The Phosphosite Library

To identify phosphorylation sites that might regulate AlphaFold-Multimer's protein-protein interaction (PPI) predictions in our Genome maintenance dataset, we mapped experimentally determined phosphosites onto the predicted interfaces. We then constructed a set of phosphosites which we predict stimulate PPIs with our novel deep learning classifier, PhosNet, which uses spatial and chemical features from the phosphosite’s local structure prediction. In addition, we applied simple screening criteria to find phosphosites which might inhibit our PPI predictions. The table below of predicted phospho-regulated interactions can be sorted by diverse features of each protein pair. Clicking on a pair retrieves an interactive structure viewer and information about the putative regulatory kinase.

Filters add clear

Select genotoxic treatment

Aphidicolin (APH) Gemcitabin (GEM) Hydroxyurea (HU) Camptothecin (CPT) Etoposide (ETO) Ionizing radiation (X-ray) Formaldehyde (FA) Methy methanesulfonate (MMS) Ultraviolet radiation (UV-C) Hydrogen peroxide (H2O2) Sodium Arsenite (AsO2)

Chain A / Chain B i	phos_chain i	phos_res i	contact i	SPOC i	models i	avg_PAE i	avg_pLDDT i	ht-lt_refs i	PhosNet i	kinase i	treatment i