An Integrated Machine Learning Model To Spot Peptide Binding Pockets in 3D Protein Screening

Abstract

The prediction of peptide−protein binding sites is of utmost importance to tackle the onset of severe neurodegenerative diseases and cancer. In this work, we detail a novel machine learning model based on Linear Discriminant Analysis (LDA) demonstrating to be highly predictive in detecting the putative protein binding regions of small peptides. Starting from 439 high-quality pockets derived from peptide−protein crystallographic complexes, three sets of well-established peptidebinding regions were first selected through a Partitioning Around Medoids (PAM) clustering algorithm based on morphological and energetic 3D GRID-MIF molecular descriptors. Next, the best combination between all the putative interacting peptide pockets and related GRID-MIF scores was automatically explored by using the LDA-based protocol implemented in BioGPS. This approach proved successful to recognize the actual interacting peptide regions (that is, AUC = 0.86 and partial ROC enrichment at 5% of 0.48) from all the other pockets of the protein. Validated on two external collections sets, including 445 and 347 crystallographic peptide−protein complexes, our LDA-based model could be effective to further run peptide−protein virtual screening campaigns.

TIRESIA: An eXplainable Artificial Intelligence Platform for Predicting Developmental Toxicity

“…In recent years, the usage of Artificial Intelligence (AI) approaches to disclose Quantitative Structure–Activity Relationships (QSARs) has gained increasing importance. Recent successful applications can be found in predictive toxicology, − drug discovery, − molecular optimization − and de novo design, just to mention a few. An aspect deserving attention relates to the explainability of AI approaches, especially when dealing with high dimensional data and nonlinear relationships.…”

Growth hormone secretagogues modulate inflammation and fibrosis in mdx mouse model of Duchenne muscular dystrophy

“…The width of a given MIF depended on the energy of the interacting groups involved. This implied that the larger the MIF, the higher was the chance to find a complementary partner group in that molecular region ( 50 ). For the hydrophobic residue pairs, both S1’ pocket binding residues interacted with the hydrophobic CRY probe by producing a large MIF (dark green surface isocontours), which intercepted a region with a high possibility to detect hydrophobic groups (that were the indole groups of JMV2894 in both binding sites of ADAMTS-5 and MMP-9 crystal structures).…”

Small Molecules for the Treatment of Long-COVID-Related Vascular Damage and Abnormal Blood Clotting: A Patent-Based Appraisal

“…US11583517B2, ramatroban ( Figure 5 ), a dual DP2/TrP antagonist, blocking the deleterious effects of both PGD2 and TXA2 ( Figure 5 ) may be beneficial against COVID-19, as recently suggested by Gupta A. et al [ 45 ] ( Figure 6 ). Ramatroban ( 4 ) binds to thromboxane A2 receptors (TPrs) in human platelets ( Figure 7 ) with a K i value of 10–13 nM, and was found to antagonize the PGD2 receptor, significantly inhibiting the binding of [ 3 H]PGD2 with an IC 50 value of 100 nM [ 53 , 54 ]. It has a well-established safety profile and has been used orally as a treatment for allergic rhinitis in Japan since 2000 [ 55 ].…”