Comprehensive benchmarking of large language models for RNA secondary structure prediction
This work provides a comparative analysis for researchers in computational biology, but it is incremental as it applies existing methods to a specific domain task.
The authors benchmarked several pre-trained large language models (LLM) for RNA secondary structure prediction, finding that two LLM outperformed others but faced challenges in low-homology scenarios.
Inspired by the success of large language models (LLM) for DNA and proteins, several LLM for RNA have been developed recently. RNA-LLM uses large datasets of RNA sequences to learn, in a self-supervised way, how to represent each RNA base with a semantically rich numerical vector. This is done under the hypothesis that obtaining high-quality RNA representations can enhance data-costly downstream tasks. Among them, predicting the secondary structure is a fundamental task for uncovering RNA functional mechanisms. In this work we present a comprehensive experimental analysis of several pre-trained RNA-LLM, comparing them for the RNA secondary structure prediction task in an unified deep learning framework. The RNA-LLM were assessed with increasing generalization difficulty on benchmark datasets. Results showed that two LLM clearly outperform the other models, and revealed significant challenges for generalization in low-homology scenarios.