Corpus Engine Towards a reproducible .science file.

Extraction

Turning published research into structured science files.

  1. 1. Retrieve the paper

    Collect the available source text.

  2. 2. Extract the science

    Identify methods, data and results.

  3. 3. Check quotations

    Match supporting quotes to the source.

  4. 4. Publish the file

    .sci / .science
    Structured findings with source references.

Progress through PubMed

9,746 / approximately 40 million records

Published as .science files · approximately 0.0244% of PubMed

0.0244%

Based on NLM’s figure of over 40 million PubMed records.

Processing details
Papers fetched and sectioned
9,996
Sections indexed
272,742
Records with structured methods
8,434