Machine Learning Meta-analysis of Large Metagenomic Datasets: Tools and Biological Insights
Open accessThe DOI matches its Crossref record; 70 of the 71 references it cites were found in a registry or an index.
Abstract
Shotgun metagenomic analysis of the human associated microbiome provides a rich set of microbial features for prediction and biomarker discovery in the context of human diseases and health conditions. However, the use of such high-resolution microbial features presents new challenges, and validated computational tools for learning tasks are lacking. Moreover, classification rules have scarcely been validated in independent studies, posing questions about the generality and generalization of disease-predictive models across cohorts. In this paper, we comprehensively assess approaches to metagenomics-based prediction tasks and for quantitative assessment of the strength of potential microbiome-phenotype associations. We develop a computational framework for prediction tasks using quantitative microbiome profiles, including species-level relative abundances and presence of strain-specific markers. A comprehensive meta-analysis, with particular emphasis on generalization across cohorts, was performed in a collection of 2424 publicly available metagenomic samples from eight large-scale studies. Cross-validation revealed good disease-prediction capabilities, which were in general improved by feature selection and use of strain-specific markers instead of species-level taxonomic abundance. In cross-study analysis, models transferred between studies were in some cases less accurate than models tested by within-study cross-validation. Interestingly, the addition of healthy (control) samples from other studies to training sets improved disease prediction capabilities. Some microbial species (most notably Streptococcus anginosus) seem to characterize general dysbiotic states of the microbiome rather than connections with a specific disease. Our results in modelling features of the "healthy" microbiome can be considered a first step toward defining general microbial dysbiosis. The software framework, microbiome profiles, and metadata for thousands of samples are publicly available at http://segatalab.cibio.unitn.it/tools/metaml.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Figures






![Fig 7 Cross-study analysis in multiple gut datasets for (a) T2D discrimination and (b) disease discrimination (independently from the type of disease). For (a), we included all the healthy (controls) and diabetes (cases) samples, whereas samples labelled as other diseases were not considered. For (b), we instead included all the samples where samples with one of the considered diseases were put together in the same "diseases" class. The * denotes cross-validation results (with the margin of error reported in parenthesis). In the other cases, the model was generated on all the datasets other than the dataset considered for testing, a “leave-one-dataset-out” cross-study validation [51]. For the testing datasets with only healthy samples, prediction accuracy was evaluated in terms of overall accuracy (OA). In bold we report the best value for each scenario and feature type (i.e., species abundance or marker presence), and circled are the absolute best value for each scenario.](https://figures.cochat.ai/paper-figures/5bb56b13e2b047c5a1c14f0db710b338/machine-learning-meta-analysis-of-large-metagenomic-datasets-tools-and-figure-7.jpg)
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The cited source says what the claim attributes to it.
In the citing paper
“We chose to use RF as our classifier, given its, on average, superior performance using microbial feature sets in previous benchmarking studies 31 .”
From A gut microbial signature for combination immune checkpoint blockade across cancer types (2024) · cites this paper as [31]
In the citing paper
“As in previous studies (Pasolli et al., 2016; Thomas et al., 2019), RFs using functional features performed similarly (0.69 cross validation and 0.71 LODO ROC AUC on pathways relative abundance), indicating a tight link between strain-specific taxonomy and gene carriage in this setting.”
From Integrating taxonomic, functional, and strain-level profiling of diverse microbial communities with bioBakery 3 (2021) · cites this paper as Pasolli et al., 2016
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Breiman, L. (2001) · Machine Learning · doi:10.1023/a:1010933404324 · Cited by 133,538
Scikit-learn: Machine Learning in Python
Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., Blondel, M., Müller, A., Nothman, J., Louppe, G., et al. (2012) · arXiv · doi:10.48550/arxiv.1201.0490 · Cited by 71,016 · cited 1×
“The classifiers were implemented using the scikit-learn python package [61].”
MethodCortes, C., & Vapnik, V. (1995) · Machine Learning · doi:10.1007/bf00994018 · Cited by 59,421
Regression shrinkage and selection via the lasso
Tibshirani, R. (1996) · Journal of the Royal Statistical Society Series B: Statistical Methodology · doi:10.1111/j.2517-6161.1996.tb02080.x · Cited by 51,482 · cited 13×
“Lasso is based on a multiple logistic regression trained with L1-norm penalized likelihood, while both L1 and L2 norms are penalized in ENet.”
BackgroundMethodInfluentialNeural Networks: A Comprehensive Foundation
Haykin, S. (1998) · Cited by 30,070
Wu, Y. (1999) · Technometrics · doi:10.1080/00401706.1999.10485951 · Cited by 27,221
The Elements of Statistical Learning
Hastie, T., Friedman, J., & Tibshirani, R. (2001) · Springer Series in Statistics · doi:10.1007/978-0-387-21606-5 · Cited by 24,162
Regularization and Variable Selection Via the Elastic Net
Zou, H., & Hastie, T. (2005) · Journal of the Royal Statistical Society Series B: Statistical Methodology · doi:10.1111/j.1467-9868.2005.00503.x · Cited by 20,957
Classification and Regression Trees.
Ryzin, J. V., Breiman, L., Friedman, J. H., Olshen, R. A., & Stone, C. J. (1986) · Journal of the American Statistical Association · doi:10.2307/2288003 · Cited by 20,825
Metagenomic biomarker discovery and explanation
Segata, N., Izard, J., Waldron, L., Gevers, D., Miropolsky, L., Garrett, W. S., & Huttenhower, C. (2011) · Genome Biology · doi:10.1186/gb-2011-12-6-r60 · Cited by 16,192 · cited 1×
“The approach proposed here (RF with embedded feature selection) focuses on the set of features with the most discriminating power rather than on strictly statistical assessments [46] or statistical assessment coupled with effect size [47].”
MethodSelf-citation
Research works printed but not linked: 1
- J Qin, 2012, A metagenome-wide association study of gut microbiota in type 2 diabetes, Nature, 490, 55, 10.1038/nature11450 · DOI · Paper page
Cited by
Items 101–200 of 666
Manghi, P., Bhosle, A., Wang, K., Marconi, R., Selma‐Royo, M., Ricci, L., Asnicar, F., Golzato, D., Ma, W., Hang, D., et al. (2024) · Nature Microbiology · doi:10.1038/s41564-024-01858-9 · Cited by 90
“We then used metAML (version 1.1) 31 to assess the link with the microbiome between combinations of the following coffee categories: ‘never’, ‘moderate’ and ‘high’.”
InfluentialSelf-citationHuman Gut Microbiota Predicts Susceptibility to Vibrio cholerae Infection
Midani, F. S., Weil, A. A., Chowdhury, F., Begum, Y. A., Khan, A. I., Debela, M. D., Durand, H. K., Reese, A. T., Nimmagadda, S. N., Silverman, J. D., et al. (2018) · The Journal of Infectious Diseases · doi:10.1093/infdis/jiy192 · Cited by 86
Gut butyrate-producers confer post-infarction cardiac protection
Chen, H.-C., Liu, Y.-W., Chang, K., Wu, Y., Chen, Y., Chao, Y.-K., You, M.-Y., Lundy, D. J., Lin, C., Hsieh, M. L., et al. (2023) · Nature Communications · doi:10.1038/s41467-023-43167-5 · Cited by 82
“The performance of our model surpasses a previous report using microbiota database of 951 participants from the American Gut Project to predict CVDs 20 .”
InfluentialKok, C. R., Rose, D., & Hutkins, R. (2023) · Annual Review of Food Science and Technology · doi:10.1146/annurev-food-060721-015516 · Cited by 73
Next-Generation Metagenomics: Methodological Challenges and Opportunities
Laudadio, I., Fulci, V., Stronati, L., & Carissimi, C. (2019) · OMICS A Journal of Integrative Biology · doi:10.1089/omi.2019.0073 · Cited by 92
“…et al., 2015; Ji et al., 2017; Wang et al., 2015), metagenomic reads binning to reference genomes (Vervier et al., 2018), gene prediction (Al-Ajlan and El Allali, 2018), gene function prediction (Li et al., 2018), and analyses of large datasets (LaPierre et al., 2019; Pasolli et al., 2016).”
Electronic Cigarette Use Promotes a Unique Periodontal Microbiome
Thomas, S. C., Xu, F., Pushalkar, S., Lin, Z., Thakor, N., Vardhan, M., Flaminio, Z., Khodadadi‐Jamayran, A., Vasconcelos, R., Akapo, A., et al. (2022) · mBio · doi:10.1128/mbio.00075-22 · Cited by 96
TaxoNN: ensemble of neural networks on stratified microbiome data for disease prediction
Sharma, D., Paterson, A. D., & Xu, W. (2020) · Bioinformatics · doi:10.1093/bioinformatics/btaa542 · Cited by 75
A meta-analysis study of the robustness and universality of gut microbiome-metabolome associations
Muller, E., Algavi, Y. M., & Borenstein, E. (2021) · Microbiome · doi:10.1186/s40168-021-01149-z · Cited by 85
“vice versa, following previous microbiome-related metaanalysis studies [25, 77, 78].”
BackgroundFilippis, F. D., Valentino, V., Sequino, G., Borriello, G., Riccardi, M. G., Pierri, B., Cerino, P., Pizzolante, A., Pasolli, E., Esposito, M., et al. (2024) · Nature Communications · doi:10.1038/s41467-024-48739-7 · Cited by 92
Self-citationVangay, P., Hillmann, B. M., & Knights, D. (2019) · GigaScience · doi:10.1093/gigascience/giz042 · Cited by 72
Metagenomics and artificial intelligence in the context of human health
Wani, A. K., Roy, P., Gahlawat, V. K., & Mir, T. U. G. (2022) · Infection Genetics and Evolution · doi:10.1016/j.meegid.2022.105267 · Cited by 79
Challenges and opportunities in sharing microbiome data and analyses
Huttenhower, C., Finn, R. D., & McHardy, A. C. (2023) · Nature Microbiology · doi:10.1038/s41564-023-01484-x · Cited by 78
“Instead, only certain of these features are directly comparable, or divergent studies must be combined meta-analytically 26 .”
BackgroundImmunotherapy for hepatocellular carcinoma: Recent advances and future targets
Yu, S. J. (2023) · Pharmacology & Therapeutics · doi:10.1016/j.pharmthera.2023.108387 · Cited by 73
Associations between the gut microbiome and metabolome in early life
Nguyen, Q. P., Karagas, M. R., Madan, J. C., Dade, E. F., Palys, T. J., Morrison, H. G., Pathmasiri, W., McRitche, S., Sumner, S., Frost, H. R., et al. (2021) · BMC Microbiology · doi:10.1186/s12866-021-02282-3 · Cited by 80
“It is unlikely that model choice was driving the lack of predictability, since all chosen methods had been shown to be suited for microbiome-associated prediction tasks [28, 31] as well as covering both linear and nonlinear associations.”
MethodMarzal, L. N. P., Rojas-Velázquez, D., Rigters, D., Prince, N., Garssen, J., Kraneveld, A. D., Perez‐Pardo, P., & Lopez‐Rincon, A. (2024) · Scientific Reports · doi:10.1038/s41598-023-50601-7 · Cited by 77
Deep learning methods in metagenomics: a review
Roy, G., Prifti, E., Belda, E., & Zucker, J. (2024) · Microbial Genomics · doi:10.1099/mgen.0.001231 · Cited by 96
“Some interesting reviews of these methods have already been published [70, 71].”
BackgroundConsistent signatures in the human gut microbiome of old- and young-onset colorectal cancer
Qin, Y., Tong, X., Mei, W.-J., Cheng, Y., Zou, Y., Han, K., Yu, J., Jie, Z., Zhang, T., Zhu, S., et al. (2024) · Nature Communications · doi:10.1038/s41467-024-47523-x · Cited by 74
“The random forest algorithm, known for its superior performance among various machine learning tools 61 , has been successfully employed in previous large-scale CRC microbiome studies 13,17 .”
Han, X., Liu, Q., Li, Y., Zhang, M., Liu, K., Kwok, L.-Y., Zhang, H., & Zhang, W. (2025) · Trends in Food Science & Technology · doi:10.1016/j.tifs.2025.104938 · Cited by 64
Physicochemical Drivers of Microbial Community Structure in Sediments of Lake Hazen, Nunavut, Canada
Ruuskanen, M. O., Pierre, K. A. S., Louis, V. L. S., Aris‐Brosou, S., & Poulain, A. J. (2018) · Frontiers in Microbiology · doi:10.3389/fmicb.2018.01138 · Cited by 72
“To address these limitations, machine learning methods have been used, for instance to predict disease progression from human gut microbiomes (Pasolli et al., 2016), to determine the factors affecting microbial diversity in soil (Ge et al.”
MethodA Genomic Outlook on Bioremediation: The Case of Arsenic Removal
Plewniak, F., Crognale, S., Rossetti, S., & Bertin, P. N. (2018) · Frontiers in Microbiology · doi:10.3389/fmicb.2018.00820 · Cited by 69
“Indeed, power analysis and sample size requirements estimation for high-throughput sequencing data demand computations of much higher complexity than classical statistical analyses and must be fine-tuned to the type of problem that is being addressed (Pasolli et al., 2016; Li et al., 2017).”
BackgroundScientific novelty beyond the experiment
Hallsworth, J. E., Udaondo, Z., Pedrós‐Alió, C., Höfer, J., Benison, K. C., Lloyd, K. G., Cordero, R. J. B., de Campos, C. B. L., Yakimov, M. M., & Amils, R. (2023) · Microbial Biotechnology · doi:10.1111/1751-7915.14222 · Cited by 62
Diffusion-based tractography atlas of the human acoustic radiation
Maffei, C., Sarubbo, S., & Jovicich, J. (2019) · Scientific Reports · doi:10.1038/s41598-019-40666-8 · Cited by 63
Anatomy and physiology of word-selective visual cortex: from visual features to lexical processing
Caffarra, S., Karipidis, I. I., Yablonski, M., & Yeatman, J. D. (2021) · Brain Structure and Function · doi:10.1007/s00429-021-02384-8 · Cited by 78
Zhao, L., Walkowiak, S., & Fernando, W. G. D. (2023) · Plants · doi:10.3390/plants12091852 · Cited by 70
“Learning and predicting the health status of plants with metagenome samples is not common, but an insightful study has been conducted for clinical metagenomics studies with more than 2400 metagenome samples [55].”
Ghaffari, P., Shoaie, S., & Nielsen, L. K. (2022) · Journal of Translational Medicine · doi:10.1186/s12967-022-03365-z · Cited by 63
“A variety of machine learning approaches, such as logistic regression, neural networks, and support vector machines, have been used to identify microbial features present in several disease states [91].”
BackgroundPredictive Modeling of Microbiome Data Using a Phylogeny-Regularized Generalized Linear Mixed Model
Xiao, J. Y., Chen, L., Johnson, S. P., Yu, Y., Zhang, X., & Chen, J. (2018) · Frontiers in Microbiology · doi:10.3389/fmicb.2018.01391 · Cited by 66
“…machine learning methods, such as Random Forest and Support Vector Machine, as well as sparse regression models, such as Lasso (Tibshirani, 1996), MCP (Zhang, 1996), and Elastic Net (Zou and Trevor, 2005), have been applied (Knights et al., 2011; Statnikov et al., 2013; Pasolli et al., 2016).”
MethodShanahan, E. R., McMaster, J. J., & Staudacher, H. M. (2021) · Journal of Human Nutrition and Dietetics · doi:10.1111/jhn.12868 · Cited by 56
“This highlights the utility of placing datasets from individual studies in context, such as through meta- analysis, to delineate trends from type1 errors.(79,107)”
MethodMachine Learning and Deep Learning Applications in Metagenomic Taxonomy and Functional Annotation
Mathieu, A., Leclercq, M., Sanabria, M., Périn, O., & Droit, A. (2022) · Frontiers in Microbiology · doi:10.3389/fmicb.2022.811495 · Cited by 69
“Learning and prediction of disease status of patient-related metagenomic samples have been rarely explored, but a successful application has been proposed using more than 2,400 metagenomic samples from clinical metagenomic studies (Pasolli et al., 2016).”
BackgroundElucidating the Beta-Diversity of the Microbiome: from Global Alignment to Local Alignment
Su, X. (2021) · mSystems · doi:10.1128/msystems.00363-21 · Cited by 65
“inflammatory bowel disease (14) and colorectal cancer (15), in other cases like type 1 diabetes (16) and autism spectrum disorder (17), only a small part of signature taxa play crucial roles that can be determined by statistical tests (18) or supervised machine learning (19) but are missed by the end-to-end comparison…”
BackgroundMulenga, M., Kareem, S. A., Sabri, A. Q. M., Seera, M., Govind, S. K., Samudi, C., & Mohamad, S. B. (2021) · IEEE Access · doi:10.1109/access.2021.3050838 · Cited by 56
“For instance, work in [16] used conventional ML algorithms to process shotgunmetagenomic data, consisting of 2424 sampleswhich were pooled from 8 different studies.”
MethodTierney, B. T., Tan, Y., Kostic, A. D., & Patel, C. J. (2021) · Nature Communications · doi:10.1038/s41467-021-23029-8 · Cited by 64
Intestinal Permeability Is a Mechanical Rheostat in the Pathogenesis of Liver Cirrhosis
Nishimura, N., Kaji, K., Kitagawa, K., Sawada, Y., Furukawa, M., Ozutsumi, T., Fujinaga, Y., Tsuji, Y., Takaya, H., Kawaratani, H., et al. (2021) · International Journal of Molecular Sciences · doi:10.3390/ijms22136921 · Cited by 68
“The role of dysbiosis in patients with CLDs has been reported by a study investigating the alteration in gut microbial diversity by 16S sequencing and a meta-analysis in some diseases such as liver cirrhosis, colorectal cancer, inflammatory bowel disease, obesity, and type 2 diabetes mellitus [5].”
BackgroundDeciphering the Ecology of Cystic Fibrosis Bacterial Communities: Towards Systems-Level Integration
Bevivino, A., Bacci, G., Dřevínek, P., Nelson, M. T., Hoffman, L. R., & Mengoni, A. (2019) · Trends in Molecular Medicine · doi:10.1016/j.molmed.2019.07.008 · Cited by 66
“Even if resultant findings cannot be used directly in clinical settings, such association is essential for developingmechanistic models that can be used for biomarker development or applied directly in clinical settings [90].”
BackgroundPapadopoulos, L., Lynn, C. W., Battaglia, D. A., & Bassett, D. S. (2020) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1008144 · Cited by 58
Pietrucci, D., Teofani, A., Milanesi, M., Fosso, B., Putignani, L., Messina, F., Pesole, G., Desideri, A., & Chillemi, G. (2022) · Biomedicines · doi:10.3390/biomedicines10082028 · Cited by 63
“Regarding the RF algorithms, the “Gini Impurity Decrease Index” was used to evaluate the importance of each feature to the classification [62].”
MethodNogal, A., Asnicar, F., Vijay, A., Kouraki, A., Visconti, A., Louca, P., Wong, K. E., Baleanu, A.-F., Giordano, F., Wolf, J., et al. (2023) · Gut Microbes · doi:10.1080/19490976.2023.2240050 · Cited by 62
“We selected RF-based methods a priori as it has been repeatedly shown to be particularly suitable and robust to the statistical challenges inherent to microbiome abundance data.(42) A cross-validation approach was implemented, based on 100 bootstrap iterations and an 80/20 random split into training and testing folds.”
MethodSelf-citationBang, S., Yoo, D., Kim, S., Jhang, S. Y., Cho, S., & Kim, H. (2019) · Scientific Reports · doi:10.1038/s41598-019-46249-x · Cited by 61
The critical role of tumor microbiome in cancer immunotherapy
Yang, L., Wang, Q., He, L., & Sun, X. (2024) · Cancer Biology & Therapy · doi:10.1080/15384047.2024.2301801 · Cited by 55
“68 Furthermore, the integration of multi-omics data with clinical and immunologic information can expedite the creation of predictive models and biomarkers, thereby ultimately enabling personalized cancer treatment strategies.”
BackgroundA review of machine learning methods for cancer characterization from microbiome data
Teixeira, M., Silva, F., Ferreira, R. M., Pereira, T., Figueiredo, C., & Oliveira, H. P. (2024) · npj Precision Oncology · doi:10.1038/s41698-024-00617-7 · Cited by 61
Fernández-Edreira, D., Liñares-Blanco, J., & Fernandez-Lozano, C. (2021) · Expert Systems with Applications · doi:10.1016/j.eswa.2021.115648 · Cited by 55
Bakır-Güngör, B., Hacılar, H., Jabeer, A., Nalbantoğlu, Ö. U., Aran, O., & Yousef, M. (2022) · PeerJ · doi:10.7717/peerj.13205 · Cited by 60
A novel deep learning method for predictive modeling of microbiome data
Wang, Y., Bhattacharya, T., Jiang, Y., Qin, X., Wang, Y., Liu, Y., Saykin, A. J., & Chen, L. (2020) · Briefings in Bioinformatics · doi:10.1093/bib/bbaa073 · Cited by 58
Can Gut Microbiota Be a Good Predictor for Parkinson’s Disease? A Machine Learning Approach
Pietrucci, D., Teofani, A., Unida, V., Cerroni, R., Biocca, S., Stefani, A., & Desideri, A. (2020) · Brain Sciences · doi:10.3390/brainsci10040242 · Cited by 51
“ML algorithms are currently used for building predictive models for the classification of biological data, and identify biomarkers through a training procedure [20,21].”
MethodPoulsen, C. S., Ekstrøm, C. T., Aarestrup, F. M., & Pamp, S. J. (2022) · Microbiology Spectrum · doi:10.1128/spectrum.00090-22 · Cited by 50
“With the increasing amount of metagenomic data in public repositories, meta-analysis and pooling of data from different studies are exciting new opportunities to gain further insight into the microbial world (10 – 12, 24, 33).”
BackgroundChang, D., Gupta, V. K., Hur, B., Cobo-López, S., Cunningham, K. Y., Han, N. S., Lee, I., Kronzer, V. L., Teigen, L. M., Karnatovskaia, L. V., et al. (2024) · Nature Communications · doi:10.1038/s41467-024-51651-9 · Cited by 50
“Additionally, including gut microbial information from all taxonomic ranks could uncover more features that accurately predict host phenotypes 14,15 .”
Temporal Stability of the Ruminal Bacterial Communities in Beef Steers
Clemmons, B. A., Martino, C., Schneider, L. G., Lefler, J., Embree, M. M., & Myer, P. R. (2019) · Scientific Reports · doi:10.1038/s41598-019-45995-2 · Cited by 54
Dogra, S. K., Martin, F., Donnicola, D., Julita, M., Berger, B. W., & Sprenger, N. (2021) · Microorganisms · doi:10.3390/microorganisms9091939 · Cited by 52
“Therefore, we initially explored the data using multiple classification methods (Random Forests with Boruta feature selection, Support Vector Machines [58] and sPLS-DA), all based on different algorithms, and observed some features that were consistently selected by the different methods.”
MethodManghi, P., Blanco‐Míguez, A., Manara, S., Nabinejad, A., Cumbo, F., Beghini, F., Armanini, F., Golzato, D., Huang, K. D., Thomas, A. M., et al. (2023) · Cell Reports · doi:10.1016/j.celrep.2023.112464 · Cited by 51
Self-citationHuang, C. D., Gin, C., Fettweis, J. M., Foxman, B., Gelaye, B., MacIntyre, D. A., Subramaniam, A., Fraser, W. D., Tabatabaei, N., & Callahan, B. J. (2023) · BMC Biology · doi:10.1186/s12915-023-01702-2 · Cited by 62
The Computational Diet: A Review of Computational Methods Across Diet, Microbiome, and Health
Eetemadi, A., Rai, N., Pereira, B. M. P., Kim, M., Schmitz, H. H., & Tagkopoulos, I. (2020) · Frontiers in Microbiology · doi:10.3389/fmicb.2020.00393 · Cited by 55
“A recent increase in the number of microbiome studies with publicly available data has enabled cross-study data integration (Pasolli et al., 2016, 2017; Duvallet et al., 2017; Wang et al., 2018; Thomas et al., 2019; Wirbel et al., 2019).”
BackgroundMicrobiome-based disease prediction with multimodal variational information bottlenecks
Grazioli, F., Siarheyeu, R., Alqassem, I., Henschel, A., Pileggi, G., & Meiser, A. (2022) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1010050 · Cited by 53
Microbiome of the first stool after birth and infantile colic
Korpela, K., Renko, M., Paalanne, N., Vänni, P., Salo, J., Tejesvi, M. V., Koivusaari, P., Pokka, T. M.-L., Kaukola, T., Pirttilä, A. M., et al. (2020) · Pediatric Research · doi:10.1038/s41390-020-0804-y · Cited by 51
“Machine learning is a novel way to analyse microbiome data (38,39).”
MethodInterpretable and accurate prediction models for metagenomics data
Prifti, E., Chevaleyre, Y., Hanczar, B., Belda, E., Danchin, A., Clément, K., & Zucker, J. (2020) · GigaScience · doi:10.1093/gigascience/giaa010 · Cited by 50
Massive metagenomic data analysis using abundance-based machine learning
Harris, Z. N., Dhungel, E., Mosior, M., & Ahn, T.-H. (2019) · Biology Direct · doi:10.1186/s13062-019-0242-0 · Cited by 50
“For example, MetAML [46] was developed for metagenomics-based prediction tasks and for quantitative assessment of the strength of potential microbiomephenotype associations.”
MethodLiu, B., Sträuber, H., Saraiva, J., Harms, H., Silva, S. G., Kasmanas, J. C., Kleinsteuber, S., & da Rocha, U. N. (2022) · Microbiome · doi:10.1186/s40168-021-01219-2 · Cited by 50
“Finally, other studies that used 16S rRNA sequencing data in machine learning solutions also reported random forest to show good prediction performance [43, 50, 51].”
ResultByndloss, M. X., Devkota, S., Duca, F. A., Niess, J. H., Nieuwdorp, M., Orho‐Melander, M., Sanz, Y., Tremaroli, V., & Zhao, L. (2024) · Diabetologia · doi:10.1007/s00125-024-06198-1 · Cited by 61
“…lifestyle, and other demographic characteristics, it will be important to perform studies in large populations and to include healthy individuals/control participants from different studies as references; these approaches have been shown to increase disease prediction accuracy [5, 6].”
BackgroundCharacterising the respiratory microbiome
Watson, R. L., de Koff, E. M., & Bogaert, D. (2019) · European Respiratory Journal · doi:10.1183/13993003.01711-2018 · Cited by 47
Sharpton, T., Lyalina, S., Luong, J., Pham, J., Deal, E. M., Armour, C., Gaulke, C., Sanjabi, S., & Pollard, K. S. (2017) · mSystems · doi:10.1128/msystems.00036-17 · Cited by 45
“However, 338 our results are consistent with microbiome case-control studies that found significant differences in the 339 taxonomic (11, 83-87) and functional (13, 16, 22) profiles of IBD patients compared to healthy controls, 340 especially in Crohn’s disease.”
ResultAdvancing microbiome research with machine learning: key findings from the ML4Microbiome COST action
D’Elia, D., Truu, J., Lahti, L., Berland, M., Παπουτσόγλου, Γ., Ceci, M., Zomer, A., Lopes, M. B., Ibrahimi, E., Gruca, A., et al. (2023) · Frontiers in Microbiology · doi:10.3389/fmicb.2023.1257002 · Cited by 50
“Pasolli et al. (2016) have demonstrated that the performance of ML models may vary substantially depending on the disease addressed in the dataset.”
BackgroundModeling Microbial Communities: Perspective and Challenges
Raajaraam, L., & Raman, K. (2024) · ACS Synthetic Biology · doi:10.1021/acssynbio.4c00116 · Cited by 45
Recent advances on DNA and omics‐based technology in Food testing and authentication: A review
Kumar, P., Rani, A., Singh, S., & Kumar, A. (2022) · Journal of Food Safety · doi:10.1111/jfs.12986 · Cited by 42
Zuo, W., Michail, S. K., & Sun, F. (2022) · Frontiers in Cellular and Infection Microbiology · doi:10.3389/fcimb.2022.918010 · Cited by 45
“Since the random forests algorithm (Breiman, 2001) has been proven to have better performance than other machine learning models, especially on microbial abundance data (Pasolli et al., 2016; Gao et al., 2022), all experiments were carried out using the random forests classifier implemented by the python package sciki…”
MethodLin, Y., Lau, H. C.-H., Liu, C., Ding, X., Sun, Y., Rong, J., Zhang, X., Wang, L., Yuan, K., Miao, Y., et al. (2025) · Cell Host & Microbe · doi:10.1016/j.chom.2025.03.012 · Cited by 46
Tackmann, J., Arora, N., Schmidt, T., Rodrigues, J. F. M., & von Mering, C. (2018) · Microbiome · doi:10.1186/s40168-018-0565-6 · Cited by 45
Eosinophilic Granulomatosis With Polyangiitis
Berti, A., Cornec, D., Moura, M. C., Smyth, R., Dagna, L., Specks, U., & Keogh, K. A. (2020) · CHEST Journal · doi:10.1016/j.chest.2019.11.045 · Cited by 44
A microbial signature following bariatric surgery is robustly consistent across multiple cohorts
Fouladi, F., Carroll, I. M., Sharpton, T. J., Bulik‐Sullivan, E. C., Heinberg, L. J., Steffen, K. J., & Fodor, A. A. (2021) · Gut Microbes · doi:10.1080/19490976.2021.1930872 · Cited by 45
“Machine learning approaches have often been used in metagenomic analyses to determine whether a microbial signature in one study can be used to predict the phenotype of samples from another study [33].”
BackgroundHsu, C.-K., Su, S., Chang, L., Shao, S., Yang, K.-J., Chen, C., Chen, Y.-T., & Wu, I. (2021) · International Journal of Medical Sciences · doi:10.7150/ijms.66451 · Cited by 46
“Streptococcus anginosus, which increased in condition of unhealthy microbiome [35], was associated with tumorigenesis of intestinal tract, such as colon or esophageal cancer, and lupus activity [36-38].”
BackgroundState-dependent evolutionary models reveal modes of solid tumour growth
Lewinsohn, M., Bedford, T., Müller, N. F., & Feder, A. F. (2023) · Nature Ecology & Evolution · doi:10.1038/s41559-023-02000-4 · Cited by 58
Sun, W., Zhang, Y., Guo, R., Sha, S., Chen, C., Ullah, H., Zhang, Y., Ma, J., You, W., Meng, J., et al. (2024) · npj Biofilms and Microbiomes · doi:10.1038/s41522-024-00567-9 · Cited by 51
Vilne, B., Ķibilds, J., Siksna, I., Lazda, I., Valciņa, O., & Krūmiņa, A. (2022) · Frontiers in Microbiology · doi:10.3389/fmicb.2022.627892 · Cited by 46
“Pasolli et al. (2016) utilized 2,424 shotgun metagenomic samples from eight studies to assess the potential of the (mainly gut) microbiome species-level abundances to be used in order to differentiate healthy vs. unhealthy (including obese and T2D) individuals and compare the prediction accuracy of…”
MethodYoung, C., Wood, H. M., Balaguer, A. F., Bottomley, D., Gallop, N., Wilkinson, L., Benton, S. C., Brealey, M., John, C., Burtonwood, C., et al. (2021) · Clinical Cancer Research · doi:10.1158/1078-0432.ccr-20-3807 · Cited by 47
Liu, Y., Zhu, J., Wang, H., Lu, W., LEE, Y. K., Zhao, J., & Zhang, H. (2022) · BMC Genomics · doi:10.1186/s12864-022-09087-2 · Cited by 37
“Specifically, we used Decision Tree (DT), Random Forest (RF), Gradient boosted regression trees (GBDT), eXtreme Gradient Boosting (XGBoost, XGBRF), Adaptive Boosting (AdaBoost), and LGB (Light Gradient Boosting Machine (LGB) to select the features through embedding in the modeling phase based on the scikit-learn and X…”
MethodLarge-scale metagenomic analysis of oral microbiomes reveals markers for autism spectrum disorders
Manghi, P., Filosi, M., Zolfo, M., Casten, L. G., García-Valiente, A., Mattevi, S., Heidrich, V., Golzato, D., Perini, S., Thomas, A. M., et al. (2024) · Nature Communications · doi:10.1038/s41467-024-53934-7 · Cited by 52
Self-citationLi, P., Li, M., & Chen, W. (2025) · Gut Microbes · doi:10.1080/19490976.2025.2489074 · Cited by 46
“29 For instance, among machine learning algorithms, methods based on Random Forest and Least Absolute Shrinkage and Selection Operator (Lasso) logistic regression are particularly popular due to their advantages, which include high performance with small sample sizes (e.g., fewer than 50),…”
Meta-analysis of cellular toxicity for graphene via data-mining the literature and machine learning
Ma, Y., Wang, J., Wu, J., Tong, C., & Zhang, T. (2021) · The Science of The Total Environment · doi:10.1016/j.scitotenv.2021.148532 · Cited by 40
Dynamic task-linked switching between brain networks – A tri-network perspective
Shaw, S. B., McKinnon, M. C., Heisz, J. J., & Becker, S. (2021) · Brain and Cognition · doi:10.1016/j.bandc.2021.105725 · Cited by 49
Current Status and Future Therapeutic Options for Fecal Microbiota Transplantation
Ткач, С. М., Dorofeyev, A., Kuzenko, I., Boyko, N., Falalyeyeva, T. M., Boccuto, L., Scarpellini, E., Kobyliak, N. M., & Abenavoli, L. (2022) · Medicina · doi:10.3390/medicina58010084 · Cited by 41
Kim, H., Nelson, P., Nzabarushimana, É., Shen, J., Jensen, J., Bhosle, A., Li, C., Lee, N., Everett, C., Berdy, B. M., et al. (2025) · Nature Metabolism · doi:10.1038/s42255-025-01318-6 · Cited by 42
“Among classical machine-learning algorithms, random forest, linear regression models and support vector machines have been shown to perform well on metagenomic data 56–58 .”
Information Theory: New Look at Oncogenic Signaling Pathways
Zielińska, K. A., & Katanaev, V. L. (2019) · Trends in Cell Biology · doi:10.1016/j.tcb.2019.08.005 · Cited by 40
Yang, F., & Zou, Q. (2020) · Database · doi:10.1093/database/baaa050 · Cited by 38
“While autoML systems are mature enough to enhance the prediction accuracy, few of them are designed to meet the specific requirements of microbial classification tasks (8).”
BackgroundTrivieri, N., Pracella, R., Cariglia, M. G., Panebianco, C., Parrella, P., Visioli, A., Giani, F., Soriano, A. A., Barile, C., Canistro, G., et al. (2020) · Journal of Experimental & Clinical Cancer Research · doi:10.1186/s13046-020-01801-w · Cited by 46
Bakır-Güngör, B., Bulut, O., Jabeer, A., Nalbantoğlu, Ö. U., & Yousef, M. (2021) · Frontiers in Microbiology · doi:10.3389/fmicb.2021.628426 · Cited by 49
“On the T2D-associated metagenomics dataset, by only using 40 species (features) that are selected using Lasso feature selection, Pasolli et al. (2016) obtained 0.70 AUC using RF classifier, as shown in Supplementary Figures 2 , 3 .”
BackgroundMethodInfluentialToward a health-associated core keystone index for the human gut microbiome
Goel, A., Shete, O., Goswami, S., Samal, A., C.B., L., Kedia, S., Ahuja, V., O’Toole, P. W., Shanahan, F., & Ghosh, T. S. (2025) · Cell Reports · doi:10.1016/j.celrep.2025.115378 · Cited by 37
Understanding the Functionality of Probiotics on the Edge of Artificial Intelligence (AI) Era
Asar, R., Erenler, S., Devecioglu, D., Ispirli, H., Karbancioglu-Guler, F., Ozturk, H. I., & Dertli, E. (2025) · Fermentation · doi:10.3390/fermentation11050259 · Cited by 37
“The application of these artificial intelligence techniques has yielded very successful results, especially in health status monitoring and early disease prediction of microbiome data [165].”
Computational frameworks transform antagonism to synergy in optimizing combination therapies
Chen, J., Lin, A., Jiang, A., Qi, C., Liu, Z., Cheng, Q., Yuan, S., & Luo, P. (2025) · npj Digital Medicine · doi:10.1038/s41746-025-01435-2 · Cited by 46
“Simultaneously, by employing machine learning and other techniques to identify key microbial biomarkers and metabolic pathways, we can guide the optimization of drug combinations and personalized therapies 80 .”
Predicting microbiomes through a deep latent space
García-Jiménez, B., Muñoz, J., Cabello, S., Medina, J., & Wilkinson, M. D. (2020) · Bioinformatics · doi:10.1093/bioinformatics/btaa971 · Cited by 42
“For example, MetAML (Pasolli et al., 2016), a computational tool for metagenomics-based prediction tasks and for microbiome-phenotype associations; q2-sample-classifier (Bokulich et al., 2018), a plugin for QIIME 2 for supervised classification; and Microbiome Learning Repo (ML Repo) (Vangay et al., 2019) which provid…”
MethodComparison of vaginal microbiota sampling techniques: cytobrush versus swab
Mitra, A., MacIntyre, D. A., Mahajan, V., Lee, Y. S., Smith, A., Marchesi, J. R., Lyons, D., Bennett, P. R., & Kyrgiou, M. (2017) · Scientific Reports · doi:10.1038/s41598-017-09844-4 · Cited by 40
Giliberti, R., Cavaliere, S., Mauriello, I. E., Ercolini, D., & Pasolli, E. (2022) · PLoS Computational Biology · doi:10.1371/journal.pcbi.1010066 · Cited by 41
“We calculated the 95% confidence interval on the difference in AUC performance between two classifiers as done in [14] using the t-distribution with df = 9:”
BackgroundMethodResultInfluentialSelf-citationA toolbox of machine learning software to support microbiome analysis
Marcos-Zambrano, L. J., López-Molina, V. M., Bakır-Güngör, B., Frohme, M., Karađuzović-Hadžiabdić, K., Klammsteiner, T., Ibrahimi, E., Lahti, L., Lončar-Turukalo, T., Dhamo, X., et al. (2023) · Frontiers in Microbiology · doi:10.3389/fmicb.2023.1250806 · Cited by 50
“Here, SVMs, RFs, Lasso, Elastic Net, and other classifiers are implemented in this ML software framework for metagenome-based prediction tasks (Pasolli et al., 2016).”
MethodMeta-analysis identifies common gut microbiota associated with multiple sclerosis
Lin, Q., Dorsett, Y., Mirza, A., Tremlett, H. L., Piccio, L. M., Longbrake, E. E., Choileáin, S. N., Hafler, D. A., Cox, L. M., Weiner, H. L., et al. (2024) · Genome Medicine · doi:10.1186/s13073-024-01364-x · Cited by 46
“We chose random forest (RF) classifier for MS and con-trol classification as RF was reported to perform well for microbiota data [28].”
MethodJiao, N., Loomba, R. S., Yang, Z., Wu, D., Fang, S., Bettencourt, R., Lan, P., Zhu, R., & Zhu, L. (2021) · Physiological Genomics · doi:10.1152/physiolgenomics.00011.2021 · Cited by 38
Casimiro‐Soriguer, C. S., Loucera, C., Peña-Chilet, M., & Dopazo, J. (2022) · Scientific Reports · doi:10.1038/s41598-021-04182-y · Cited by 42
“In particular, recent evidence suggests that the human gut microbiome could be a relevant factor in human diseases 18,19 .”
BackgroundLee, Y., Cappellato, M., & Camillo, B. D. (2022) · GigaScience · doi:10.1093/gigascience/giad083 · Cited by 45
Obesity and the gut microbiota: implications of neuroendocrine and immune signaling
Romaní‐Pérez, M., Liébana-García, R., Flor‐Duro, A., Bonillo‐Jiménez, D., Bullich‐Vilarrubias, C., Olivares, M., & Sanz, Y. (2024) · FEBS Journal · doi:10.1111/febs.17249 · Cited by 37
“Studies using shotgun metagenomic sequencing to analyze the relationship between the microbiota and obesity are limited [19,20], but a recent study including two large cohorts showed that these data allow the classification of subject’s features important for defin-ing metabolic health, such as BMI…”
Aasmets, O., Lüll, K., Lang, J. M., Pan, C., Kuusisto, J., Fischer, K., Laakso, M., Lusis, A. J., & Org, E. (2021) · mSystems · doi:10.1128/msystems.01191-20 · Cited by 39
“Although machine learning has become popular in microbiome studies as 79 well, the ability of the algorithms to provide robust results remains unclear (9, 10).”
BackgroundXiang, L., Jin, X., Liu, Y., Ma, Y., Jian, Z., Wei, Z., Li, H., Li, Y., & Wang, K. (2021) · World Journal of Urology · doi:10.1007/s00345-021-03801-7 · Cited by 34
Role of Digital Health and Artificial Intelligence in Inflammatory Bowel Disease: A Scoping Review
Majidova, K., Handfield, J., Kafi, K., Martin, R., & Kubinski, R. (2021) · Genes · doi:10.3390/genes12101465 · Cited by 37
How Diet and Lifestyle Can Fine-Tune Gut Microbiomes for Healthy Aging
Tamayo, M., Olivares, M., Ruas‐Madiedo, P., Margollés, A., Espín, J. C., Medina, I., Moreno-Arribas, M. V., Canals, S., Mirasso, C. R., Ortín, S., et al. (2024) · Annual Review of Food Science and Technology · doi:10.1146/annurev-food-072023-034458 · Cited by 44
The Microbiome and Hepatocellular Carcinoma
Rattan, P., Minacapelli, C. D., & Rustgi, V. K. (2020) · Liver Transplantation · doi:10.1002/lt.25828 · Cited by 34
“(35) Computational analysis with machine-based learning using 16S rRNA databases has reaffirmed these observations of compositional imbalance in liver disease.(49) This tool allows assessment of the predictive power of microbiome composition by large-scale shotgun metagenomics and provides an accurate and practical us…”
MethodBiomarkers and Future Perspectives for Hepatocellular Carcinoma Immunotherapy
He, Y., Lu, M., Che, J., Chu, Q., Zhang, P., & Chen, Y. (2021) · Frontiers in Oncology · doi:10.3389/fonc.2021.716844 · Cited by 37
Metadata
- Authors
- Pasolli, E., Truong, D. T., Malik, F., Waldron, L., & Segata, N.
- Affiliations
- University of Trento (IT)
- City University of New York (US)
- Corresponding author
- Segata, Nicola
- Venue
- PLOS Computational Biology
- Publisher
- Public Library of Science (PLoS)
- Published
- 2016-07-11
- Volume
- 12
- Issue
- 7
- Pages
- e1004977
- ISSN
- 1553-734X (ISSN-L)
- 1553-7358
- PMC
- PMC4939962
- OpenAlex
- W2473355215
- Cited by
- 659
- References
- 71
- Citations by year
- 2026: 81 · 2025: 66 · 2024: 83 · 2023: 64 · 2022: 84 · 2021: 93 · 2020: 80 · 2019: 44 · 2018: 38 · 2017: 24 · 2016: 1
- FWCI
- 3.23
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- PLoS Computational Biology (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- Iris (University of Trento) (submittedVersion, cc-by)
- RePEc: Research Papers in Economics (submittedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
- Figshare (submittedVersion, cc-by)
- Funders
- Ministero dell’Istruzione, dell’Università e della Ricerca (RBFR13EWWI)
- Seventh Framework Programme (BE) (PCIG13-618833)
- Fondazione Caritro (Rif.Int.2013.0239)
- Terme di Comano
- National Science Foundation (CNS-958379, CNS-0855217, ACI-1126113)
- City University of New York High Performance Computing Center
- National Institute of Allergy and Infectious Diseases (1R21AI121784-01)
- Topics
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Gut microbiota and health
- Health Sciences › Dentistry › Periodontics › Oral microbiology and periodontitis research
- Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Genomics and Phylogenetic Studies
- Keywords
- metagenomics, human microbiome, disease prediction, biomarker discovery, strain-specific markers, dysbiosis, Streptococcus anginosus, feature selection, cross-validation
- MeSH terms
- Gastrointestinal Microbiome, Machine Learning, Humans, Obesity, Software, Colorectal Neoplasms, Inflammatory Bowel Diseases, Computational Biology, Metagenome, Metagenomics
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.