Rootstock-mediated salinity resilience in cucumber (Cucumis sativus L.): integrating physiological traits, genomic...

Salt stress is a major abiotic constraint in cucumber (Cucumis sativus L.), reducing biomass, photosynthesis, and genomic stability. Grafting onto salt-tolerant Cucurbita rootstocks is a promising strategy to enhance plant resilience. Recently, machine learning (ML) has provided new opportunities to capture complex trait interactions and identify key predictors of stress tolerance.

We evaluated two cucumber cultivars (Cagla F1, Minimix F1) grafted onto four interspecific Cucurbita maxima × Cucurbita moschata rootstocks (TZ148, Devrim, Cremna, Kublai) under 0 vs. 100 mM NaCl for 30 days in a soilless fertigation system. Morphological, physiological, and molecular traits were evaluated, including biomass accumulation, chlorophyll content (SPAD) and incident photosynthetically active radiation (PAR), and genomic template stability (GTS) using ISSR markers. Salt stress reduced growth and biomass (leaf FW - 56%, root DW - 74%) and lowered SPAD and relative water content (RWC); grafting-especially with TZ148 (and to a lesser extent Kublai)-mitigated these losses by maintaining chlorophyll content (SPAD) and biomass under salinity. Grafted combinations, especially TZ148/Cagla, maintained higher stability (GTS: 88%, GC: 0.07), confirming the protective role of grafting. ML approaches, including Principal Component Analysis (PCA) and Random Forest (RF), clearly separated control vs. salinity and, while grafting types showed only partial separation, RF consistently ranked root/stem fresh weight, SPAD, leaf area, and fruit weight as top predictors.

Grafting significantly improved cucumber tolerance to salinity by sustaining biomass, photosynthetic capacity proxies (SPAD), and genomic integrity. ML-based analyses added predictive power and biological interpretation, confirming grafting with appropriate rootstocks as a sustainable strategy for cucumber production in saline nutrient solution conditions.