Predicting stability of nucleic acid biopharmaceutical products using advanced computational modeling techniques

Modified oligonucleotides increase efficacy of nucleic acid drugs. These modifications impact degradation resistance, enhance thermodynamic stability, and render oligonucleotides unrecognizable to the immune system. However, not much is known about how these chemical modifications affect structure-function relationships. In theory, specific modifications can be leveraged on a sequence basis to increase efficacy by better mimicking the unmodified RNA structure and dynamics. To date, siRNA and most ASO RNA drug products are relatively short, < 40 base pairs in length. Based on previous research (Bergonzo and Grishaev, 2019, JBNMR and JSB), we can generate conformational ensembles of dsRNA which reproduce experimental observables from methods such as NMR and SAXS.

The first step in modeling modified oligonucleotides is creating potential functions for molecular dynamics force fields. Recently completed work introduced modXNA, a program that facilitates parameterization by modularizing the nucleic acid into three groups: a phosphate backbone, a ribose or deoxyribose sugar, and a nucleobase. We have pre-parameterized modifications for each of these three groups and created a procedure for stitching together one module from each group to create the full modified residue, ready for incorporation into the nucleic acid strand and further dynamics. We have additionally submitted a “living” tutorial for its use: https://doi.org/10.33011/livecoms.6.1.4545. This work is ongoing – we are parameterizing new modules to reflect changing industry modifications and therapeutic applications.