A CNN model for predicting binding affinity changes between SARS-CoV-2 spike RBD variants and ACE2 homologues

Abstract

The cellular entry of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) involves the association of its receptor binding domain (RBD) with human angiotensin converting enzyme 2 (hACE2) as the first crucial step. Efficient and reliable prediction of RBD-hACE2 binding affinity changes upon amino acid substitutions can be valuable for public health surveillance and monitoring potential spillover and adaptation into non-human species. Here, we introduce a convolutional neural network (CNN) model trained on protein sequence and structural features to predict experimental RBD-hACE2 binding affinities of 8,440 variants upon single and multiple amino acid substitutions in the RBD or ACE2. The model achieves a classification accuracy of 83.28% and a Pearson correlation coefficient of 0.85 between predicted and experimentally calculated binding affinities in five-fold cross-validation tests and predicts improved binding affinity for most circulating variants. We pro-actively used the CNN model to exhaustively screen for novel RBD variants with combinations of up to four single amino acid substitutions and suggested candidates with the highest improvements in RBD-ACE2 binding affinity for human and animal ACE2 receptors. We found that the binding affinity of RBD variants against animal ACE2s follows similar trends as those against human ACE2. White-tailed deer ACE2 binds to RBD almost as tightly as human ACE2 while cattle, pig, and chicken ACE2s bind weakly. The model allows testing whether adaptation of the virus for increased binding with other animals would cause concomitant increases in binding with hACE2 or decreased fitness due to adaptation to other hosts.

Early warning of emerging infectious diseases based on multimodal data

“… [60] SAS https://www.biosino.org/sas A platform which can predict the resistant effect of emerging variants and the dynamic coverage of SARS-CoV-2 antibodies among circulating strains. [61] STHAM https://github.com/uofu-ccts/prisms-comp-model-stham Spatio-temporal human activity model (STHAM), a kind of extended agent-based model which enables to simulate SARS-CoV-2 transmission dynamics. [67] …”

“…Spike protein Antigenicity for the SARS-CoV-2 (SAS) platform provided by Zhang et al [60] can predict the antigenicity of SARS-CoV-2 variants and help researchers better target immunobinding experiments to monitor the effectiveness of vaccine antibodies. Besides, Chen et al [61] predicted binding affinity of SARS-CoV-2 receptor binding domains or ACE2 with different amino acid substitutions by introducing a convolutional neural network (CNN) model for training on protein sequence and structural characteristics.…”

SARS-CoV-2 Omicron Subvariants Do Not Differ Much in Binding Affinity to Human ACE2: A Molecular Dynamics Study

“…The interaction between ACE2 and RBD of Omicron BA.1, BA.2, and BA.3 was analyzed using docking and molecular dynamics (MD) simulations . Using an artificial intelligence model and docking simulation, it was shown that due to a higher binding ability to the host cell, Omicron is more contagious than the WT virus. , The artificial intelligence techniques were used to predict binding affinity changes between ACE2 and SARS-CoV-2 variants . The deep learning method was utilized to obtain the change in binding free energy caused by mutations in Omicron subvariants .…”

“… 29 , 30 The artificial intelligence techniques were used to predict binding affinity changes between ACE2 and SARS-CoV-2 variants. 31 The deep learning method was utilized to obtain the change in binding free energy caused by mutations in Omicron subvariants. 32 Combination of MD simulation with the molecular mechanics Poisson–Boltzmann/generalized Born surface area (MM-PB/GBSA) methods can explain why Omicron BA.1 binds to ACE2 more strongly than the ancestral strain.…”