However, the emergence of a growing number of experimentally validated inhibitors of PPIs (i-PPIs) allows drug designers to use chemoinformatics and machine learning technologies to unravel the nature of the chemical space covered by the reported compounds.
For example, if there are 300 PPIs from the STRING predicted functional associations dataset are in score range 150~160, and among these 300 PPIs there are 30 PPIs overlapping with the B2H PPI dataset, then in this score range 150~160 the "overlap PPI number ratio" is 30/300 = 0.1.
AImed extracts the most PPIs (2957), and LLL extracts 1871 PPIs despite only containing 300 annotated PPIs.
Functional associations between genes can be inferred from two types of protein protein interaction (PPI) data sources: (i) small/medium scale PPIs (LC) derived from curated PPI databases such as IntAct (20), DIP (21), MINT (22), BioGRID (23); (ii) large-scale high-throughput PPIs (HT) inferred by affinity purification-mass spectrometry (24 26) and yeast two hybrid analysis (27).
To explain this difference researchers have noted that PPIs cause prolonged, profound hypochlorhydria [ 34] whereas the effect of H2 blockers is short lived – a transient hypochlorhydria [ 35].
Focusing on datasets of a particular complex type, combining interface classification with interaction likelihood prediction, and integrating learning of the different properties of the interface core and rim into PPIS prediction are just a few examples of promising areas that are currently under investigation.
We built the interactome network from the PPIs, and computed the centrality measures of PPIs (edges) and the participant proteins (nodes).
Differential protein protein interaction (PPI) networks (DNs) have been used to delineate PPIs (edges) among differentially regulated nodes (DRNs), such as up- or downregulated genes or proteins.
Yet, PPIs show characteristic pharmacokinetic properties that may specifically influence the potency of cardiac effects.
To test the utility of accounting for core-rim differences, the authors used the Intervor [ 240] algorithm (which computes distance to the interface rim via Voronoi shelling order) with PresCont, detecting differences in propensities (as found previously [ 198]), but ultimately not recommending use of core residues alone for training when the full PPIS is desired.
