Using VIS-NIR spectroscopy and multi-omics analysis to compare mango anthracnose under natural and inoculated conditions.

Current studies on the detection and analysis of anthracnose in mangoes using optical technology mostly rely on inoculation methods. However, to what extent the inoculation (InI) can represent the biological and metabolic differences of the naturally infected (NaI) diseases remains unknown. Therefore, this study systematically compared microbial community composition, metabolite profiles, and visible near-infrared (VIS-NIR) spectral characteristics to evaluate whether InI can serve as a reliable substitute for NaI in laboratory research. The results revealed distinct microbial and metabolic differences between the two infection modes. In the InI group, Colletotrichum-xanthorrhoeae dominated (99.6 %), whereas the NaI group exhibited a more diverse microbial composition, with Colletotrichum-xanthorrhoeae (66.7 %) coexisting with Botryosphaeria agaves (32.9 %). Metabolomic analysis identified 255 differential metabolites, with only three shared among the top 20 most significant ones, indicating substantial biochemical variations between infection types. Spectral analysis in the 400-1000 nm range demonstrated that the effective wavelength regions differed between InI and NaI in the early stages, with In-I-early at 786-798 nm and Na-I-early at 631-637 nm. Spectral reflectance differences between the two infection modes may stem from variations in metabolite composition and pigment accumulation, affecting optical absorption and scattering, especially in the unique spectral features with phenolic compounds, flavonoids, and organic acids of NaI. In addition, the Partial Least Squares Discriminate Analysis (PLS-DA) model was used to discriminate two types of diseased mangoes. The detection accuracy rate for the early-stage of InI is as high as 100.00 %, while the early stage of NaI is 89.92 %. In conclusion, the findings indicate that inoculation may not fully replicate the physiological and biochemical complexity of natural infection, emphasizing the need to consider natural disease models when developing non-destructive optical detection techniques for anthracnose in mangoes.