Nicholas Jacob, Diones Almeida, Tiago Fernandes, Anders Skaarud, Alejandro Tola Alvarez, Rajesh Joshi
Aquaculture (2026), https://doi.org/10.1016/ j.aquaculture.2026.744117
Abstract
Selective breeding has transformed tilapia aquaculture, yet comparisons of strain performance across production systems remain challenging due to differences in system type, growth conditions, and evaluation metrics.
The primary objective of this study was to compare the performance of two Nile tilapia strains under commercial pond and cage production systems in Brazil using a combination of conventional performance metrics and newly established key performance indicators (KPIs). A common garden experiment was conducted to evaluate the GenoMar 1000 strain against a widely used local commercial strain.
Performance was assessed using standard KPIs, including growth, survival, and uniformity. We developed new composite KPIs that integrate growth rate, stocking density, survival, and feed conversion ratio to enable comparisons across production systems. In addition, image-based phenotyping using YOLOv8 keypoint detection was applied to estimate body weight and assess morphological differences as a complementary highthroughput phenotyping tool.
GenoMar 1000 consistently outperformed the local strain, exhibiting higher growth (+29%), improved survival (+6%), and greater uniformity (24% lower coefficient of variation) at harvest. These differences were reflected in the newly established KPI values, which were 33-42% higher for GenoMar 1000 across production systems. Image-derived body weight estimates closely matched measured weights, supporting the utility of automated phenotyping in large-scale trials.
Economic calculations based on these results showed that GenoMar 1000 delivered 24% and 35% higher annual profits in ponds and cages respectively, despite a higher fingerling purchase cost. Overall, this study demonstrates clear performance differences between strains across production environments, while highlighting the value of integrated KPIs and image-based phenotyping as supporting tools for strain comparison.
