A comparative review of red, green, and brown seaweed: Bioactive components, health effects, and machine learning...

Edible seaweed subcategorized into Chlorophyta (green algae), Phaeophyceae (brown algae) and Rhodophyta (red algae) are considered sustainable functional foods with marketing challenges and regulatory issues. Seaweed-associated safety concerns continue to garner attention. Underdefined similarities and differences among different categorized seaweed hinder their distinct developments in food innovation. This review comparatively analyzed Chlorophyta, Phaeophyceae and Rhodophyta, focusing on profiles of functional, bioactive, and undesirable substances. Mechanistic insights into different and unique physicochemical characteristics and biological functions were discussed for each type of seaweed. Chlorophyta, Phaeophyceae and Rhodophyta can be distinguished by distinct phenotypic traits, chemical components, physicochemical characteristics, sensory properties, and bioactive potencies. Rhodophyta are the richest in dietary fiber, protein, agar, vitamin (C and B6) and phycobiliprotein. Alginate, fucoidan, and laminarin, phlorotannin, fucoxanthin and iodine in Phaeophyceae, and Chlorophyta ulvan are well-recognized. Rhodophyta, Chlorophyta and Phaeophyceae exhibited antioxidative, antimicrobial, anti-inflammatory, neuroprotective, anticancer and antidiabetic properties. Phaeophyceae excels in metal chelation, free radical scavenging and neuroprotection. Heavy metal contamination of Phaeophyceae is concerning. Small sample size, non-representativeness in sampling, and non-standardized extraction and analytical methods introduce uncertainty into comparing findings across/within studies. Comparative data regarding macromolecular crowding, safety criteria, bioactive potencies range, large-scale clinical trials, and commercial viability remain insufficient. Promisingly, machine learning identifies/predicts distinct aspects, including morphologies, chemical compositions, physicochemical properties and bioactivities. Solving the challenges of limited data quantity and quality, insufficient model validation and poor model interpretability will promote machine learning applications. The findings of this review provide fundamental knowledge for developing functional foods with seaweed-derived ingredients.