AI-assisted experimental planning for two-stage cultivation to enhance photosynthetic pigment production in Dunaliella salina DSTA20

Abstract

Photosynthetic pigments derived from microalgae are increasingly valued for their diverse bioactivities. In this study, we improved carotenoid production in the halophilic microalga Dunaliella salina by combining two-stage cultivation with Artificial Intelligence (AI)-assisted experimental planning. ChatGPT-4 supported experimental planning, including Response Surface Methodology (RSM) design comparison and factor-range selection. Culture conditions were optimized using Central Composite Design-based RSM, which predicted 24.24°C and 56.36 ppt as the optimal temperature and salinity conditions. Cultures grown under these conditions in Stage I (from lag to early stationary phase) were subsequently subjected to nine temperature–salinity stress combinations (18.00, 24.20, or 30.00°C × 30.00, 56.36, or 80.00 ppt) for 3 days in Stage II (late stationary phase). Pigment profiling showed marked condition-dependent differences in chlorophyll a and b and the carotenoids including lutein, violaxanthin, β-carotene, and zeaxanthin. Exposure to low-temperature and high-salinity stress (18.00°C, 80.00 ppt) produced a maximal total pigment content representing a 2.36-fold increase compared to the minimum conditions (low-temperature and low salinity, 18.00°C, 30.00 ppt). These findings demonstrate that simultaneous temperature and salinity stress within a two-stage cultivation regime enhances pigment accumulation in D. salina and suggest that AI-assisted experimental planning may support efficient optimization strategies in algal biotechnology.