Capability of urine analysis for the disease blastomycosis using infrared spectroscopy with eigenvector signal to noise assessment

Blastomyces dermatitidis is the etiological agent of the relatively common disease blastomycosis found in certain regions of the United States. In order to discern the presence of blastomycosis, microorganisms present in biofluids of infected individuals must be incubated and cultured and then identified by microscopic observation. This procedure is time consuming due to difficult incubation conditions and necessitates special safeguards. An enzyme-linked immunosorbent assay(ELISA) was developed in an attempt to quicken determination. The immunological assay, however, also requires incubations as well as special reagents. This paper explores the potential of using infrared spectroscopy for determination of blastomycosis in urine samples from infected dogs. The regression methods of principal component regression (PCR) and partial least squares (PLS) are used to develop models. Recently introduced regression diagnostics in conjunction with new diagnostic tools described in this paper are used to determine final models. One of the new tools allows for an estimate of the signal (information) to noise ratio, i.e., reliability, for each eigenvector's contribution to the model. Based on diagnostic information, it is concluded that PCR and PLS are generating the same models.