Computational Prediction of the Comprehensive SARS-CoV-2 vs. Human Interactome to Guide the Design of Therapeutics

Abstract

Understanding the disease pathogenesis of the novel coronavirus, denoted SARS-CoV-2, is critical to the development of anti-SARS-CoV-2 therapeutics. The global propagation of the viral disease, denoted COVID-19 ("coronavirus disease 2019"), has unified the scientific community in searching for possible inhibitory small molecules or polypeptides. Given the known interaction between the human ACE2 ("Angiotensin-converting enzyme 2") protein and the SARS-CoV virus (responsible for the coronavirus outbreak circa. 2003), considerable focus has been directed towards the putative interaction between the SARS-CoV-2 Spike protein and ACE2. However, a more holistic understanding of the SARS-CoV-2 vs. human inter-species interactome promises additional putative protein-protein interactions (PPI) that may be considered targets for the development of inhibitory therapeutics.To that end, we leverage two state-of-the-art, sequence-based PPI predictors (PIPE4 & SPRINT) capable of generating the comprehensive SARS-CoV-2 vs. human interactome, comprising approximately 285,000 pairwise predictions. Of these, we identify the high-scoring subset of human proteins predicted to interact with each of the 14 SARS-CoV-2 proteins by both methods, comprising 279 highconfidence putative interactions involving 225 human proteins. Notably, the Spike-ACE2 interaction was the highest ranked for both the PIPE4 and SPRINT predictors, corroborating existing evidence for this PPI. Furthermore, the PIPE-Sites algorithm was used to predict the putative subsequence that might mediate each interaction and thereby inform the design of inhibitory polypeptides intended to disrupt the corresponding host-pathogen interactions.We hereby publicly release the comprehensive set of PPI predictions and their corresponding PIPE-Sites landscapes in the following DataVerse repository: 10.5683/SP2/JZ77XA. All data and metadata are released under a CC-BY 4.0 licence. The information provided represents theoretical modeling only and caution should be exercised in its use. It is intended as a resource for the scientific community at large in furthering our understanding of SARS-CoV-2. : bioRxiv preprint Promisingly, many computational approaches have been rapidly deployed to increase our understanding of SARS-CoV-2, including protein function, three-dimensional (3D) protein structures, and possible target regions for small inhibitory molecules [2,3]. Through the use of publication preprint platforms, this information can be immediately disseminated, albeit, with the disclaimer of "non-peer-reviewed" research. Two notable examples include the use of DeepMind's recently published AplhaFold protein structure predictor [2] to predict the 3D protein structure of each of the SARS-CoV-2 proteins, and the use of the SUMMIT high-performance computing (HPC) infrastructure to perform large-scale virtual docking simulations as a form of high-throughput screening to identify small inhibitory molecules [3]. Given that the Spike protein from the original SARS coronavirus, SARS-CoV, is ...

Multi-schema computational prediction of the comprehensive SARS-CoV-2 vs. human interactome

“…We highlight those 279 pairs within the predicted all interactome, the 129 pairs within the predicted proximal schema, and the 539 pairs within the predicted RP-PPI schema. All data are published in the following DataVerse repository, for broader use by the scientific community ( Dick, Biggar & Green, 2020 ).…”

“…For example, among the tables for the all schema, the Molecular Functions exhibiting a fold enrichment greater than 3 are reported; the Biological Processes exhibiting a fold enrichment greater than 50 are reported; and the Cellular Components exhibiting a fold enrichment greater than 15 are reported. The fold enrichment cut-offs were selected to limit the size of the tables; the complete tables are available in the appendix of the Supplemental Information and at at public repository, Dick, Biggar & Green (2020) .…”

“…In the present study, of the ∼285,000 host-virus pairs, we leverage three prediction schema and two independent PPI predictors to select a highly conservative set of predicted interactions for each of the 14 SARS-CoV-2 proteins considered in this study resulting in the identification of several putative human protein targets. We have publicly released these predictions and related meta-data for use by the broader scientific community in the following DataVerse repository: https://www.doi.org/10.5683/SP2/JZ77XA ( Dick, Biggar & Green, 2020 ).…”

Bioinformatics resources facilitate understanding and harnessing clinical research of SARS-CoV-2

“…Besides using the experimental results of V-H PPIs, computationally predicted and curated PPIs are also being used for analyzing the pathogenesis of SARS-CoV-2 [ 33 ]. Several computational strategies, such as structure-, sequence- and evolution-based methods, have been conducted to analyze and quantify the intra-viral and V-H PPIs [ 31 , 32 , 34 ]. Structure-based methods for PPIs require the information on the 3D structure of both proteins.…”

“…Dick et al . [ 32 ] have used two sequence-based PPI prediction approaches such as PIPE4 (Protein-Protein Interaction Prediction Engine) and SPRINT (Scoring PRotein INTeractions) to predict host protein interactions with 14 SARS-CoV-2 proteins and identified 279 putative interactions connecting with 225 human proteins. To train the predictors and infer new putative V-H interactions, experimentally elucidated human-virus PPIs of several viral organisms were used those obtained from the VirusMentha database ( https://virusmentha.uniroma2.it/ ).…”