Sugi Atlas
52K+ deterministic gene/drug/disease reference pages
Last verified:
What is Sugi Atlas?
Sugi Atlas is a biomedical reference atlas built from a knowledge graph that mines around 70 primary biomedical sources into more than 52,000 canonical pages on human genes, drugs, and diseases. It is designed as a deterministic, reproducible catalog rather than a model-generated summary, so each page is assembled from fixed query plans and the same entity is covered with the same breadth on every rebuild.
The tool organizes its content into three linked corpora: genes, drugs, and diseases. Gene pages include identifiers, protein products, variants, expression, pathways, clinical evidence, and pharmacology; drug pages include mechanism targets, indications, clinical trials, and evidence; disease pages include associated genes, molecular evidence, clinical features, therapeutics, and trials. Relationships are woven into a cross-entity mesh so users can navigate from a gene to related drugs and diseases, or from a disease to its linked genes and treatments.
The site is built for researchers and AI agents that need grounded biomedical reference data. It emphasizes source traceability, stable page structure, schema.org metadata, and openly published pages that can be browsed or consumed programmatically. It is especially useful when you need a current, connected view of a single biomedical entity without manually reconciling dozens of databases.
Sugi Atlas also includes deterministic validation and refresh workflows so the catalog can be regenerated as upstream data changes. The pages are uniform, machine-readable, and openly available, making the atlas useful both as a human-readable reference and as a structured grounding layer for automated systems.
Sugi Atlas pricing
Pricing model: Freemium
The website describes Sugi Atlas as openly available, and it does not present a paid pricing page on the site content reviewed. The atlas itself is published openly at the main site, while the underlying pipeline is open-source software released under the MIT license. The site text does not list a free tier, subscription tiers, or feature-limited plans; instead, access appears to be open and public. The source and documentation are available in the project repository, and the catalog can be regenerated as the underlying graph is refreshed.
Sugi Atlas pros
- 52,000+ canonical reference pages
- Covers genes, drugs, and diseases
- Built from 70 primary sources
- Every figure traces to source
- Deterministic page generation
- Reproducible across builds
- Uniform breadth for every entity
- Stable page structure
- Cross-entity links between corpora
- Gene-to-drug-to-disease navigation
- Human-readable markdown pages
- Schema.org structured metadata
- Machine-readable and agent-friendly
- Updated at least monthly
- Openly published catalog
- Covers proteins with gene pages
- Includes variants and expression data
- Includes clinical evidence and trials
- Includes pathway and GO enrichment
- Validates pages before publication
Sugi Atlas cons
- Not a full literature review
- Some sections show no-data states
- Many disease pages are thin
- Top-ranked records only on page
- Complete lists require the BioBTree API
- Human pages only, not all species
- Depends on BioBTree availability
- Coverage limited to curated relationships
- Off-label uses are not shown
- Non-curated edges are excluded
- Salt-form children may be removed
- Some disease molecular sections vary by evidence
- Not every entity has rich annotation
- No model-driven exploratory query mode
- Depth is constrained by source databases
Frequently asked questions about Sugi Atlas
What is Sugi Atlas?
Sugi Atlas is a biomedical reference atlas that consolidates genes, drugs, and diseases into a single browsable catalog. It uses a knowledge graph to mine and organize data from around 70 primary biomedical sources into more than 52,000 canonical pages. The result is a structured reference layer meant for both researchers and AI agents.
What kinds of pages does it include?
It includes three main page types: gene pages, drug pages, and disease pages. Gene pages cover identifiers, protein products, variants, expression, pathways, clinical evidence, and related pharmacology. Drug pages cover targets, indications, clinical trials, and clinical evidence, while disease pages cover associated genes, molecular evidence, clinical features, therapeutics, and trials.
How is Sugi Atlas built?
The atlas is built deterministically from the BioBTree knowledge graph using fixed query plans. Instead of letting a model improvise each page at runtime, the system traverses the graph in a reproducible way so every entity is covered consistently. This makes the catalog easier to refresh, validate, and reproduce from a given graph snapshot.
How current is the data?
The site states that the atlas is updated at least monthly. Because it is built from a knowledge graph that re-ingests upstream sources, the catalog can be regenerated as the source databases change. That design is meant to keep the reference content current without relying on manual rewrite cycles.
How are gene, drug, and disease pages connected?
Sugi Atlas builds a cross-entity mesh that links curated relationships across the three corpora. A relationship recorded on one page can be navigated from the related gene, drug, or disease page, while keeping the relationship type distinct. This lets users move from a gene to its implicated drugs and diseases, or from a disease to its associated genes and treatments.
Does Sugi Atlas show source evidence?
Yes. The site emphasizes that every figure traces to source, and the pages are generated from primary biomedical databases rather than model-created summaries. It also includes machine-readable metadata such as schema.org records so consumers can inspect and reuse the data more reliably.
What makes it different from a generic biomedical search site?
Sugi Atlas is not just a search interface or a scraped directory. It is a deterministic catalog with stable page contracts, explicit no-data states, and curated cross-links between genes, drugs, and diseases. That makes it better suited to reproducible reference work and programmatic grounding than a loosely aggregated search result page.
Who is Sugi Atlas for?
It is aimed at researchers and AI agents that need a grounded, connected biomedical reference. The site is useful when someone wants a consolidated view of a single gene, drug, or disease without manually reconciling many databases. Its structured pages also make it useful for automated workflows and retrieval systems.
What are the limitations of the pages?
Some pages show only top-ranked records rather than exhaustive lists on the page itself, and the full list may require the underlying BioBTree API. The atlas also has explicit no-data states for under-annotated entities, so not every page is equally rich. In addition, it focuses on curated and reproducible content rather than broad speculative connections.
Is the project open source?
Yes. The site text says the Sugi Atlas pipeline is open-source software released under the MIT license. The catalog is openly published, and the source code and documentation are available in the project repository.