Brucellosis Breakthrough: Bangladeshi Researchers Leverage Machine Learning to Combat Zoonotic Threat
A team of veterinary researchers from the Faculty of Veterinary Science at Bangladesh Agricultural University (BAU) has made a groundbreaking discovery in identifying, assessing, and preventing brucellosis—a bacterial zoonotic disease with global prevalence. For the first time in Bangladesh, the team has utilized machine learning techniques to extract crucial data, according to Professor Dr. Md. Siddiqur Rahman, the lead researcher from the Department of Medicine.
The research was conducted with the collaboration of PhD candidate Colonel (Retd.) S.M. Azizul Karim Hossaini, under the co-supervision of Dr. Heinrich Neubauer from the Friedrich Loeffler Institute in Germany. Technical support was provided by King Faisal University in Saudi Arabia.
The findings, focusing on the deadly disease that affects domestic animals, wildlife, and marine mammals, have been published in the Asian Journal of Agriculture and Biology, a Scopus-indexed journal with an impact factor of 1.6.
The study identified four of the twelve known Brucella species—B. abortus, B. suis, B. melitensis, and B. canis—as the most virulent. "Currently used live vaccines like B. abortus S-19 and RB51 carry potential side effects. However, our research proves that killed vaccines are safer and more effective," said Professor Siddiqur Rahman. He added that this technology holds promise for developing an innovative brucellosis vaccine in Bangladesh.
“There is no effective cure for this disease, which has caused immense distress among livestock farmers. Using five machine learning algorithms, we successfully identified the key risk factors for brucellosis. Among them, MLP, Deep Learning 4J, AdaBoostM1, and J48 Tree were the most effective,” the professor noted.
He further stated, “Machine learning algorithms are currently used worldwide to detect heart disease, kidney problems, diabetes, and respiratory illnesses in humans. However, in our country, due to financial limitations and religious sensitivities, culling infected animals is not feasible. Instead, animals that are highly valued or meant for export can be selectively treated.”
PhD researcher Colonel (Retd.) Azizul Karim Hossaini said, “Under the direct supervision of Professor Siddiqur Rahman, we successfully used machine learning to identify the transmission process and preventive strategies for the disease. Combined use of oxytetracycline, streptomycin, and benzylpenicillin yielded promising results in treatment.”
“We followed precise protocols and achieved the expected outcomes,” he added. “This success will remain a matter of pride for the nation.”
Brucellosis is a globally widespread zoonotic disease that poses serious health threats to domestic livestock, wildlife, and marine mammals. It causes significant economic losses for livestock farmers and is highly transmissible to humans. Infected animals often experience reduced milk production, abortions, and a dramatic decline in productivity.
