Researchers at the University of Georgia (UGA) have developed a method combining microbiome data with artificial intelligence (AI) to determine the geographic origin of garlic, potentially offering a cost-effective approach to detecting food fraud.

The research, led by Xiangyu Deng, Ph.D., UGA Center for Food Safety Professor, began with an examination of the microbiota present on the surface of garlic bulbs. The researchers found that garlic carries a distinct bacterial signature reflective of the soil in which it was grown, enabling geographic identification based on microbial composition.

Garlic Fraud and Economic Drivers

Garlic has been identified as a target for food fraud due to trade protections and price differentials. The U.S. imposes a 376 percent tariff on garlic imported from China, incentivizing some suppliers to engage in trans-shipment, a practice in which products are routed through third countries and mislabeled to avoid tariffs.

Globally, food fraud has been estimated to result in annual economic losses of up to $15 billion.

Microbiome and Machine Learning Approach

The UGA researchers analyzed bacteria present on garlic by washing samples in a sterile solution and sequencing the microbial DNA. These microbial profiles were then input into a machine learning model trained to associate specific bacterial communities with geographic regions.

According to the researchers, the model correctly identified the origin of garlic samples approximately 90 percent of the time.

Because garlic grows underground and is not typically washed prior to distribution, it retains microbial characteristics of the surrounding soil, a concept commonly referred to as terroir.

Implications for Food Fraud Detection

The researchers noted that altering the microbial composition of garlic without compromising product quality would be difficult, making microbiome-based authentication a potentially robust tool against fraud.

The method was described as faster and more cost-effective than traditional analytical techniques and may support enforcement efforts by agencies such as U.S. Customs and Border Protection in verifying product origin claims.

Existing methods for determining garlic origin typically rely on detecting chemical or elemental markers, which require specialized equipment and extensive sample preparation.

Ongoing Research

Dr. Deng said the research team plans to continue exploring applications of microbiome analysis in food authentication.

He noted that, compared to established tools such as whole genome sequencing (WGS), practical applications for food microbiome analysis remain less defined, and additional work is needed to identify scalable use cases.