Besides the ~20,000 human proteins, DeepMind provides open access to proteomes of 20 other organisms, totalling over 350,000 protein structures.

True if the AlphaFold database release of July 2021 made freely available, without restriction, the human proteome (about 20,000 proteins) plus the proteomes of 20 other organisms, with a combined total of more than 350,000 structures. Mostly true if the release matches except for a small slip: the number of other organisms was 19 to 21, or the total was just under 350,000 (above about 340,000). Mixed if one major part is substantially wrong, such as far fewer organisms, while the rest holds. Mostly false if a release happened but both the organism count and the total are substantially wrong. False if no such open-access release of these proteomes happened in July 2021.

Set before the research began.

1. E1 DeepMind and EMBL release the most complete database of predicted 3D structures of human proteins EMBL-EBI · 22 July 2021

   “Partners use AlphaFold, the AI system recognised last year as a solution to the protein structure prediction problem, to release more than 350,000 protein structure predictions including the entire human proteome to the scientific community”

2. E2 DeepMind and EMBL release the most complete database of predicted 3D structures of human proteins EMBL-EBI · 22 July 2021

   “the data will be freely and openly available to the scientific community”

3. E3 DeepMind and EMBL release the most complete database of predicted 3D structures of human proteins EMBL-EBI · 22 July 2021

   “In addition to the human proteome, the database launches with ~350,000 structures including 20 biologically-significant organisms such as E.coli, fruit fly, mouse, zebrafish, malaria parasite and tuberculosis bacteria.”

4. E4 CHANGELOG.txt · LiteFold/AlphaFoldDB at ad648d3 Hugging Face (LiteFold dataset repository)

   “* Initial release of 21 model organisms”

EMBL-EBI's launch announcement of 22 July 2021 says the partners released more than 350,000 protein structure predictions, including the entire human proteome (E1). The same announcement says the database launched with the human proteome plus 20 biologically-significant organisms (E3), and that the data would be freely and openly available to the scientific community (E2). A changelog for the database records an initial release of 21 model organisms (E4), which fits human plus 20 others. Together these sources support the release of the human proteome plus 20 other organisms, more than 350,000 structures in total, made freely and openly available to the scientific community; they do not give the size of the human proteome.