) [3], Pehrsson, E. et al. (2016) [4], Shao, Y. et. al (2019) [5], Vatanen, T. et al. (2016) [6], Yassour, M. et al. (2018) [7], Bonham, K. et al. (2023) [8]; and
  • 4 studies from the Wellcome Leap 1kD program: Fatori, D. et al. (2024) [9], Hemmingway, A. et al. (2020) [10], O'Sullivan, J. et al. (2024) [11] and the publication accompanying this dataset, Bottino, G. et al. (2024), which introduces the samples from the Khula study, Zieff, M. et al. (2024) [12]
• Pre-formatted selected functional EC profiles for the samples from the Khula study

Description of the data and file structure

This dataset contains 4 files:

• full_taxonomic_inputs.csv is the file containing all taxonomic profiles prior to prevalence filtering and removal of samples with no reads assigned to prevalence-filtered taxa. filtered_taxonomic_inputs.csv is the post-filter version of the same file, which contains only the samples and features utilized to train the Machine Learning models and to perform downstream analysis. These files have the following structure:
  • study_name: String, identified by publication , i.e. BonhamK_2023.
  • datagroup: String, meta-detagroup of samples. Options are ECHO, LEAP and CMD.
  • site: String, ISO 3166-1 alpha-3 three-letter code for country of origin of samples.
  • datacolor: String, manuscript-wide color adopter for data source.
  • datasource: String, major data source for the sample. Options are ECHO-RESONANCE, CMD-DIABIMMUNE, CMD-OTHER, 1kDLEAP-KHULA, 1kDLEAP-GERMINA, 1kDLEAP-COMBINE, 1kdLEAP-M4EFAD.
    • CMD-DIABIMMUNE and CMD-OTHER are combined into the single CMD meta category on most downstream analyses.
  • visit: String, nominal timepoint of stool sample collection, when available, on the original annotation format of each study. For a better understanding of acronyms in this column, please refer to the visit_annotation_glossary.xlsx file.
  • westernized_cat: String, categorization of site as a westernized country.
  • subject_id: String, unique study participant identifier
  • ageMonths: Numeric, age at stool sample collection, in months.
  • sample: String, unique sample identifier
  • Multiple columns of taxa abundances encoded with species name in the form Genus_species, i.e. Escherichia_coli or Klebsiella_pneumoniae. Values in relative abundances on the scale 0.0:100.0
  • Shannon_index: alpha-diversity of the sample measured as the Shannon index
• ecs.arrow is the file containing all species-stratified functional EC profiles for the 426 prevalence-filtered Khula samples from the prefious filtered_inputs.csv.
  • Columns are samples and rows are ECs with the format X.X.X.X: Function|Genus_species, i.e. 1.1.1.88: Hydroxymethylglutaryl-CoA reductase|s__Staphylococcus_aureus
  • Metadata can be obtained by joining on ecs_metadata.csv with column names matching the sample column
• AgeModel_FullCV_Results.jld2 are the results of the model training experiments with cross-validation, for full cross-cohort CV with hyperparameter grid search. Additional Leave-One-Source-Out crossvalidation results are also given on .jld format, for each of the data sources mentioned in the main text: regression_Age_LeaveCMDOut.jld, regression_Age_LeaveCombineOut.jld, regression_Age_LeaveEchoOut.jld, regression_Age_LeaveGerminaOut.jld, regression_Age_LeaveKhulaOut.jld, regression_Age_LeaveM4EFADOut.jld; to be read with the JLD2.jl library and the classes from the Leap.jl julia package available on Zenodo (opens in new window)
• We have also provided inputs_with_testdata.csv, which is a copy of full_taxonomic_inputs.csv with the held-out independent test samples (Ennis, 2024) [13] appended at the end. We similarly made available regression_Age_LeaveEnnisOut.jld as the test data results.

Sharing/Access information

Code for loading the data and performing analyses is available on Zenodo (opens in new window)

Code/Software

See the Readme and setup instructions for using code available on Zenodo (opens in new window)

1. Asnicar, F. et al. Studying vertical microbiome transmission from mothers to infants by strain-level metagenomic profiling. mSystems 2, (2017).
2. Bäckhed, F. et al. Dynamics and stabilization of the human gut microbiome during the first year of life. Cell Host Microbe 17, 690–703 (2015).
3. Kostic, A. D. et al. The dynamics of the human infant gut microbiome in development and in progression toward type 1 diabetes. Cell Host Microbe 17, 260–273 (2015).
4. Pehrsson, E. C. et al. Interconnected microbiomes and resistomes in low-income human habitats. Nature 533, 212–216 (2016).
5. Shao, Y. et al. Stunted microbiota and opportunistic pathogen colonization in caesarean-section birth. Nature 574, 117–121 (2019).
6. Vatanen, T. et al. Variation in microbiome LPS immunogenicity contributes to autoimmunity in humans. Cell 165, 842–853 (2016).
7. Yassour, M. et al. Natural history of the infant gut microbiome and impact of antibiotic treatment on bacterial strain diversity and stability. Sci. Transl. Med. 8, 343ra81 (2016).
8. Bonham, K. S. et al. Gut-resident microorganisms and their genes are associated with cognition and neuroanatomy in children. Sci Adv 9, eadi0497 (2023).
9. Fatori, D. et al. Identifying biomarkers and trajectories of executive functions and language development in the first 3 years of life: design, methods and initial findings of the Germina cohort study. Preprint at https://doi.org/10.31219/osf.io/ed4fb (2024).
10. Hemmingway, A. et al. A detailed exploration of early infant milk feeding in a prospective birth cohort study in Ireland: combination feeding of breast milk and infant formula and early breast-feeding cessation. Br. J. Nutr. 124, 440–449 (2020).
11. O’Sullivan, J. et al. Alterations in gut microbiota composition, plasma lipids, and brain activity, suggest inter-connected pathways influencing malnutrition-associated cognitive and neurodevelopmental changes. Preprint at https://doi.org/10.21203/rs.3.rs-4115616/v1 (2024).
12. Zieff, M. R. et al. Characterizing developing executive functions in the first 1000 days in South Africa and Malawi: The Khula Study. Wellcome Open Research 9, (2024).
13. Ennis, D., Shmorak, S., Jantscher-Krenn, E. & Yassour, M. Longitudinal quantification of Bifidobacterium longum subsp. infantis reveals late colonization in the infant gut independent of maternal milk HMO composition. Nat. Commun. 15, 894 (2024).