{
"$type": "site.standard.document",
"bskyPostRef": {
"cid": "bafyreicutvacbvl7pegp7lw5n3al23my3jm4tvxicohu2bja64e2hjw774",
"uri": "at://did:plc:2gbt2dlwaqovtnmxkat3tyke/app.bsky.feed.post/3mhcr2kxwty52"
},
"coverImage": {
"$type": "blob",
"ref": {
"$link": "bafkreigajr4sbxguuwrpl7kq7lydstmrflq3ifjbwxdchieu4awwpv7z3u"
},
"mimeType": "image/jpeg",
"size": 68812
},
"path": "/articles/d41586-026-00787-3",
"publishedAt": "2026-03-18T05:30:47.307Z",
"site": "https://www.nature.com",
"tags": [
"doi:10.1038/d41586-026-00787-3"
],
"textContent": "Nature, Published online: 17 March 2026; doi:10.1038/d41586-026-00787-3\n\nThe database of 200 million protein-structure predictions now includes homodimers, adding new biological relevance.",
"title": "AlphaFold hits ‘next level’: the AI database now includes protein pairing",
"updatedAt": "2026-03-17T00:00:00.000Z"
}