Gene-based microbiome representation enhances host phenotype classification
ABSTRACT With the concomitant advances in both the microbiome and machine learning fields, the gut microbiome has become of great interest for the potential discovery of biomarkers to be used in the classification of the host health status. Shotgun metagenomics data derived from the human microbiome...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
American Society for Microbiology
2023-08-01
|
| Series: | mSystems |
| Subjects: | |
| Online Access: | https://journals.asm.org/doi/10.1128/msystems.00531-23 |
| _version_ | 1797730300998975488 |
|---|---|
| author | Thomas Deschênes Fred Wilfried Elom Tohoundjona Pier-Luc Plante Vincenzo Di Marzo Frédéric Raymond |
| author_facet | Thomas Deschênes Fred Wilfried Elom Tohoundjona Pier-Luc Plante Vincenzo Di Marzo Frédéric Raymond |
| author_sort | Thomas Deschênes |
| collection | DOAJ |
| description | ABSTRACT With the concomitant advances in both the microbiome and machine learning fields, the gut microbiome has become of great interest for the potential discovery of biomarkers to be used in the classification of the host health status. Shotgun metagenomics data derived from the human microbiome is composed of a high-dimensional set of microbial features. The use of such complex data for the modeling of host-microbiome interactions remains a challenge as retaining de novo content yields a highly granular set of microbial features. In this study, we compared the prediction performances of machine learning approaches according to different types of data representations derived from shotgun metagenomics. These representations include commonly used taxonomic and functional profiles and the more granular gene cluster approach. For the five case-control datasets used in this study (Type 2 diabetes, obesity, liver cirrhosis, colorectal cancer, and inflammatory bowel disease), gene-based approaches, whether used alone or in combination with reference-based data types, allowed improved or similar classification performances as the taxonomic and functional profiles. In addition, we show that using subsets of gene families from specific functional categories of genes highlight the importance of these functions on the host phenotype. This study demonstrates that both reference-free microbiome representations and curated metagenomic annotations can provide relevant representations for machine learning based on metagenomic data. IMPORTANCE Data representation is an essential part of machine learning performance when using metagenomic data. In this work, we show that different microbiome representations provide varied host phenotype classification performance depending on the dataset. In classification tasks, untargeted microbiome gene content can provide similar or improved classification compared to taxonomical profiling. Feature selection based on biological function also improves classification performance for some pathologies. Function-based feature selection combined with interpretable machine learning algorithms can generate new hypotheses that can potentially be assayed mechanistically. This work thus proposes new approaches to represent microbiome data for machine learning that can potentiate the findings associated with metagenomic data. |
| first_indexed | 2024-03-12T11:42:17Z |
| format | Article |
| id | doaj.art-18e7e7187a95493599fc20eac3097c7f |
| institution | Directory Open Access Journal |
| issn | 2379-5077 |
| language | English |
| last_indexed | 2024-03-12T11:42:17Z |
| publishDate | 2023-08-01 |
| publisher | American Society for Microbiology |
| record_format | Article |
| series | mSystems |
| spelling | doaj.art-18e7e7187a95493599fc20eac3097c7f2023-08-31T13:00:43ZengAmerican Society for MicrobiologymSystems2379-50772023-08-018410.1128/msystems.00531-23Gene-based microbiome representation enhances host phenotype classificationThomas Deschênes0Fred Wilfried Elom Tohoundjona1Pier-Luc Plante2Vincenzo Di Marzo3Frédéric Raymond4Centre Nutrition, Santé et Société (NUTRISS) – Institut sur la Nutrition et les Aliments Fonctionnels (INAF), Université Laval , Québec, CanadaCentre Nutrition, Santé et Société (NUTRISS) – Institut sur la Nutrition et les Aliments Fonctionnels (INAF), Université Laval , Québec, CanadaCentre Nutrition, Santé et Société (NUTRISS) – Institut sur la Nutrition et les Aliments Fonctionnels (INAF), Université Laval , Québec, CanadaCentre Nutrition, Santé et Société (NUTRISS) – Institut sur la Nutrition et les Aliments Fonctionnels (INAF), Université Laval , Québec, CanadaCentre Nutrition, Santé et Société (NUTRISS) – Institut sur la Nutrition et les Aliments Fonctionnels (INAF), Université Laval , Québec, CanadaABSTRACT With the concomitant advances in both the microbiome and machine learning fields, the gut microbiome has become of great interest for the potential discovery of biomarkers to be used in the classification of the host health status. Shotgun metagenomics data derived from the human microbiome is composed of a high-dimensional set of microbial features. The use of such complex data for the modeling of host-microbiome interactions remains a challenge as retaining de novo content yields a highly granular set of microbial features. In this study, we compared the prediction performances of machine learning approaches according to different types of data representations derived from shotgun metagenomics. These representations include commonly used taxonomic and functional profiles and the more granular gene cluster approach. For the five case-control datasets used in this study (Type 2 diabetes, obesity, liver cirrhosis, colorectal cancer, and inflammatory bowel disease), gene-based approaches, whether used alone or in combination with reference-based data types, allowed improved or similar classification performances as the taxonomic and functional profiles. In addition, we show that using subsets of gene families from specific functional categories of genes highlight the importance of these functions on the host phenotype. This study demonstrates that both reference-free microbiome representations and curated metagenomic annotations can provide relevant representations for machine learning based on metagenomic data. IMPORTANCE Data representation is an essential part of machine learning performance when using metagenomic data. In this work, we show that different microbiome representations provide varied host phenotype classification performance depending on the dataset. In classification tasks, untargeted microbiome gene content can provide similar or improved classification compared to taxonomical profiling. Feature selection based on biological function also improves classification performance for some pathologies. Function-based feature selection combined with interpretable machine learning algorithms can generate new hypotheses that can potentially be assayed mechanistically. This work thus proposes new approaches to represent microbiome data for machine learning that can potentiate the findings associated with metagenomic data.https://journals.asm.org/doi/10.1128/msystems.00531-23microbiomemachine learningmetagenomicsshotgun microbiomefeature selectiongene clusters |
| spellingShingle | Thomas Deschênes Fred Wilfried Elom Tohoundjona Pier-Luc Plante Vincenzo Di Marzo Frédéric Raymond Gene-based microbiome representation enhances host phenotype classification mSystems microbiome machine learning metagenomics shotgun microbiome feature selection gene clusters |
| title | Gene-based microbiome representation enhances host phenotype classification |
| title_full | Gene-based microbiome representation enhances host phenotype classification |
| title_fullStr | Gene-based microbiome representation enhances host phenotype classification |
| title_full_unstemmed | Gene-based microbiome representation enhances host phenotype classification |
| title_short | Gene-based microbiome representation enhances host phenotype classification |
| title_sort | gene based microbiome representation enhances host phenotype classification |
| topic | microbiome machine learning metagenomics shotgun microbiome feature selection gene clusters |
| url | https://journals.asm.org/doi/10.1128/msystems.00531-23 |
| work_keys_str_mv | AT thomasdeschenes genebasedmicrobiomerepresentationenhanceshostphenotypeclassification AT fredwilfriedelomtohoundjona genebasedmicrobiomerepresentationenhanceshostphenotypeclassification AT pierlucplante genebasedmicrobiomerepresentationenhanceshostphenotypeclassification AT vincenzodimarzo genebasedmicrobiomerepresentationenhanceshostphenotypeclassification AT fredericraymond genebasedmicrobiomerepresentationenhanceshostphenotypeclassification |