Search

Objective: Most existing fine-tuned biomedical large language models (LLMs) focus on enhancing performance in monolingual biomedical question answering and conversation tasks. To investigate the effectiveness of the fine-tuned LLMs on diverse biomedical NLP tasks in different languages, We present Taiyi, a bilingual fine-tuned LLM for diverse biomedical tasks. Materials and Methods: We first curated a comprehensive collection of 140 existing biomedical text mining datasets (102 English and 38 Chinese datasets) across over 10 task types. Subsequently, a two-stage strategy is proposed for supervised fine-tuning to optimize the model performance across varied tasks. Results: Experimental results on 13 test sets covering named entity recognition, relation extraction, text classification, question answering tasks demonstrate that Taiyi achieves superior performance compared to general LLMs. The case study involving additional biomedical NLP tasks further shows Taiyi's considerable potential for bilingual biomedical multi-tasking. Conclusion: Leveraging rich high-quality biomedical corpora and developing effective fine-tuning strategies can significantly improve the performance of LLMs within the biomedical domain. Taiyi shows the bilingual multi-tasking capability through supervised fine-tuning. However, those tasks such as information extraction that are not generation tasks in nature remain challenging for LLM-based generative approaches, and they still underperform the conventional discriminative approaches of smaller language models.

Transcription factors (TFs) activate enhancers to drive cell-specific gene programs in response to signals, but our understanding of enhancer assembly during signaling events is incomplete. Here, we show that Androgen Receptor (AR), a steroid hormone-regulated transcription factor, forms condensates through multivalent interactions in response to androgen signaling to orchestrate enhancer assembly. We demonstrate that the intrinsically disordered N-terminal domain (NTD) of AR drives 1,6-Hexanediol-sensitive condensate formation and that NTD deletion or aromatic residue mutation reduces AR self-association and abolishes AR transcriptional activity. AR NTD can be substituted by intrinsically disordered regions (IDRs) from selective proteins for AR condensation capacity and transactivation function. Surprisingly, strengthened AR condensation capacity caused by extending the polyQ tract within AR NTD also leads to impaired transcriptional activity without affecting AR binding on enhancers. Furthermore, either NTD deletion or polyQ extension reduces heterotypic multivalent interactions between AR and other enhancer components. These results suggest the importance of an optimal level of AR condensation in mediating AR-AR homotypic and AR-cofactor heterotypic interactions to regulate enhancer assembly in response to signals. Our study supports the notion that alteration of the fine-tuned multivalent IDR-IDR interactions might underlie AR-related human pathologies, thereby providing novel molecular insights for potential therapeutic strategies to treat prostate cancer and other AR-involved diseases by targeting AR multivalent interactions.