Comparative phyto-biochemical profiling, antibacterial efficacy, and culturable rhizosphere bacterial dynamics of turmeric (Curcuma longa L.) variants from eastern India

Different turmeric variants possess distinct medicinal and ecological attributes, yet a comprehensive comparative analysis of their phyto-biochemical profiles and associated culturable rhizobacterial profiles is lacking. To resolve this, the current study evaluated yellow, black, and white turmeric variants, revealing distinct phyto-biochemical constituent profiles characterized by peak phenolics in white turmeric (121.63 mg GAE/g), maximum flavonoids (49.85 mg QE/g) and proteins (9.65%) in black turmeric, and primary carbohydrate (63.52%) and carotenoid (35.44 µg/g) reserves in yellow turmeric. Multivariate analysis established marked negative correlations between protein and carotenoid content ( r = − 0.991), and total phenolics and carbohydrates ( r = − 0.970), with a two-component PCA model accounting for 100% total variance (PC1 = 59.42%, PC2 = 40.58%). Antibacterial evaluation against Staphylococcus aureus and Escherichia coli demonstrated the highest inhibitory activity in the ethanolic extract of yellow turmeric compared to the other tested variants; however, this activity remained significantly lower than that of ampicillin. In contrast, white turmeric extracts exhibited no antibacterial activity. Phylogenetic model selection based on the lowest Bayesian Information Criterion (BIC) designated the Jukes–Cantor model with invariant sites (JC + I; BIC = 5,279.42) as the best fit for Bacillus sp. AKU 5 (putative Bacillus thuringiensis affinity), whereas the Tamura 3-parameter model (T92; BIC = 9,815.55) was optimal for Bacillus sp. AM 1 (putative B. cereus affinity). Maximum likelihood analyses revealed contrasting evolutionary patterns, with B. thuringiensis exhibiting a highly constrained substitution rate (α = 0.04; Log Likelihood = − 2006.341) and pronounced C→A substitution bias, whereas B. cereus showed a more uniform rate distribution (α = 30.23; Log Likelihood = − 4234.930) characterized by symmetric A↔G and T↔C transition biases, emphasizing preliminary nucleotide substitution variations within the turmeric rhizosphere. In-silico thermodynamic profiling of 16S rRNA amplicon secondary structures revealed distinct computational folding stabilities between B. thuringiensis AKU 5 (minimum free energy (MFE) = − 511.90 kcal/mol) and B. cereus AM 1 (MFE = − 506.50 kcal/mol). These baseline structural disparities reflect inherent differences in sequence length and GC-rich base-pairing density rather than in vivo physiological adaptations within the turmeric rhizosphere. Collectively, these findings establish an in-silico sequence modeling rather than verified genome-wide host-driven evolution, providing insights into preliminary turmeric chemotype-culturable rhizobacterial associations. Future research should focus on high-throughput metagenomic profiling to unravel whole-community dynamics, isolation of specific bioactive compounds, and exploiting target rhizosphere strains for agricultural biofertilizer or therapeutic formulations.