SMMPPI: a machine learning-based approach for prediction of modulators of protein–protein interactions and its application for identification of novel inhibitors for RBD:hACE2 interactions in SARS-CoV-2

Abstract

Small molecule modulators of protein–protein interactions (PPIs) are being pursued as novel anticancer, antiviral and antimicrobial drug candidates. We have utilized a large data set of experimentally validated PPI modulators and developed machine learning classifiers for prediction of new small molecule modulators of PPI. Our analysis reveals that using random forest (RF) classifier, general PPI Modulators independent of PPI family can be predicted with ROC-AUC higher than 0.9, when training and test sets are generated by random split. The performance of the classifier on data sets very different from those used in training has also been estimated by using different state of the art protocols for removing various types of bias in division of data into training and test sets. The family-specific PPIM predictors developed in this work for 11 clinically important PPI families also have prediction accuracies of above 90% in majority of the cases. All these ML-based predictors have been implemented in a freely available software named SMMPPI for prediction of small molecule modulators for clinically relevant PPIs like RBD:hACE2, Bromodomain_Histone, BCL2-Like_BAX/BAK, LEDGF_IN, LFA_ICAM, MDM2-Like_P53, RAS_SOS1, XIAP_Smac, WDR5_MLL1, KEAP1_NRF2 and CD4_gp120. We have identified novel chemical scaffolds as inhibitors for RBD_hACE PPI involved in host cell entry of SARS-CoV-2. Docking studies for some of the compounds reveal that they can inhibit RBD_hACE2 interaction by high affinity binding to interaction hotspots on RBD. Some of these new scaffolds have also been found in SARS-CoV-2 viral growth inhibitors reported recently; however, it is not known if these molecules inhibit the entry phase.

pdCSM-PPI: Using Graph-Based Signatures to Identify Protein–Protein Interaction Inhibitors

“…The performances of each target-specific predictive model were further evaluated on nonredundant blind test sets, and outcomes were compared with those reported for PPIM-pred and SMMPPI . PPIM-pred is a ML-based predictor for inhibitors targeting three distinct PPIs: Bcl2/Bak, Mdm2/P53, and c-Myc/Max.…”

“…Experimentally characterized PPI inhibitors were retrieved from TIMBAL, iPPI-DB, and 2P2I-DB v2 (Figure S16), comprising 4965 small molecules targeting 51 distinct PPIs (Table S35). Inhibitors for each PPI were clustered separately using the Butina algorithm with a Tanimoto similarity cutoff of 0.8, based on Morgan molecular fingerprints (1024 bits) with a radius size of 2, with one compound selected as a cluster representative to minimize redundancy. PPIs with at least 20 nonredundant inhibitors were selected to build class-specific predictors.…”

“…Since then, many efforts to increase data availability for small molecules targeting PPIs have been proposed, most notably 2P2I-DB with over 274 inhibitors targeting 27 different PPIs and iPPI-DB with over 2378 inhibitors acting on 46 PPI families. The emergence of large collections of data prompted the development of more robust data-driven methods such as PPIMpred and SMMPPI . The former used 11 chemical descriptors with SVM to build predictive models to identify inhibitors for three oncogenic PPIs: Bcl2/Bak, mdm2/P53, and c-Myc/Max.…”

Prediction of coronavirus 3C-like protease cleavage sites using machine-learning algorithms

“…The RF algorithm is a robust machine learning algorithm that can be used for a variety of tasks including regression and classification. It was extensively used in coronavirus research such as infection risk prediction ( Qiang et al, 2020 ), disease diagnosis ( Rosado et al, 2021 ), origin identification ( El Boujnouni et al, 2021 ), drug development ( Gupta and Mohanty, 2021 ), and so on. The RF model developed here used the cleavage sites on polyproteins from 14 coronaviruses for modeling which were more than three times to that used in Kiemer’s study ( Kiemer et al, 2004 ).…”

Aurora A and AKT Kinase Signaling Associated with Primary Cilia

“…Binding of AurkinA causes mislocalization of AURKA from the microtubules in mitotic spindles and inhibits its catalytic activity without affecting ATP binding [ 11 ]. Novel approaches are currently being developed to find chemical spaces in AURKA that can modulate its protein–protein interactions [ 199 , 200 , 201 ].…”