Bee-based sensor using apis mellifera for detection of Andrographis paniculata volatile compounds

Honey bees can be trained to be sniffer bees due to the presence of high number of odorant receptors (170 odorant receptors) in their heads. This unique characteristic enables them to detect scent down to part per trillion level. In this study, localized honey bees (Apis mellifera) were trained by using the method of classical Pavlovian conditioning. The proboscis extension reflex (PER) of sniffer bees towards the target odor was observed and recorded. The phytochemical profile of Andrographis paniculata (A. paniculata) was constructed using headspace solid-phase microextraction coupled with gas chromatography integrated with mass spectrometer method. The volatile marker compounds were identified. The sniffing capacity of the sniffer bees was determined by varying the heating temperature from 50-120 °C, the weight of plant material from 20-100 mg and the percentage (20-100%) of the target herbal sample in the mixture of A. paniculata and Clinacanthus nutans (C. nutans). C. nutans is an herbal plant which is morphologically similar to A. paniculata and it also belongs to the Acanthaceae family. The efficiency, accuracy and sensitivity of sniffer bees were analyzed and validated statistically. The success rate of sniffer bees for heating temperatures was approximately 90 %. The success rate for minimum weight of plant sample, 20 mg was 50 %. The success rate percentage of target herbal sample increased when the percentage of A. paniculata was proportionally increased. Compounds such as caryophyllene, ß-elemene, 3,3-dimethylhexane, apiol, 6,10,14-trimethyl-2-pentadecanone and dihydroactinidiolide were detected in the gaseous mixture. The kinetics of volatile marker compounds released from the plant samples were studied to predict the concentrations of the volatile marker compounds for sniffer bee detection at 85 ?. Second-order and two-site kinetic models were selected because of the kinetic data of these volatile marker compounds fitted well to these models (R2 > 0.9)