Fine scale structural information substantially improves multivariate regression model for mRNA in-vial degradation prediction
Yi, S.; Ali, S.; Jadeja, Y.; Davis, J. W.; Metkar, M.
Show abstract
The success of COVID-19 mRNA vaccines has made optimizing mRNAs for in-vial stability a key objective. However, we still lack a complete understanding of the sequence metrics that influence mRNA stability in solution. RNA secondary structure plays a central role in protecting against hydrolysis, the primary degradation pathway under storage conditions. Yet, the structural metrics that best guide stability-focused mRNA design remain unclear. Global metrics like minimum free energy and average unpaired probability have improved mRNA stability but fail to capture local structural variation relevant to degradation. We show that base-pairing probability, in terms of log odds, provide fine-scale, orthogonal insight that complements global metrics and improves stability modeling. By combining local and global features into a four-feature regression model, dubbed STRAND (Stability Regression Analysis using Nucleotide-Derived features), we achieve substantial gains in predictive performance over current methods. This compact and interpretable model provides a practical framework for designing mRNAs with enhanced in-solution stability.
Matching journals
The top 5 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Learning cis-regulatory principles of ADAR-based RNA editing from CRISPR-mediated mutagenesis 96%
- Differential Analysis of RNA Structure Probing Experiments at Nucleotide Resolution: Uncovering Regulatory Functions of RNA Structure 96%
- Generative and predictive neural networks for the design of functional RNA molecules 96%
Similar papers in this journal
Similar papers in this journal
Similar papers in this journal
- Global Importance Analysis: An Interpretability Method to Quantify Importance of Genomic Features in Deep Neural Networks 96%
- RNA structure prediction using positive and negative evolutionary information 95%
- Structural comparison of protein-RNA homologous interfaces reveals widespread overall conservation contrasted with versatility in polar contacts 92%
"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.