Multi-response optimization of spice-enriched butterfly pea beverages by ANN and chemometric approaches: linking color attributes, functional properties, and in vitro bioaccessibility
Quality Assurance and Safety of Crops and Foods, cilt.18, sa.3, ss.378-410, 2026 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 18 Sayı: 3
- Basım Tarihi: 2026
- Doi Numarası: 10.15586/qas.v18i3.1650
- Dergi Adı: Quality Assurance and Safety of Crops and Foods
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, ABI/INFORM, Compendex, Food Science & Technology Abstracts, Natural Science Collection (ProQuest), Engineering Source (EBSCO), Health Research Premium Collection (ProQuest)
- Sayfa Sayıları: ss.378-410
- Anahtar Kelimeler: artificial neural network (ANN), bioactive components, Clitoria ternatea L, in vitro bioaccessibility, principal component analysis (PCA)
- Açık Arşiv Koleksiyonu: AVESİS Açık Erişim Koleksiyonu
- Bursa Uludağ Üniversitesi Adresli: Evet
Özet
Clitoria ternatea L. flowers are a rich natural source of anthocyanins and phenolic compounds. This study comparatively evaluated 10 butterfly pea tea formulations containing mint, rosemary, fenugreek, cardamom, garlic, chili, and their combinations, and investigated the effects of in vitro gastrointestinal digestion on their physicochemical and functional properties. Significant formulation-dependent differences were observed in pH, °Brix, color parameters (L*, a*, b*, C*, and h° [hue angle]), total phenolic content (TPC: 756.30–5,299.33 mg GAE/100 g dry matter (dm)), total monomeric anthocyanin content (TMA: 1.39–4.45 mg/kg), and total antioxidant capacity (DPPH: 237.54–1,799.48 µmol Trolox/g dm and FRAP: 161.80–1,114.53 µmol Trolox/g dm). Multivariate analyses successfully differentiated the formulations, with the first two principal components explaining 65.9% of the total variance and revealing strong associations between phenolic recovery, antioxidant responses, and color attributes. A feed-forward artificial neural network (ANN) model demonstrated strong predictive performance, with a high overall correlation between experimental and predicted responses (R ≈ 0.97). ANN-based optimization identified the mint-enriched formulation (MI) as the best overall formulation (desirability score: 0.4955), whereas CA-PG exhibited the highest TPC after intestinal digestion (12,813.69 mg GAE/100 g dm).