FGfuncLLM

Functional annotation of fusion genes is critical for understanding the role of the chimeric genes and developing targeted therapeutic strategies. Summarizing the function of known fusion genes requires extensive reading of the reported studies on these genes, and inferring the function of never-studied fusion genes requires imagination based on the extensive reading of fusion gene partners and related biomolecules. Due to the limited number of studied fusion genes (only ~ 1%), utilizing a large language model (LLM) is a feasible approach to help researchers understand the functional role of fusion genes. To maximize the potential of accurate functional annotation of LLM, human assistance is needed, such as providing a reasoning flow in a well-structured format, data based on the fluent biological knowledgebases. The major component of the fusion gene mechanism of action (MoA) is on the retained or lost protein functional features with abnormally regulated fusion gene expression, which are annotated in three levels of the central dogma of biology, like transcription factor swapping in DNA-level, miRNA regulation avoidance or swapping in RNA-level, and retained protein functional features in protein-level. We constructed a reasoning flow for the LLM to generate annotations of the potential mechanism of action and output the hallmark of cancer for the input fusion gene in five steps. For each step, we generated prompts with related biological knowledge from multiple reference resources. Our knowledge-based functional annotation of targetable human fusion genes (FGfuncLLM) provides all potential scenarios of MoAs of ~ 8K drug targetable fusion genes. Users can choose their most likely scenario; for the chosen one, FGfuncLLM provides all evidence and references for every step of reasoning and every scenario of the hallmark of cancer. We hope FGfuncLLM can help annotate the impact on tumorigenesis of human fusion genes and their oncogenic roles.

Search