Near infrared spectroscopy with aquaphotomics as a novel approach was assessed for the diagnosis of soybean plants (Glycine max) infected with soybean mosaic virus (SMV) at latent symptomless stage of the disease. Near infrared (NIR) leaf spectra (in the range of 730-1025 nm) acquired from soybean plants with and without the inoculation of SMV were used. Leaf samples from all plants were assayed with enzyme-linked immunosorbent assay (ELISA) to confirm the infection. Previously reported NIR band for water at 970 nm and two new bands at 910 nm and 936 rim in the water specific region of NIR were found to be markedly sensitive to the SMV infection 2 weeks prior to the appearance of visual symptoms on infected leaves. The spectral calibration model soft independent modeling of class analogy (SIMCA), predicted the disease with 91.6% sensitivity and 95.8% specificity when the second order derivative of the individual plant averaged spectra were used. The study shows the potential of NIR spectroscopy with its novel approach to elucidate latent biochemical and biophysical information of an infection as it allowed successful discrimination of SMV infected plant from healthy at the early symptomless stage of the disease. (C) 2010 Elsevier Inc. All rights reserved.