A brand-new study called “Biomedical generative pre-trained based transformer language model for age-related disease target discovery” was published in the journal Aging.
The creation of novel treatments and diagnostics depends on the identification of targets. However, the effectiveness, specificity, and scalability of current techniques are frequently constrained, making the investigation of cutting-edge methods for locating and confirming disease-relevant targets necessary. Natural language processing innovations have opened up new ways to identify possible treatment targets for a range of disorders.
Researchers from Insilico Medicine Diana Zagirova, Stefan Pushkov, Geoffrey Ho Duen Leung, Bonnie Hei Man Liu, Anatoly Urban, Denis Sidorenko, Aleksandr Kalashnikov, Ekaterina Kozlova, Vladimir Naumov, Frank W. Pun, Ivan V. Ozerov, Alex Aliper, and Alex Zhavoronkov present a novel method for anticipating therapeutic targets using an LLM in their most recent research.
The study authors stated, “We trained a domain-specific BioGPT model on a large corpus of biomedical literature consisting of grant text and developed a pipeline for generating target prediction.”
The results of this study show that pre-training the LLM model with task-specific texts increases its performance. The researchers used the created pipeline to obtain projected aging and age-related disease targets and demonstrated that these proteins corresponded to the database data.
Furthermore, they suggest CCR5 and PTH as novel dual-purpose anti-aging and disease targets that had not previously been identified as age-related but were highly rated in their methodology.
The study authors further added, “Overall, our work highlights the high potential of transformer models in novel target prediction and provides a roadmap for future integration of AI approaches for addressing the intricate challenges presented in the biomedical field.”
Zagirova, D., et al. (2023) Biomedical generative pre-trained based transformer language model for age-related disease target discovery. Aging. doi:10.18632/aging.205055