Dataset
Retrieval coverage, per compound
One row per compound in this portfolio's corpus: the literature query that was run for it, how many records that query returned, how many rows the corpus actually carries, and the difference between the two.
77 rows12 columnsGenerated 13 October 2026
The file
77 rows, 12 columns, 7 KB. Comma separated, UTF-8, one header row, quoted where a value contains a comma or a quote. Download corpus-coverage-gap.csv.
How it was derived
Each row joins two generated files for the same compound. retrieved_hits and query come from data/studies/<slug>.json, which records what Europe PMC returned at retrieval time. indexed_rows and not_indexed come from data/evidence/<slug>.json. Coverage is computed here as rows divided by hits, to four decimal places, because the smallest value in the set is under one hundredth of one per cent.
Regenerate it with node scripts/sentinel-analysis.mjs --write. The same run produces every figure in 12 rows out of 188,556, so the page and the file cannot drift apart.
Columns
| Column | What it holds |
|---|---|
| slug | The compound's identifier in this corpus. Joins to every other dataset here. |
| name | Display name as the corpus holds it. |
| query | The exact Europe PMC query string that produced the hit count. |
| retrieved | Date the query was run, ISO format. |
| retrieved_hits | Europe PMC's own hit count for that query on that date. A hit is a record whose title or abstract contains the phrase. |
| indexed_rows | How many records the corpus carries for this compound, after the twelve-row cap. |
| not_indexed | retrieved_hits minus indexed_rows. The records the query returned and the corpus does not hold. |
| coverage_pct | indexed_rows as a percentage of retrieved_hits, four decimal places. |
| distinct_journals | Distinct journal names among the carried rows. |
| year_from | Publication year of the oldest carried row. |
| year_to | Publication year of the newest carried row. Low values indicate a citation-ranked sample, not a recent one. |
| named_in_title | How many carried rows name the compound in the title rather than only in the abstract. |
What it cannot be used for
retrieved_hits is a count of records matching a phrase, not a count of studies about the compound. For common endogenous molecules most of those records concern something else entirely, so a large not_indexed value is not a list of missing peptide papers.
The limits of the analysis built on this file are stated on the investigation itself, and the rules this publication applies to all of its counting are on the methods page.