Researchers at the Carl R. Woese Institute for Genomic Biology have successfully employed generative artificial intelligence (AI) to design new mitochondrial targeting sequences (MTSs) [1, 4, 5]. This innovation promises to enhance the study and manipulation of mitochondria, the cell's energy-producing organelles [1, 5, 7].
The AI identified key characteristics of effective MTSs, such as a positive charge and a propensity to form an α-helix [1, 2]. The team then created a million AI-generated MTSs and experimentally assessed the mitochondrial targeting capabilities of 41 of these sequences, achieving a 50% to 100% success rate across yeast, plant, and mammalian cells [1, 6].
These AI-designed sequences have been successfully applied in metabolic engineering and protein delivery, opening new possibilities for therapeutic applications and a deeper understanding of mitochondrial evolution [1, 2, 5]. This study, published in May 2025, underscores AI's potential in synthetic biology and biotechnology by addressing the limited availability of diverse MTSs, which is crucial for efficient research and development in mitochondrial biology [1, 4, 7].