The knowledge graph

Everything this project knows about the water bear lives in one graph. Every fact points to the paper, page, figure or dataset it came from. The questions papers leave open are in it, so are the tests we ran against data (failures included), and so is every assumption we had to make where no source says. It is yours to search, download and build on.

Fly through it → — the whole graph drawn at once, filterable by type, in the browser.

Download: nodes.jsonl.gz and edges.jsonl.gz (one JSON object per line), or the summary. Licence: CC BY 4.0 for our work; the facts belong to their sources, which are cited on every line. Please cite Tardigraph.

How it is built

• Claims are single facts, each with a source and a locator (page, figure, table, supplementary file or record), how it was established (reported, measured, inferred, assumed) and a confidence.
• Sources are papers, datasets, books and web pages; questions are the open problems sources pose, quoted; tests compare a model or claim with data and say pass, partial or fail in words; assumptions are claims with evidence "assumed", listed so nothing is hidden.
• Edges say how things relate: cites, supports, contradicts (both sides kept), refines, answers, checks, in_species (anything from another species is marked as such) and more.

Screened for the public copy: local file paths and bookkeeping fields, and cross-references to our private data-quality register (a flag is not an accusation until it is checked). The papers themselves are not included.