Microbial Discovery Forge¶
AI-accelerated discovery for microbial research within the K-BERDL platform
The Microbial Discovery Forge is an integrated, AI-native environment within K-BERDL designed to accelerate hypothesis generation, annotation, and discovery across microbial and metagenomic datasets. It combines harmonized BER data with automated analysis pipelines and AI-assisted reasoning to support researchers working at the frontier of microbial science.
Overview¶
The Forge provides a unified workspace for:
- Exploring microbial genomes and metagenomes across BER program datasets
- Running automated annotation and functional classification pipelines
- Generating AI-assisted hypotheses from cross-program multi-omics data
- Visualizing metabolic pathways, gene clusters, and phenotypic relationships
Key Capabilities¶
| Capability | Description |
|---|---|
| Genome Annotation | Automated structural and functional annotation of microbial genomes using curated reference databases |
| Metagenomic Assembly | Scalable assembly and binning workflows for environmental metagenomes |
| Comparative Genomics | Cross-dataset comparison of gene content, phylogenetics, and metabolic potential |
| AI Hypothesis Engine | AI agents that surface candidate genes, pathways, and organisms of interest based on user queries |
| Knowledge Graph Integration | Links microbial entities to existing knowledge in JGI, NMDC, and KBase databases |
Supported Data Types¶
- Microbial isolate genomes (FASTA, GFF3)
- Metagenomic assemblies and bins (MAGs)
- Functional annotations (COG, KEGG, PFAM)
- Transcriptomics and proteomics data
- Environmental metadata (soil, sediment, water)
Integration with K-BERDL¶
The Microbial Discovery Forge is deeply integrated with the broader K-BERDL platform:
- Data Plane — Reads and writes directly to tenant Delta Lake schemas
- AI Integration — Leverages the K-BERDL Agent SDK for automated reasoning workflows
- Metadata Catalog — All outputs are registered in the unified BER metadata catalog with full provenance
- KBase Narratives — Analysis results can be exported directly into KBase Narratives for collaboration and publication
Getting Started¶
Documentation for onboarding, pipeline configuration, and API access is coming soon.
For early access, contact the K-BERDL platform team.