AI-Powered Tool DECIPHAER Unveils Molecular Mechanisms of Tuberculosis Drug Action

AI-powered analysis reveals how drugs kill tuberculosis at the molecular level

Tuberculosis (TB) is the world's deadliest infectious disease -- and one of the hardest to cure. Standard treatment requires a cocktail of multiple drugs over at least six months, and one in five patients have a type of TB that resists these first-line medications. Now, a new study offers a powerful AI-assisted method for uncovering exactly how TB drugs kill the bacteria, opening the door to smarter treatment combinations that could work faster. Developing a more effective -- and shorter -- TB treatment is a global priority. "We need a better multidrug regimen: three to five new drugs that work even for what is currently drug-resistant TB," says Bree Aldridge, senior author of the study and a professor in molecular biology and microbiology at Tufts University School of Medicine and professor in biomedical engineering at Tufts University School of Engineering. But progress has been slow, in part, because scientists lacked tools to see precisely how drugs work and therefore how they could best work together to attack TB bacteria. "TB likely has multiple Achilles' heels that we could hit all at once," explains Aldridge, who also is associate director of the Stuart B. Levy Center for Integrated Management of Antimicrobial Resistance at Tufts. "But it's surprisingly difficult to figure out exactly how a drug kills its target cell." It's like walking into a room and spying bruised faces, an overturned chair, and a shattered lamp; you can tell that a fight happened but not who started it or how it unfolded. In the same way, scientists can tell when a drug has killed target cells, but often not the exact chain of molecular events, aka, "mechanism of death." Aldridge and her collaborators from Tufts University School of Medicine and other institutions have now found a way to understand that mechanism. In a new study in Cell Systems, they demonstrated how their new AI-assisted tool -- called DECIPHAER (decoding cross-modal information of pharmacologies via autoencoders) -- can reveal, in molecular detail, how potential TB drugs kill the bacteria. The tool builds on the team's earlier research that captured high-resolution images of TB bacteria as they die during treatment. These snapshots reveal clues -- for example, changes in the bacterial cells' shape or internal structure -- caused by a drug's mode of attack. Scientists use this "morphological profiling" as a kind of crime scene investigation for cells: They dose TB bacteria with a new drug, freeze them at the moment of death, and compare the resulting cellular damage with patterns seen from known antibiotics. "If you treat TB bacteria with a new drug and it goes splat in the same way it does for other drugs that destroy the cell wall, then you may assume it destroys the cell wall as well," says Aldridge. Using AI, the team has now gone a step further, linking these visual clues to detailed readouts of bacterial gene activity, known as transcriptional profiles. The researchers trained a model to spot which molecular changes, such as bacterial genes switching on or off, occur alongside specific visual changes. "Before, we could only say roughly how a drug killed TB using morphological profiling. Now we can bring more exact insights into how drugs are impacting the cells and why the bacteria are dying," says Aldridge. For example, in testing DECIPHAER, she says the team found that a TB drug in clinical development didn't work as expected. "Based on similar existing compounds, we had assumed the drug worked by destroying the cell wall," she says. "But it actually kills TB bacteria by impairing the respiratory chain and cells' ability to make energy." Because the AI tool can predict a drug's molecular impact from images alone -- which is far cheaper than using RNA sequencing -- it can faster reveal how potential TB treatments work in different growth conditions, genetic strains, or drug combinations. "We plan to keep using it in our own lab's drug combination studies and hope it will support collaborations worldwide to accelerate the development of new TB drugs," says Aldridge. While the need is especially urgent for TB, she adds that DECIPHAER's approach also could be applied to other infectious diseases and cancer.

AI tool reveals how TB drugs kill bacteria at the molecular level

Tufts UniversityAug 25 2025 Tuberculosis (TB) is the world's deadliest infectious disease -- and one of the hardest to cure. Standard treatment requires a cocktail of multiple drugs over at least six months, and one in five patients have a type of TB that resists these first-line medications. Now, a new study offers a powerful AI-assisted method for uncovering exactly how TB drugs kill the bacteria, opening the door to smarter treatment combinations that could work faster. Developing a more effective -- and shorter -- TB treatment is a global priority. "We need a better multidrug regimen: three to five new drugs that work even for what is currently drug-resistant TB," says Bree Aldridge, senior author of the study and a professor in molecular biology and microbiology at Tufts University School of Medicine and professor in biomedical engineering at Tufts University School of Engineering. But progress has been slow, in part, because scientists lacked tools to see precisely how drugs work and therefore how they could best work together to attack TB bacteria. "TB likely has multiple Achilles' heels that we could hit all at once," explains Aldridge, who also is associate director of the Stuart B. Levy Center for Integrated Management of Antimicrobial Resistance at Tufts. "But it's surprisingly difficult to figure out exactly how a drug kills its target cell." It's like walking into a room and spying bruised faces, an overturned chair, and a shattered lamp; you can tell that a fight happened but not who started it or how it unfolded. In the same way, scientists can tell when a drug has killed target cells but often not the exact chain of molecular events, aka, "mechanism of death." Aldridge and her collaborators from Tufts University School of Medicine and other institutions now have found a way to understand that mechanism. In a new study in Cell Systems, they demonstrated how their new AI-assisted tool -- called DECIPHAER (decoding cross-modal information of pharmacologies via autoencoders) -- can reveal, in molecular detail, how potential TB drugs kill the bacteria. The tool builds on the team's earlier research that captured high-resolution images of TB bacteria as they die during treatment. These snapshots reveal clues -- for example, changes in the bacterial cells' shape or internal structure -- caused by a drug's mode of attack. Scientists use this "morphological profiling" as a kind of crime scene investigation for cells: They dose TB bacteria with a new drug, freeze them at the moment of death, and compare the resulting cellular damage with patterns seen from known antibiotics. If you treat TB bacteria with a new drug and it goes splat in the same way it does for other drugs that destroy the cell wall, then you may assume it destroys the cell wall as well." Bree Aldridge, senior author of the study Using AI, the team now has gone a step further, linking these visual clues to detailed readouts of bacterial gene activity, known as transcriptional profiles. The researchers trained a model to spot which molecular changes, such as bacterial genes switching on or off, occur alongside specific visual changes. "Before, we could only say roughly how a drug killed TB using morphological profiling. Now we can bring more exact insights into how drugs are impacting the cells and why the bacteria are dying," says Aldridge. For example, in testing DECIPHAER, she says the team found that a TB drug in clinical development didn't work as expected. "Based on similar existing compounds, we had assumed the drug worked by destroying the cell wall," she says. "But it actually kills TB bacteria by impairing the respiratory chain and cells' ability to make energy." Because the AI tool can predict a drug's molecular impact from images alone -- which is far cheaper than using RNA sequencing -- it can faster reveal how potential TB treatments work in different growth conditions, genetic strains, or drug combinations. "We plan to keep using it in our own lab's drug combination studies and hope it will support collaborations worldwide to accelerate development of new TB drugs," says Aldridge. While the need is especially urgent for TB, she adds that DECIPHAER's approach also could be applied to other infectious diseases and cancer. William C. Johnson, a Ph.D. student in molecular microbiology at Tufts Graduate School of Biomedical Sciences, is the first author. Research reported in this article was supported in part by the Gates Foundation and by the National Institutes of Health under award number T32AI007422. Complete information on authors, funders, methodology, limitations, and conflicts of interest is available in the published paper. The content is solely the responsibility of the authors and does not necessarily represent the official views of the funders. Tufts University Journal reference: Johnson, W. C., et al. (2025). Integration of multi-modal measurements identifies critical mechanisms of tuberculosis drug action. Cell Systems. doi.org/10.1016/j.cels.2025.101348

New AI Tool Reveals How Drugs Kill Tuberculosis | Newswise

Newswise -- Tuberculosis (TB) is the world's deadliest infectious disease -- and one of the hardest to cure. Standard treatment requires a cocktail of multiple drugs over at least six months, and one in five patients have a type of TB that resists these first-line medications. Now, a new study offers a powerful AI-assisted method for uncovering exactly how TB drugs kill the bacteria, opening the door to smarter treatment combinations that could work faster. Developing a more effective -- and shorter -- TB treatment is a global priority. "We need a better multidrug regimen: three to five new drugs that work even for what is currently drug-resistant TB," says Bree Aldridge, senior author of the study and a professor in molecular biology and microbiology at Tufts University School of Medicine and professor in biomedical engineering at Tufts University School of Engineering. But progress has been slow, in part, because scientists lacked tools to see precisely how drugs work and therefore how they could best work together to attack TB bacteria. "TB likely has multiple Achilles' heels that we could hit all at once," explains Aldridge, who also is associate director of the Stuart B. Levy Center for Integrated Management of Antimicrobial Resistance at Tufts. "But it's surprisingly difficult to figure out exactly how a drug kills its target cell." It's like walking into a room and spying bruised faces, an overturned chair, and a shattered lamp; you can tell that a fight happened but not who started it or how it unfolded. In the same way, scientists can tell when a drug has killed target cells but often not the exact chain of molecular events, aka, "mechanism of death." Aldridge and her collaborators from Tufts University School of Medicine and other institutions now have found a way to understand that mechanism. In a new study in Cell Systems, they demonstrated how their new AI-assisted tool -- called DECIPHAER (decoding cross-modal information of pharmacologies via autoencoders) -- can reveal, in molecular detail, how potential TB drugs kill the bacteria. The tool builds on the team's earlier research that captured high-resolution images of TB bacteria as they die during treatment. These snapshots reveal clues -- for example, changes in the bacterial cells' shape or internal structure -- caused by a drug's mode of attack. Scientists use this "morphological profiling" as a kind of crime scene investigation for cells: They dose TB bacteria with a new drug, freeze them at the moment of death, and compare the resulting cellular damage with patterns seen from known antibiotics. "If you treat TB bacteria with a new drug and it goes splat in the same way it does for other drugs that destroy the cell wall, then you may assume it destroys the cell wall as well," says Aldridge. Using AI, the team now has gone a step further, linking these visual clues to detailed readouts of bacterial gene activity, known as transcriptional profiles. The researchers trained a model to spot which molecular changes, such as bacterial genes switching on or off, occur alongside specific visual changes. "Before, we could only say roughly how a drug killed TB using morphological profiling. Now we can bring more exact insights into how drugs are impacting the cells and why the bacteria are dying," says Aldridge. For example, in testing DECIPHAER, she says the team found that a TB drug in clinical development didn't work as expected. "Based on similar existing compounds, we had assumed the drug worked by destroying the cell wall," she says. "But it actually kills TB bacteria by impairing the respiratory chain and cells' ability to make energy." Because the AI tool can predict a drug's molecular impact from images alone -- which is far cheaper than using RNA sequencing -- it can faster reveal how potential TB treatments work in different growth conditions, genetic strains, or drug combinations. "We plan to keep using it in our own lab's drug combination studies and hope it will support collaborations worldwide to accelerate development of new TB drugs," says Aldridge. While the need is especially urgent for TB, she adds that DECIPHAER's approach also could be applied to other infectious diseases and cancer. William C. Johnson, a Ph.D. student in molecular microbiology at Tufts Graduate School of Biomedical Sciences, is the first author. Research reported in this article was supported in part by the Gates Foundation and by the National Institutes of Health under award number T32AI007422. Complete information on authors, funders, methodology, limitations, and conflicts of interest is available in the published paper. The content is solely the responsibility of the authors and does not necessarily represent the official views of the funders.