Texas A&M AgriLife Research team's predicted protein partners could speed research on diseases tied to mitochondria
Introduction Artificial intelligence is helping Texas A&M AgriLife Research scientists quickly find answers to a long-standing question in human biology: What do the hundreds of mysterious proteins within mitochondria actually do? For years, scientists have known that about 1,100 proteins exist within mitochondria, the energy centers of human cells, but the function of many remains unknown. This lack of knowledge is a significant problem because mutations in nearly 30% of mitochondrial proteins have been linked to various human diseases, including common conditions like cancer, Parkinson’s, and Alzheimer’s, as well as rare inherited metabolic disorders. Pairing up proteins for answers Understanding the precise biochemical function of these proteins is crucial to comprehending the causes of related diseases. One effective method to infer a protein’s function is to identify other proteins it interacts with. In their study, “The predicted interactome of the human mitochondrial proteome,” published in Nature Communications, the research team utilized an AI tool to screen over 630,000 potential pairings among approximately 1,100 human mitochondrial proteins. The AI’s predictions identified nearly 2,900 likely interactions, linking many disease-associated proteins to previously unknown biological pathways. This breakthrough points to potential diagnostic and therapeutic targets. From a basic science perspective, the study also provided the first probable partners for 85 previously uncharacterized proteins, offering vital clues about their roles within cells. Abhinav Swaminathan, Ph.D., the study’s lead author and an AgriLife Research scientist, emphasized that without knowing the function of individual proteins, determining effective disease treatments remains challenging. Repurposing an AI tool that predicts protein shape Researchers globally have been working to identify the functions of these mysterious mitochondrial proteins, but traditional laboratory methods are inherently slow. To accelerate this process, Swaminathan and Gohil initiated a project nearly three years ago to test an artificial intelligence tool known as AlphaFold-Multimer (AFM). Google DeepMind originally developed and released AFM in 2022 to accurately predict the three-dimensional shapes that proteins form when they interact. Swaminathan explained that their lab lacked reliable experimental tools for evaluating direct mitochondrial protein interactions, especially for the mitochondrial copper pathway proteins they studied. A trial run using the AI approach proved highly successful, prompting them to apply it to all human mitochondrial proteins. Out of the 630,003 possible protein pairs, the team cataloged 2,895 likely interactions, with more than half of these being entirely new discoveries never before reported. Confirming AI predictions in the lab Following the AI predictions, about two years of intensive laboratory experiments were conducted. These experiments successfully confirmed several of the predicted protein partnerships in both yeast and human cells, with some validations performed in collaboration with researchers at Washington University School of Medicine in St. Louis. The findings demonstrated that AI predictions can serve as an invaluable roadmap, guiding scientists toward laboratory experiments most likely to yield significant results efficiently. To further aid the scientific community, the team has made its predictions publicly available as a free online resource called MitoMatch, enabling researchers worldwide to identify likely partners for the proteins they are studying. As the team continues to experiment with the vast dataset, they are also actively collaborating with other researchers to deepen the understanding of cell function and pave the way for new insights into human diseases. While cures for mitochondria-linked diseases may still be distant, Swaminathan considers the team’s recent success with their AI-driven approach a promising start. He noted that their lab, traditionally focused on biochemistry and genetics, achieved this success by creatively addressing the protein function problem with accessible AI resources, providing a solution for broader use. This groundbreaking study received support from the National Institutes of Health, Texas A&M High Performance Research Computing, and The Welch Foundation.
Introduction
Artificial intelligence is helping Texas A&M AgriLife Research scientists quickly find answers to a long-standing question in human biology: What do the hundreds of mysterious proteins within mitochondria actually do? For years, scientists have known that about 1,100 proteins exist within mitochondria, the energy centers of human cells, but the function of many remains unknown. This lack of knowledge is a significant problem because mutations in nearly 30% of mitochondrial proteins have been linked to various human diseases, including common conditions like cancer, Parkinson’s, and Alzheimer’s, as well as rare inherited metabolic disorders.
Pairing up proteins for answers
Understanding the precise biochemical function of these proteins is crucial to comprehending the causes of related diseases. One effective method to infer a protein’s function is to identify other proteins it interacts with. In their study, “The predicted interactome of the human mitochondrial proteome,” published in Nature Communications, the research team utilized an AI tool to screen over 630,000 potential pairings among approximately 1,100 human mitochondrial proteins. The AI’s predictions identified nearly 2,900 likely interactions, linking many disease-associated proteins to previously unknown biological pathways. This breakthrough points to potential diagnostic and therapeutic targets. From a basic science perspective, the study also provided the first probable partners for 85 previously uncharacterized proteins, offering vital clues about their roles within cells. Abhinav Swaminathan, Ph.D., the study’s lead author and an AgriLife Research scientist, emphasized that without knowing the function of individual proteins, determining effective disease treatments remains challenging.
Repurposing an AI tool that predicts protein shape
Researchers globally have been working to identify the functions of these mysterious mitochondrial proteins, but traditional laboratory methods are inherently slow. To accelerate this process, Swaminathan and Gohil initiated a project nearly three years ago to test an artificial intelligence tool known as AlphaFold-Multimer (AFM). Google DeepMind originally developed and released AFM in 2022 to accurately predict the three-dimensional shapes that proteins form when they interact. Swaminathan explained that their lab lacked reliable experimental tools for evaluating direct mitochondrial protein interactions, especially for the mitochondrial copper pathway proteins they studied. A trial run using the AI approach proved highly successful, prompting them to apply it to all human mitochondrial proteins. Out of the 630,003 possible protein pairs, the team cataloged 2,895 likely interactions, with more than half of these being entirely new discoveries never before reported.
Confirming AI predictions in the lab
Following the AI predictions, about two years of intensive laboratory experiments were conducted. These experiments successfully confirmed several of the predicted protein partnerships in both yeast and human cells, with some validations performed in collaboration with researchers at Washington University School of Medicine in St. Louis. The findings demonstrated that AI predictions can serve as an invaluable roadmap, guiding scientists toward laboratory experiments most likely to yield significant results efficiently. To further aid the scientific community, the team has made its predictions publicly available as a free online resource called MitoMatch, enabling researchers worldwide to identify likely partners for the proteins they are studying. As the team continues to experiment with the vast dataset, they are also actively collaborating with other researchers to deepen the understanding of cell function and pave the way for new insights into human diseases. While cures for mitochondria-linked diseases may still be distant, Swaminathan considers the team’s recent success with their AI-driven approach a promising start. He noted that their lab, traditionally focused on biochemistry and genetics, achieved this success by creatively addressing the protein function problem with accessible AI resources, providing a solution for broader use. This groundbreaking study received support from the National Institutes of Health, Texas A&M High Performance Research Computing, and The Welch Foundation.
