Researchers have uncovered thousands of preserved metabolic molecules inside fossilized bones millions of years old, offering a new window into prehistoric life. The findings, published January 3, 2026, by New York University researchers, reveal details about ancient animals' diets, diseases, and surrounding climates.

The team successfully examined metabolism-related molecules preserved inside fossilized bones from animals that lived between 1.3 and 3 million years ago. This approach offers a novel way to reconstruct ancient ecosystems, according to the researchers. One fossil even showed signs of a parasite still known today.

"These preserved molecules act like tiny time capsules, providing direct evidence of what these animals ate, the illnesses they suffered, and the conditions they endured," said Timothy Bromage, a lead researcher at NYU Dentistry. "It's like reading a diary written in the very bones of these creatures."

The analysis involved using advanced mass spectrometry techniques coupled with AI-powered algorithms to identify and interpret the complex mixture of molecules extracted from the fossils. The AI was crucial in differentiating between genuine metabolic signals and background noise or contamination, explained Bin Hu, another researcher involved in the study. This process is similar to how AI is used in medical diagnostics to detect subtle patterns indicative of disease.

The implications of this research extend beyond paleontology. By understanding the metabolic responses of ancient animals to environmental changes, scientists can gain insights into how modern species might adapt to current challenges like climate change and emerging diseases. The discovery of a parasite still present today, for example, highlights the long-term evolutionary relationships between hosts and pathogens.

The ability to extract and analyze these ancient molecules relies on advances in analytical chemistry and computational biology. The AI algorithms used in this study were trained on vast datasets of known metabolic compounds, allowing them to identify even trace amounts of these substances in the fossil samples. This approach is akin to how AI is used in drug discovery to identify potential therapeutic molecules.

The research team plans to expand their analysis to a wider range of fossils from different geological periods and geographic locations. They also aim to refine their AI algorithms to extract even more detailed information from the molecular data. This could potentially reveal insights into the genetic makeup of extinct species and the evolutionary history of life on Earth.