Funding in, findings out

For any dataset: the NCI award that paid to create it, and the awards that got a published finding out of it afterwards.

Pick a dataset

332 of 602 datasets have at least one award resolved. 31 have awards on both sides.

Or start from one of these - awards that paid to create it awards that funded work using it

TCGA-THYM 17 → 22 ORGANOID-PANCREATIC 13 → 49 CPTAC-3 10 → 36 WCDT-MCRPC 9 → 43 TCGA-STAD 18 → 10 BEATAML1.0-COHORT 7 → 33

Pick a dataset to see its funding both ways

Whichever of the 332 you choose, the answer is drawn as one chain: the awards credited on the dataset’s own marker paper, the dataset, the articles that went on to quote its accession, and the awards those articles were funded by.

1. Awards that paid to create this dataCredited on the dataset's own marker paper
2. The dataset
3. Articles that used itThe accession appears in their methods section
4. Awards those articles were funded byNCI money spent on data it did not pay to generate

31 of them have awards on both sides, which is the fullest version of the picture. The rest say which side is empty and why.

Where the links come from

NIH RePORTER maps a PMID to the awards that reported it. On a dataset’s marker paper that gives the awards which created it; on an article that used the data, the awards which funded that use.

The two sides are not measured alike. Generation is complete: a dataset has one marker paper and every award on it is resolved. Reuse is a sample - the articles using a dataset can number in the thousands, and only the few dozen strongest are retrieved and looked up. Every award shown here funded reuse; there are more that are not shown.

How far it reaches

283 datasets have an award credited with creating them, 80 with using them, and 31 both.

What is missing

No citable accession means no traceable articles; no authoritative marker paper means no attributable generation award. Absence is a gap in the record. Methods.

Start from an award instead, and see every dataset it touches.