In Silico designing the future of RNA therapeutics

Ribonucleic acid (RNA) is a pivotal molecule of life, involved in almost all aspects of cell biology. RNA therapeutics comprise a rapidly expanding category of drugs that is believed to change the standard of care for many diseases and actualize personalized medicine. One example was the development of mRNA Covid-19 vaccines, adapting rapidly to a new virus that threatened human life.

The development of RNA therapeutics required that several major hurdles be overcome, specifically the rapid degradation of exogenous RNA by RNases that are ubiquitous in the environment and tissues; delivery of negatively charged RNA across hydrophobic cytoplasmic membrane; and immunogenicity of exogenous RNA that may cause adverse effects. Therefore, structure design algorithms to predict ribonucleic acid stability can effectively improve their efficiency. Moreover, RNA designs depend on a unique structure to become functional. However, designed sequences rarely fold into the intended structure either in vivo or in vitro. This is why it’s essential to hasten the search and development of robust de novo RNA design procedures to accelerate the development of practical RNA tools.

MoiRNAiFold is built on the same concept and philosophy as RNAiFold, except that it incorporates novel ideas that improve its efficacy. Several RNA design tools and algorithms have been developed over the years to solve the RNA design shortfall. Of the developed approaches, NUPACK is the most preferred in designing functional RNA sequences, although it doesn’t seem effective in benchmarks.

In light of this, scientists from Moirai Biodesign in Spain, Dr. Gerard Minuesa, Dr.  Cristina Alsina, and Dr. Ivan Dotu together with Dr. Juan Antonio Garcia-Martin, and Professor Juan Carlos Oliveros from the National Centre for Biotechnology developed MoiRNAiFold, a constraint programming-based RNA design tool with novel modeling approaches. MoiRNAiFold aims at solving complex benchmark sets better than its predecessors while introducing a myriad of design constraints and new quality standards urgently needed for functional RNAs design. Their study aimed to design RNA sequences for multiplex RNA structures while producing in vivo and in vitro validated RNA sequences simultaneously. Their research work is now published in the journal Nucleic Acid Research.

MoiRNAiFold incorporates groundbreaking variable types, restart methods, and heuristics for Large Neighborhood Search. The research team focused on toehold switches RNA constructs to experimentally validate their work. They also presented a Small Transcriptional Activating RNAs (STARs) example to validate that transcriptional control can also be effective by RNA structures by the proposed algorithm.

The authors applied MoiRNAiFold in solving EteRNA benchmark puzzles. The tool solved 91 puzzles within a day, 84 within a minute, and 12 for the first time. Its predecessors only solved 43 puzzles, clearly showing the impact of the novel approaches incorporated in MoiRNAiFold, particularly the new strategies in Large Neighborhood Search, GUPs, and the novel heuristics. MoiRNAiFold performed exemplary well in smaller benchmarks, with most structures being solved within one second. It was the only software that solved all puzzles with a probable design by existing algorithms and thermodynamic models. Gerard Minuesa and colleagues validated their work experimentally by investigating RNA gene expression regulation through RNA-RNA interactions. From the study, MoiRNAiFold designed biologically relevant and functional RNA structures successfully fixing the apparent gaps between theoretical puzzles and functional designs.

The Moirai Biodesign scientists could generate translation-activating 3-way junction repressors in vitro without necessarily preparing and analyzing a large assembly of constructs. Remarkably, the three constructs they generated outperformed those reported in literature in vivo. This further reinforced MoiRNAiFold software’s potency in developing functional RNA designs suitable for in vivo settings. This new study paves the way to efficiently generate highly structured RNA molecules that no other software can design. In a statement to Medicine Innovates, Dr Dotu said The RNA Therapeutics market has seen tremendous growth in recent years. Functional, structured RNAs have been (and still are) designed either by modifying naturally occurring RNAs, with a few structural rules and through extensive trial and error, or via SELEX (mostly for aptamers). Here, we provide a versatile tool for fast and exhaustive design of novel structured RNAs with dozens of design constraints and functionality measures. The goal is to reduce time and cost from concept to market.