Development of an in Vitro Organ-Specific Regeneration System for Scrophularia Ningpoensis and Predictive Modeling of Culture Conditions Using Machine Learning

Scrophularia ningpoensis relies mainly on vegetative propagation, a practice that can accelerate germplasm degradation and pathogen accumulation, thereby limiting the large-scale production of healthy seedlings. Here, we established an in vitro regeneration system using leaf and petiole explants cultured on Murashige and Skoog (MS) medium supplemented with different combinations of 6-benzylaminopurine (6-BA) and naphthaleneacetic acid (NAA), and applied machine learning (ML) to model the regeneration responses within the tested experimental space. Regeneration showed clear organ-specific patterns: Leaf explants exhibited strong rooting capacity, reaching 96.7% +/- 3.3% at 1.0 mg/L 6-BA + 1.0 mg/L NAA, whereas petiole explants showed higher shoot regeneration, with a maximum of 50.0% +/- 11.6% at 0.5 mg/L 6-BA + 0.2 mg/L NAA. Factorial analysis of variance and ordinary least squares regression identified explant type and NAA concentration as the major determinants of regeneration. Random forest models showed moderate predictive performance for shoot regeneration (R2 = 0.57) and stronger predictive performance for rooting (R2 = 0.78), with explant type consistently ranked as the most influential variable. Notably, the highest-ranked rooting condition predicted by the model was the same as the best-performing treatment identified experimentally. Together, these findings establish an efficient regeneration system for S. ningpoensis, demonstrate explant-specific responses to phytohormones, and support ML as a complementary tool for quantitative interpretation and predictive analysis of tissue culture responses in nonmodel medicinal plants.