The Negatome Database


The Negatome is a collection of protein and domain pairs which are unlikely engaged in direct physical interactions. The database currently contains experimentally supported non-interacting protein pairs derived from two distinct sources: by manual curation of literature and by analysing protein complexes from the PDB. More stringent lists of non-interacting pairs were derived from these two datasets by excluding interactions from IntAct. It can be used to evaluate newly derived experimental interactions. The Negatome is much less biased towards functionally dissimilar proteins than the negative data derived by randomly selecting proteins from different cellular locations. Thus, the negatome is complementary to such random data for training protein interaction prediction algorithms.


The supplement to the paper can be downloaded here.

References and Data Used


We provide a 'Linked view' on the left side of the table below. Text download can be done on the right.
Dataset Derived from Description Number of Pairs
Manual Manual literature annotation Manually annotated literature data describing the lack of protein interaction. High-throughput data are not included. The data is restricted only to mammalian proteins. 1291
Manual-stringent Manual The Manual dataset filtered against the IntAct dataset 1162
Manual-PFAM Manual-stringent PFAM domain pairs found in the Manual dataset filtered using iPFAM and 3did 523
PDB The PDB database Protein pairs that are members of at least one structural complex but do not interact directly. Organism of origin is not restricted 809
PDB-stringent PDB The PDB dataset filtered against the IntAct dataset. 745
PDB-PFAM PDB-stringent Non-interacting PFAM domains found in the same structural complex filtered using iPFAM and 3did 458
Combined Manual and PDB-stringent A combined non-interacting Protein dataset 1892
Combined Pfam Manual-PFAM and PDB-PFAM A combined non-interacting Protein domain dataset 979

Additional Mappings

Additional mappings to IntAct, KEGG, GO for our data can be found here and as text download here .


This work was partially funded by the Biosapiens Network of Excellence. We thank Roland Arnold for the name and Alexei Antonov.