mebipred

mebipred predicts metal-binding proteins from protein sequences and annotates metal-binding ligands to support functional and ecological analyses.

Key Features:

• High accuracy: Achieves nearly 90% accuracy in identifying proteins that bind metal ions and ion-containing ligands.
• Reference-free approach: Operates without relying on sequence alignments, enabling alignment-free prediction.
• Comprehensive metal ion annotation: Annotates ten ubiquitously present metal ions and ion-containing ligands.
• Short sequence analysis: Identifies metal-binding functionalities from short sequence stretches, including translated sequencing reads from metagenomic samples.
• Microbiome-scale analysis: Analyzes microbiome-derived protein sequences to detect differences in metal usage across environments such as ocean, hot spring sediments, soil, and human host-related microbiomes and to detect shifts correlated with physiological conditions and ion concentrations.

Scientific Applications:

• Metalloprotein annotation: Annotates metal-binding proteins in genomes and metagenomes to inform protein function assignments.
• Environmental microbiology: Compares metal utilization across ecosystems (ocean, hot spring sediments, soil) to study ecological metal usage patterns.
• Microbiome and host-microbe studies: Assesses metal preferences and shifts in human-associated microbiomes in relation to physiological conditions.
• Functional and structural inference: Supports investigations into roles of metals in catalysis, DNA/RNA binding, and protein structural stability.

Methodology:

Uses a machine learning-based approach that leverages sequence-derived features, operates reference-free without sequence alignments, and is applicable to short translated sequencing reads from metagenomic samples.