Add to Chrome
Add to Firefox
Add to EdgeMessenger and Non-Coding RNA Design via Expected Partition Function and Continuous Optimization
Dec 29, 2023
The tasks of designing messenger RNAs and non-coding RNAs are discrete optimization problems, and several versions of these problems are NP-hard. As an alternative to commonly used local search methods, we formulate these problems as continuous optimization and develop a general framework for this optimization based on a new concept of "expected partition function". The basic idea is to start with a distribution over all possible candidate sequences, and extend the objective function from a sequence to a distribution. We then use gradient descent-based optimization methods to improve the extended objective function, and the distribution will gradually shrink towards a one-hot sequence (i.e., a single sequence). We consider two important case studies within this framework, the mRNA design problem optimizing for partition function (i.e., ensemble free energy) and the non-coding RNA design problem optimizing for conditional (i.e., Boltzmann) probability. In both cases, our approach demonstrate promising preliminary results. We make our code available at https://github.com/KuNyaa/RNA_Design_codebase.
