Medical researchers and scientists are increasingly turning their attention to metabolites — small molecules produced within living cells through processes like the digestion of food, chemicals, or drugs. This field of study, called metabolomics, helps scientists identify disease markers, assess the effectiveness of treatments, and understand how diet and nutrition influence health. One of the main tools used to detect these molecules is mass spectrometry, a technique that measures the weight and electrical charge of tiny particles. However, despite the vast amount of data collected, most of these particles remain unidentified. These unknown molecules are referred to as the "dark metabolome," a term highlighting the lack of understanding about their structures and roles in the body. A new tool called the Molecular Community Network (MCN), developed by Professor Vladimir Boginski and a team of researchers, is helping to shed light on this mysterious "dark matter" of metabolites. Their findings were recently published in Cell Reports Methods. According to Boginski, public databases contain approximately 8.4 million mass spectra — records of molecular weights and charges — from both known and unknown molecules. He estimates that up to 90% of these are part of the dark metabolome. "It's crucial to develop methods that can systematically explore this vast molecular space and potentially discover new molecules," he explains. Traditional methods of analyzing molecular data, known as molecular networking, have limitations. These methods group molecules based on how similar their mass spectra are, but they often fail to connect biologically related molecules that fall below a certain similarity threshold. This can fragment molecular families and limit the potential for new discoveries. The MCN, however, uses an algorithm that divides the molecular network into natural groups, or communities, where molecules are tightly connected. This approach preserves strong links within groups and keeps the communities intact. "This allows us to link almost every molecule in the network to at least one neighbor, typically from a similar molecular family," Boginski says. This method supports annotation propagation, where the identity of unknown molecules can be predicted based on their connections to known ones. As a result, nearly 95% of molecules are now connected and assigned to network communities, greatly expanding the potential for molecular discovery. Using the MCN, Boginski and his team have already identified a new class of bile acids, which are produced by gut microbes. One of these bile acids was found to be present only in young infants. The discovery was first predicted by the MCN and later confirmed in the laboratory. Boginski believes this is just the beginning. "Biomarker discovery often hits a roadblock when a molecule that differentiates sick from healthy individuals is detected but remains unidentified," he says. "MCN can help turn these roadblocks into discoveries. Because this method works with existing mass spectrum data, about 8.4 million molecules in public databases can now be reanalyzed, significantly increasing the potential for new biomarker discoveries."