Skip to main content
European Commission logo
English English
CORDIS - EU research results
CORDIS
CORDIS Web 30th anniversary CORDIS Web 30th anniversary

Illuminating Earth’s microbial diversity and origins from metagenomes with deep learning

Project description

Expanding our knowledge on Earth’s microbial diversity

Microbes are very small living things that cannot be seen with the naked eye. They are found all around us and play a crucial role in maintaining the balance of nutrients and waste products in the biosphere. Moreover, they are important for the preservation of the natural environment by regulating biogeochemical cycles. The EU-funded ERMADA project aims to analyse and elucidate Earth’s microbial diversity using bioinformatics and machine learning algorithms. Specifically, it will shed light on the composition and structure of the microbiome at different rank levels and lineages and provide a complete record of the planet’s present microbial diversity footprint.

Objective

The estimated number of microbes on our planet outnumbers the stars of the Milky Way galaxy and their biomass exceeds that of all plants and animals. Out of the 10^12 microbial species, only around 10^4 have been cultured, less than 10^5 species are represented by classified sequences, and a staggering estimated 99% of these microorganisms remain taxonomically unknown. Metagenomic shotgun sequencing has emerged as the most prevalent way of studying and classifying microorganisms from various habitats whereas genome analysis can be used to uncover the functions of genes, enzymes and metabolic pathways in a microbial community. This painstaking effort is crucial to understanding Earth's biodiversity, as microbes play important roles in regulating the planet’s biogeochemical cycles through processes that govern nutrient circulation in both terrestrial and marine environments. In this proposal, we will employ cutting edge bioinformatics and machine learning algorithms to analyze and elucidate Earth’s microbial diversity. We will use deep neural networks trained by large volumes of metagenomic sequences as well as big data methods to process hundreds of terabytes of data and taxonomically classify all uncharacterized metagenomic samples, by identifying their origins and habitats. Going beyond the capacities of conventional sequence similarity and comparison analyses, neural network models can capture higher level, abstract defining features and patterns in metagenomic sequences. The aim of this study is twofold: i) to gain a deeper understanding of the composition and structure of the microbiome at different rank levels and lineages and ii) to provide a complete record of the planet’s present microbial diversity footprint. The latter can serve as a reference dataset for future studies pertaining to microbiome evolution due to climate change or other long-term environmental factors.

Coordinator

EREVNITIKO KENTRO VIOIATRIKON EPISTIMON ALEXANDROS FLEMINGK
Net EU contribution
€ 247 628,16
Address
FLEMING STREET 34
16672 Vari-Athens
Greece

See on map

Region
Αττική Aττική Ανατολική Αττική
Activity type
Research Organisations
Links
Total cost
€ 247 628,16