Effect of a new probiotic on Artemia cysts determined by a convolutional neural network

Abstract

One of the problems in sea farming is infections that cause mass mortality of crustaceans. To fight infections and improve sanitary conditions, farmers are actively using probiotic preparations. We aimed to study the effect of a new probiotic based on Bacillus toyonensis B-13249 and Bacillus pumilus B-13250 strains on the incubation of Artemia franciscana cysts. Another purpose was to test a possibility of using a convolutional neural network for fast automatic counting of cysts, nauplii, and embryos. A pilot batch of the probiotic was prepared at the Prombiotech Engineering Center, Altai State University, from two strains of spore bacteria from the Center’s collection: B. toyonensis B-13249 and B. pumilus B-13250. The recommended amount of the probiotic was experimentally determined as 0.1 per 2 g of cysts. This concentration increased the number of hatched cysts by 1.4 and 10% in the batches from Lake Bolshoye Yarovoye (Z29.04) and from Lake Kuchuk (C9). It also increased the biomass yield to 7.40 ± 0.69 and 6.80 ± 0.43 g in these two batches, respectively, compared to the control samples where the yields were 5.30 ± 0.60 and 4.60 ± 0.50 g, respectively. The robot counter reduced the sample processing time 15 times and saved the data for further use. The probiotic based on B. toyonensis B-13249 and B. pumilus B-13250 had a positive effect on the hatching rate and biomass yield of A. franciscana. The new method for rapid counting of Artemia, which was based on the convolutional neural network and developed as an application of the Artemeter-1 robot, reduced the processing time and lowered labor costs.