Probing the eukaryotic microbes of ruminants with a deep-learning classifier and comprehensive protein databases

Area of Science:

• Microbiology
• Bioinformatics
• Eukaryotic Genomics

Background:

• Metagenomics has advanced microbial understanding, but eukaryotic microbes in microbiomes, like the mammalian gastrointestinal (GI) tract, are understudied.
• Rumen eukaryotes, including ciliates and fungi, are crucial for feed digestion and methane emissions but lack comprehensive genomic data.
• Existing bioinformatic tools have limitations in accurately identifying and classifying eukaryotic sequences from metagenomic data.

Purpose of the Study:

• To develop an advanced bioinformatics tool, GutEuk, for accurate identification and analysis of eukaryotic sequences in metagenomes.
• To improve the distinction between fungal and protozoal sequences for a deeper understanding of their roles.
• To create enhanced eukaryotic reference databases for ruminant microbiomes.

Main Methods:

• Development of GutEuk, a novel bioinformatics tool designed for high-precision identification of eukaryotic sequences.
• Optimization of GutEuk for various sequence lengths and differentiation of fungal and protozoal origins.
• Curation of ruminant-specific eukaryotic protein databases.

Main Results:

• GutEuk demonstrated improved accuracy in identifying eukaryotic sequences compared to existing tools like Tiara and EukRep.
• Analysis of over 1000 rumen metagenomes using GutEuk revealed greater genomic diversity in protozoa than previously known.
• The curated databases significantly improved the functional analysis of ruminant eukaryotes versus prokaryotes.

Conclusions:

• GutEuk provides a robust solution for the understudied area of GI tract eukaryotes.
• The tool and associated databases open new avenues for in-depth research into the ecological, physiological, and nutritional roles of ruminant eukaryotes.
• This work significantly advances our ability to study eukaryotic diversity and function within complex microbiomes.