Researchers from CSE, Professor Yue Dong and her PhD student Haz Sameen Shahgir, have developed BiRNA-BERT, a new RNA language model capable of processing full-length RNA sequences without the truncation limitations seen in prior models. This breakthrough allows the model to handle long non-coding RNAs (lncRNAs) in their entirety, achieving state-of-the-art performance in lncRNA interaction prediction.

Ongoing work aims to extend BiRNA-BERT to RNA-based homology search, opening new possibilities for studying evolutionary relationships that were previously inaccessible to language-model-based methods.