Robust computational design and evaluation of peptide vaccines for cellular immunity with application to SARS-CoV-2.

摘要: We present a combinatorial machine learning method to evaluate and optimize peptide vaccine formulations, we find for SARS-CoV-2 that it provides superior predicted display of viral epitopes by MHC class I II molecules over populations when compared other candidate vaccines. Our is robust idiosyncratic errors in the prediction considers target population HLA haplotype frequencies during optimization. To minimize clinical development time our methods validate vaccines with multiple presentation algorithms increase probability will be effective. an objective function based on likelihood diverse set peptides conditioned distribution expected epitope drift. produce separate formulations loci (HLA-A, HLA-B, HLA-C) (HLA-DP, HLA-DQ, HLA-DR) permit signal sequence cell compartment targeting using nucleic acid platforms. provide 93.21% coverage at least five peptide-HLA hits average individual (≥ 1 99.91%) all perfectly conserved across 4,690 geographically sampled genomes. 90.17% having observed mutation ≤ 0.001. 29 previously published designs evaluation tool requirement per individual, they have maximum 58.51% 71.65% given analysis. open source implementation design (OptiVax), (EvalVax), as well data used efforts.