Hierarchical graph learning for protein–protein interaction

Abstract

Protein-Protein Interactions (PPIs) are fundamental means of functions and signalings in biological systems. The massive growth in demand and cost associated with experimental PPI studies calls for computational tools for automated prediction and understanding of PPIs. Despite recent progress, in silico methods remain inadequate in modeling the natural PPI hierarchy. Here we present a double-viewed hierarchical graph learning model, HIGH-PPI, to predict PPIs and extrapolate the molecular details involved. In this model, we create a hierarchical graph, in which a node in the PPI network (top outside-of-protein view) is a protein graph (bottom inside-of-protein view). In the bottom view, a group of chemically relevant descriptors, instead of the protein sequences, are used to better capture the structure-function relationship of the protein. HIGH-PPI examines both outside-of-protein and inside-of-protein of the human interactome to establish a robust machine understanding of PPIs. This model demonstrates high accuracy and robustness in predicting PPIs. Moreover, HIGH-PPI can interpret the modes of action of PPIs by identifying important binding and catalytic sites precisely. Overall, “HIGH-PPI [https://github.com/zqgao22/HIGH-PPI]” is a domain-knowledge-driven and interpretable framework for PPI prediction studies.

Facile synthesis of a novel covalent triazine framework for recovering gold(i) from thiosulfate solution

“…7 Among them, adsorption has the advantages of economical and easy separation in terms of cost and recovery and has been widely used for metal recovery. 8,9 Shao et al 10 reported a strategy for the closed-loop recovery of Ag + using amorphous mixed-valence MoO x . MoO x exhibited an ultrahigh capture capacity (2605.91 mg g À1 ) and excellent selectivity for Ag + .…”

Machine Learning-Guided Protein Engineering

“…For example, edges corresponding to chemical bonds can guide a model toward learning chemical knowledge more rapidly or with less data than would otherwise be needed. Graph neural network (GNN) architectures have recently achieved state-of-the-art performance in multiple protein-related tasks, as exemplified by the DeepFRI and HIGH-PPI methods for predicting protein function and protein–protein interactions, respectively. Special graph-based protein representations, such as point clouds (no edges) or complete graphs (a full set of edges), are especially convenient for processing using powerful transformer models. , …”

pH-dependent adsorption of the sulfamethoxazole antibiotic on HKUST-1@CNS nanocomposite corroborating efficiency, mechanistic, and kinetic studies

“…14,15 Among the ones listed above, adsorption is an effective, easy-to-design, and cost-effective process for wastewater treatment. 16,17 It occurs at the interface between liquid and solid phases, transporting antibiotics to solid matrices like the soil, sediments, and plants. 18 For the successful removal of antibiotics from aquatic environments, several adsorbents have been created, including multiwalled carbon nanotubes, 19 nanofibers, 20,21 activated carbons, 22 zeolites, 23 cyclodextrin polymers, 24 and clay.…”