When light colour matters: Spectral quality is associated with distinct small RNA candidates in Arabidopsis thaliana

The spectral composition of incident light is perceived by plant photoreceptors and can rapidly reshape downstream gene expression programs; however, its impact on the small-RNA layer beyond annotated miRNAs remains incompletely characterized. The objective of this study was to determine whether 3 h exposure of Arabidopsis thaliana rosettes to blue, green, red, or white light at an equal photosynthetic photon flux density (PPFD) of 400 μmol m^-2 s^-1 is associated with distinct profiles of candidate small RNAs. A stringent discovery and post-processing workflow was applied to identify high-confidence, treatment-associated small RNA candidates beyond annotated miRNA. The miRkwood-based pipeline, combined with additional filtering and contextual annotation, recovered a compact set of candidates dominated by the blue-light treatment (41 candidates), compared with fewer candidates associated with red (8), green (6), and white light (3). Genomic-context analysis indicated that most candidate sites were intergenic, with candidates detected under green light being entirely intergenic, and overlap with transposon annotations was used to distinguish candidates arising from transposon-rich genomic space. Sequence-feature profiling revealed pronounced treatment-dependent terminal nucleotide biases at both 5' and 3' ends, consistent with spectrum-associated shifts in length-class and terminal-nucleotide signatures that are informative for AGO-loading hypotheses. Target prediction highlighted a subset of genes showing convergent targeting by multiple independent blue-associated candidates, and duplex summaries showed structured, plant-like complementarity patterns (including frequent 10-11 pairing). Together, these results indicate that 3 h exposure to wavelength-defined light at 400 μmol m^-2 s^-1, particularly blue light, is associated with a distinct profile of detectable candidate small RNAs in Arabidopsis leaves and identify candidate interactions for follow-up validation.