question · intermediate · partly written
Has every protein been tested against every other protein?
Short answer
No. The largest binary map, HuRI, reports 52,569 interactions among 8,275 proteins, out of a space of over 150 million pairs, and its authors put it at 2 to 11 percent of the binary interactome. The largest co-complex map, BioPlex 3.0, used 10,128 baits in one cell line. Nobody claims completeness. 12
Reported, not yet replicated
The efforts, with scale more
Yeast, 2000 and 2001: two proteome-scale yeast two-hybrid screens. Ito's reported 4,549 interactions among 3,278 proteins and said the two screens "do not largely overlap" (abstract-level reading). Human, 2014: HI-II-14, about 14,000 interactions among 4,303 proteins, covering about 40 percent of its own search space. Human, 2020: HuRI, nine yeast two-hybrid screens over a 17,408-gene clone set; its authors report no interaction found for half of the proteins tested. Human, 2021: BioPlex 3.0, affinity purification mass spectrometry with 10,128 baits, 118,162 interactions among 14,586 proteins, in 293T cells and a second line; the authors write that false negatives may prevail. 34512
Reported, not yet replicated
They test different objects more
HuRI asks whether two proteins can touch; BioPlex asks what co-purifies from a lysate. Of HuRI interactions found in complexes of known structure, 90 percent were direct contacts, against 81 percent for literature binary pairs and under 50 percent for complex maps (Luck 2020). So an all-against-all census of one kind is not a census of the other. Neither contains kinetics. 1
Reported, not yet replicated
Why nobody has finished deep
Our reasoning, not a finding in any source. Pairs grow as n squared: 17,408 genes give about 151.5 million pairs (arithmetic from the HuRI clone-set size), and dividing HuRI's 52,569 by its own 2 to 11 percent gives roughly 0.5 to 2.6 million binary interactions in total (arithmetic). One pass misses interactions: HuRI needed three yeast two-hybrid versions and repeated screens. A single negative in yeast two-hybrid is weak evidence, and the fusion terminus can hide an interface, since sensitivity fell for interfaces within 20 angstroms of the fusion terminus. AlphaFold-based predictions can rank many more pairs, but the standard confidence scores were reported not to separate real interactions reliably from false ones. 16
Our inference
Observation
What is known
What is not
The fraction of the space actually tested in HuRI is not stated in the sources opened, the derivation of the 2 to 11 percent estimate was not read, and independent false-positive rates for HuRI and BioPlex were not found. 1
Our inference
What would settle it
A published estimate of total binary interactome size with a stated tested fraction, plus an independent re-run of a sample of pairs by a method with different biases. 1
Our inference
Still open here
- What fraction of the roughly 150 million possible human pairs did HuRI actually test, and how was the 2 to 11 percent estimate derived?
- Are there independent false-positive rates for HuRI and BioPlex?
- Does any map claim completeness for any species?
- Has anything like proteome-scale kinetics (kon, koff) been measured? See questions/protein-interaction-kinetics-at-proteome-scale.
Not yet written: Hypotheses, What is missing, Who works on it.
Linked from
- Affinity purification mass spectrometry (AP-MS)
- Crosslinking mass spectrometry (XL-MS)
- Genetic interaction and epistasis screens
- Isothermal titration calorimetry (ITC)
- Proximity labelling (BioID, TurboID, APEX)
- Structural evidence of complexes (crystallography and cryo-EM)
- Surface plasmon resonance (SPR)
- Yeast two-hybrid (Y2H)
- Pathways and interaction partners
- Model a single bacterial cell, end to end
- How do we know a pathway is a bona fide pathway?
- Has anyone measured kon and koff for interactions at proteome scale?
Sources
- Luck, K., Kim, D.-K., Lambourne, L., et al.. A reference map of the human binary protein interactome. Nature (2020). paper; opened 2026-11-17 all records citing this
- Huttlin, E. L., Bruckner, R. J., et al.. Dual proteome-scale networks reveal cell-specific remodeling of the human interactome. Cell (2021). paper; opened 2026-11-17 all records citing this
- Ito T, et al.. A comprehensive two-hybrid analysis to explore the yeast protein interactome. PNAS (2001). paper; opened 2026-11-17 all records citing this
- Uetz, P., Giot, L., et al.. A comprehensive analysis of protein-protein interactions in Saccharomyces cerevisiae. Nature (2000). paper; opened 2026-11-17 all records citing this
- Rolland, T., Tasan, M., Charloteaux, B., et al.. A proteome-scale map of the human interactome network. Cell (2014). paper; opened 2026-11-17 all records citing this
- Schmid, E. W., Walter, J. C.. Predicting protein-protein interactions with AlphaFold-Multimer: metrics and the SPOC classifier. Europe PMC (2025). paper; opened 2026-11-17 all records citing this
- Amit's dictation of 2026-11-17: what the atlas must contain. Kept in the atlas repository; not published.
Where it sits on the map
History of this record (7)
This record has been changed 7 times, on the dates below. The note kept against each change is written for the people who maintain the atlas, in their own shorthand, so it is not shown here; it is in the record itself.
- Kind
- amit
- Register
- open-question
- Status
- partly written
- Last updated
- 2026-11-19
- Record
- content/questions/has-every-protein-pair-been-tested.json