What is CultureBotAI?

CultureBotAI is a research initiative at Lawrence Berkeley National Laboratory in Berkeley, California, that develops AI-driven tools and knowledge graphs for microbial cultivation and computational biology.

Who leads CultureBotAI?

CultureBotAI is led by Dr. Marcin P. Joachimiak, a scientist specializing in microbiology, knowledge graph development, and computational biology at Lawrence Berkeley National Laboratory.

Where is CultureBotAI based?

CultureBotAI is based at Lawrence Berkeley National Laboratory in Berkeley, California, within the Environmental Genomics and Systems Biology Division.

KG-Microbe is a comprehensive modular knowledge graph developed by Dr. Marcin P. Joachimiak that integrates diverse microbial data sources to enable AI-driven insights for growth prediction and culture optimization.

How can I access KG-Microbe?

KG-Microbe is available on GitHub at https://github.com/Knowledge-Graph-Hub/kg-microbe under the BSD-3-Clause license. The publication is available in GigaScience at https://doi.org/10.1093/gigascience/giag077

What does CultureBotAI work on?

CultureBotAI focuses on three main areas: (1) cultivation of isolated and novel organisms, (2) culture optimization through data-driven approaches, and (3) growth preference prediction using machine learning and AI methods.

CommunityMech: Autonomous Knowledge Factory for Microbial Communities

Overview

CommunityMech is an autonomous knowledge factory for microbial community composition, ecological interactions, cultivation conditions, environments, and supporting evidence, with human oversight. Canonical YAML records feed schema validation, static browsing, and graph exports. Repository overview.

The repository contains 515 community records and nine additional isolate records. The published browser lists 510 communities across 16 categories, matching the landing page checked on November 8, 2026. Its landing page also advertises 512-dimensional v3 embeddings; the embedding views are separate derived artifacts and can cover a different set of records from the current browser. Published site.

Explore the Published Site

Record browser — filter communities by category, ecological state, metals, and rare-earth elements.
PaCMAP embedding map — explore similarity based on community taxonomic composition.
Graph layout — explore another layout of the community embedding space.
Text map and sequence map — explore text-embedding and sequence-embedding views across communities.

Recent Additions

Five most recently added community and isolate records in the pinned repository snapshot (cafbbc8):

Bacillus subtilis 168 BS0 Brewing-Wastewater Wild-Type Isolate Baseline (CommunityMech:000530, added 2026-11-08)
Bacillus subtilis 168 BS3 Brewing-Wastewater Metal-Resistant Isolate Baseline (CommunityMech:000529, added 2026-11-08)
SO3-Delta-mtrC-BS3 Lithium-ion Battery Bioleaching Consortium (CommunityMech:000526, added 2026-11-08)
SO3-Delta-mtrC-Delta-omcA-BS3 Lithium-ion Battery Bioleaching Consortium (CommunityMech:000528, added 2026-11-08)
SO3-Delta-omcA-BS3 Lithium-ion Battery Bioleaching Consortium (CommunityMech:000527, added 2026-11-08)

See all fourteen Mechs’ recent additions on the X-Mech Suite overview.

Data Architecture

Location	Contents
`kb/communities/`	Canonical microbial community records
`data/isolates/`	Additional records using the community schema
`kb/taxa/`	Reusable taxon, genome, and gene records
`references_cache/`	Source text used by evidence checks
`history/`	Append-only curation provenance
`docs/`	Published HTML and maintainer documentation

A community record can describe taxonomic membership, functional roles, environmental context, cultivation conditions, and directed ecological interactions. Ontology identifiers are paired with labels, and evidence is attached to the assertions it supports. Schema validation checks structure; literature validation separately checks evidence against source text. Validation and evidence workflow.

Example: A Curated Community

The Yogurt Two-Species Starter Culture record, CommunityMech:000164, documents Streptococcus thermophilus and Lactobacillus delbrueckii subsp. bulgaricus. It records cross-feeding interactions with taxon and metabolite identifiers and supporting literature.

The README’s schema-validated example shows how that record represents membership, formate exchange, and evidence. Use the linked YAML for the complete current record.

Getting Started

Development and CI use Python 3.13, uv, and just. Run commands from a source checkout:

git clone https://github.com/CultureBotAI/CommunityMech.git
cd CommunityMech
just install
just test

Validate a tracked community:

community_record=kb/communities/Yogurt_TwoSpecies_Starter_Culture.yaml
just validate "$community_record"
just validate-strict "$community_record"
just validate-gtdb "$community_record"
just validate-gtdb-domain "$community_record"
just validate-terms "$community_record"

Evidence checking is a separate step and may use network access:

just validate-references-explained "$community_record"

The repository documents a corpus-wide evidence-repair backlog. A structurally valid record does not by itself establish that all of its scientific assertions have been verified. Current validation status.

Generate Pages and Graph Exports

just gen-html
just check-docs-current
just gen-browser
just kgx-export
just kgx-validate

GitHub Pages serves the committed docs/ tree. Update community YAML or rendering templates before regenerating derived HTML. Embedding-map generation additionally needs the local embedding artifact described in the visualization guide.

The KGX exporter writes nodes.tsv, edges.tsv, and manifest.json under output/kgx/; its release workflow compresses the two TSV files, although no release has been published yet. Edges carry deterministic identifiers, relation predicates, source attribution, and available publication/supporting-text evidence. These exports support downstream graph integration and community analysis. KGX contract.

Integration with the X-Mech Suite

Community records can refer to CultureMech media by culturemech_id. Ingredients that carry ChEBI terms can be matched to MediaIngredientMech records through those terms. Some records also hold older MediaIngredientMech:NNNNNN identifiers, which match no current MediaIngredientMech record. CommunityMech’s schema marks that identifier scheme as vestigial.

Repository & Documentation

Repository and published site
Schema
Curation workflow
Licensing policy: CC BY 4.0 for data; BSD-3-Clause for code. Redistributed source material retains its applicable terms.

X-Mech Suite overview - All fourteen Mechs, their shared vocabulary and cross-references, and the culturebotai-claw orchestrator
HabitatMech - Habitats harmonized from GOLD, BacDive, PREGO and Madin et al. into ontology-grounded records
TaxonMech - Microbial taxa and strains grounded in NCBI Taxonomy, harmonized with GTDB, LPSN and BacDive
TraitMech - Autonomous knowledge factory for microbial ecophysiological traits
PathwayMech - Microbial pathway mechanisms with participants, reactions, causal edges, and reference-backed evidence
CellStructureMech - Microbial cell structures, between the trait and protein layers
ProteinTraitsMech - Protein sequence, structure, and function traits
DUFMech - Domains and protein families of unknown function, with evidence gathering and triage worklists
NaturalProductMech - Natural product structures with their producer organisms and gene clusters
AntibioticMech - Antimicrobial structures harmonizing ChEBI and CARD/ARO
CMMMech - Evidence-backed microbial extraction, transformation and recovery of critical materials
AIMech - AIO-grounded AI concepts with maintained evidence and explicit ontology mappings
MediaIngredientMech - LLM-assisted ingredient curation
CultureMech - Single-organism media requirements (6,325 canonical media)
PFASCommunityAgents - AI-driven consortium design for PFAS biodegradation
kg-microbe - Central knowledge graph (864K+ species)
MicroGrowAgents - Multi-agent media design system

Publications & Citations

Primary Citation

If you use CommunityMech in your research, please cite:

Santangelo BE, et al. (2026). KG-Microbe - Building Modular and Scalable Knowledge Graphs for Microbiome and Microbial Sciences. GigaScience, giag077. doi: 10.1093/gigascience/giag077

Community assembly principles in microbial cultivation
Syntrophic partnerships in engineered systems
PFAS biodegradation consortium design

Contact & Collaboration

For questions about CommunityMech or consortium design projects:

Principal Investigator: Dr. Marcin P. Joachimiak
Email: mjoachimiak@lbl.gov
Organization: CultureBotAI
Laboratory: Environmental Genomics and Systems Biology Division, Lawrence Berkeley National Laboratory

Acknowledgments

CommunityMech development is supported by:

Lawrence Berkeley National Laboratory
U.S. Department of Energy, Office of Science
Environmental Genomics and Systems Biology Division
ABPDU (Advanced Biofuels and Bioproducts Process Development Unit)

Bibliography

Santangelo BE, Hegde H, Caufield JH, Reese J, Kliegr T, Hunter LE, Lozupone CA, Mungall CJ, Joachimiak MP. KG-Microbe — Building Modular and Scalable Knowledge Graphs for Microbiome and Microbial Sciences. GigaScience. 2026;giag077. doi:10.1093/gigascience/giag077
Naseem S, Miller MA, Martinez-Gomez NC, Sun N, Joachimiak MP. MicroGrowAgents: An Agentic AI System for Microbial Cultivation Engineering. bioRxiv. 2026. doi:10.64898/2026.06.04.729985
Máša P, Kliegr T, Joachimiak MP. Explainable rule-based prediction of cultivation media for microbes. Computational and Structural Biotechnology Journal. 2025;27:5194–5206. doi:10.1016/j.csbj.2025.10.014 · free full text

Full publication list →