Artificial Neural Network Prediction and Elicitation of Flavones and Rosmarinic Acid in Teucrium polium Calli by Methyl Jasmonate and Nano-Sized TiO₂

The present research study evaluated the effects of benzylaminopurine (BAP) in combination with naphthalene acetic acid (NAA), as well as methyl jasmonate and nano-sized TiO₂ elicitors on callus induction, pharmaceutically important flavones and rosmarinic acid in leaf-derived calli of Teucrium polium. The feed-forward Artificial Neural Network (ANN) model was also applied to predict the effects of hormones and elicitors on phenolic compounds production. The best callus induction was achieved with media fortified with 1.5 mg L^− 1 BAP plus 0.5 mg L^− 1 NAA, followed by 1.5 mg L^− 1 BAP. Biochemical features and rosmarinic acid have substantially raised by the maximum ones in calli elicited with methyl jasmonate (50 µM). The exception was phenolic acids possessed much greater content in 10 mgL^− 1 nano-sized TiO₂-treated calli. High antioxidant activities (IC₅₀ 2.066 ± 0.24 µg mL^− 1 and 383.79 ± 33.60 mg Fe^+ 2 100 g^− 1 DW), with approximately 2- and 1.5-fold increases compared to controls, were observed in hydromethanolic extracts of calli elicited by 50 µM methyl jasmonate on an MS medium containing 1.5 mg L^− 1 BAP plus 0.5 mg L^− 1 NAA. Likewise, the highest contents of anti-genotoxic flavones apigenin and luteolin were measured in callus pretreated with 50 µM methyl jasmonate on the media supplemented with 1 mg L^− 1 BAP (180.67 ± 2.52 and 201.11 ± 3.63 µg100 g^− 1 extract, respectively) and 1.5 mg L^− 1 BAP plus 0.5 mg L^− 1 NAA (206.01 ± 7 and 256.22 ± 2.1 µg 100 g^− 1 extract, respectively). The ANN with hyperbolic-tangent (10-1-3) activation function (r² 0.99, SSE 0.01, and RE 0.001) showed a strong correlation between the ANN model outputs and the experimental amounts of polyphenols, ortho-diphenols, and phenolic acids.