Back

On Correlation between Structural Properties and Viral Escape Measurements from Deep Mutational Scanning

Zhang, L.; Domeniconi, G.; Yang, C.-C.

2022-02-18 bioinformatics
10.1101/2022.02.17.480939 bioRxiv
Show abstract

Encouraged by recent efforts to map responses of SARS-CoV-2 mutations to various antibody treatments with deep mutational scanning, we explored the possibility of tying measurable structural contact information from the binding complexes of antibodies and their targets to experimentally determined viral escape responses. With just a single crystal structure for each binding complex, we find that the average correlation coefficient R is surprisingly high at 0.76. Our two methods for calculating contact information use binary contacts measured between all residues of two proteins. By varying the parameters to obtain binary contacts, we find that 3.6 [A] and 7 [A] are pivotal distances to toggle the binary step function when tallying the contacts for each method. The correlations are improved by short simulations ([~]25 ns), which increase average R to 0.78. With blind tests using the random forest model, we can further improve average R to 0.84. These easy-to-implement measurements can be utilized in computational screening of viral mutations that escape antibody treatments and potentially other protein-protein interaction problems.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.