@article{fdi:010078969,
  title = {{I}nterpretable and accurate prediction models for metagenomics data},
  author = {{P}rifti, {E}di and {C}hevaleyre, {Y}. and {H}anczar, {B}. and {B}elda, {E}. and {D}anchin, {A}. and {C}lement, {K}. and {Z}ucker, {J}ean-{D}aniel},
  editor = {},
  language = {{ENG}},
  abstract = {{B}ackground: {M}icrobiome biomarker discovery for patient diagnosis, prognosis, and risk evaluation is attracting broad interest. {S}elected groups of microbial features provide signatures that characterize host disease states such as cancer or cardio-metabolic diseases. {Y}et, the current predictive models stemming from machine learning still behave as black boxes and seldom generalize well. {T}heir interpretation is challenging for physicians and biologists, which makes them difficult to trust and use routinely in the physician-patient decision-making process. {N}ovel methods that provide interpretability and biological insight are needed. {H}ere, we introduce "predomics", an original machine learning approach inspired by microbial ecosystem interactions that is tailored for metagenomics data. {I}t discovers accurate predictive signatures and provides unprecedented interpretability. {T}he decision provided by the predictive model is based on a simple, yet powerful score computed by adding, subtracting, or dividing cumulative abundance of microbiome measurements. {R}esults: {T}ested on >100 datasets, we demonstrate that predomics models are simple and highly interpretable. {E}ven with such simplicity, they are at least as accurate as state-of-the-art methods. {T}he family of best models, discovered during the learning process, offers the ability to distil biological information and to decipher the predictability signatures of the studied condition. {I}n a proof-of-concept experiment, we successfully predicted body corpulence and metabolic improvement after bariatric surgery using pre-surgery microbiome data. {C}onclusions: {P}redomics is a new algorithm that helps in providing reliable and trustworthy diagnostic decisions in the microbiome field. {P}redomics is in accord with societal and legal requirements that plead for an explainable artificial intelligence approach in the medical field.},
  keywords = {prediction ; interpretable models ; metagenomics biomarkers ; microbial ecosystems},
  booktitle = {},
  journal = {{G}igascience},
  volume = {9},
  numero = {3},
  pages = {giaa010 [11 p.]},
  ISSN = {2047-217{X}},
  year = {2020},
  DOI = {10.1093/gigascience/giaa010},
  URL = {https://www.documentation.ird.fr/hor/fdi:010078969},
}