Two approaches for large-scale classification of fluorescence microscopy images were used to analyze subcellular protein patterns in images from the HPA Cell Atlas. An image-classification task was introduced into the online science fiction video game, EVE Online. 320,000 gamers provided more than 32 million image classifications, and the data was combined with deep learning to build a tool to classify proteins into 29 subcellular localization patterns.
Key publication
- Sullivan DP et al., Deep learning is combined with massive-scale citizen science to improve large-scale image classification. Nat Biotechnol. (2018)
PubMed: 30125267 DOI: 10.1038/nbt.4225
Other selected publications
- Peplow M., Citizen science lures gamers into Sweden's Human Protein Atlas. Nat Biotechnol. (2016)
PubMed: 27153260 DOI: 10.1038/nbt0516-452c
Figure legend: An EVE Online spaceship cruising a "universe" of cells from the HPA.
Key facts
- Deep learning was combined with massive-scale citizen science
- Project Discovery marks the first time a citizen science task was integrated into a mainstream online computer game
- In one year, 320,000 players provided 32 million image classifications and spent a total of 70 working years