The data itself is not stored in the blockchain. Only verification information about the data (SHA256 hash) is stored as a manifest in the blockchain along with the metadata. The actual data remains off-chain.
Storing large amounts of data in the blockchain is inefficient especially since some scientific datasets tend to be in the multi-terabytes size range. Storing only the comprehensive metadata of a dataset enables researchers to share large datasets or sensitive data that are stored off-chain, yet verifiable with the information stored in the blockchain.