Machine learning-enhanced UV spectrophotometric platform for multicomponent analysis in biological matrices

Designing analytical methodologies that can accurately resolve highly overlapped multicomponent systems in biological matrices while maintaining environmental sustainability remains a major objective in contemporary bioanalytical chemistry. UV spectrophotometry is an attractive analytical tool because of its operational simplicity, rapid data acquisition, and low reagent consumption. Nevertheless, its direct quantitative application is frequently hindered by extensive spectral overlap among analytes and by interference arising from complex biological matrices. In this study, a computationally engineered dual-matrix bioanalytical framework is introduced for multicomponent determination in complex systems. The platform integrates machine learning-assisted chemometric modeling with statistically optimized experimental design to enhance the resolving power of UV spectrophotometry, enabling reliable analysis in human plasma as a representative biological matrix. Calibration datasets were constructed using factorial experimental design, while validation sets were generated via Sobol quasi-random sampling to ensure uniform multidimensional coverage and reduced collinearity. Spectral dimensionality reduction and adaptive wavelength selection were achieved using a Firefly Algorithm-optimized partial least squares (FA-PLS) model, improving predictive performance compared to conventional approaches. The developed platform demonstrated high predictive accuracy and robustness against matrix interference, achieving recoveries of 98.46-100.45% with RMSEP values of 0.035-0.520 μg/mL in human plasma. Applicability was further verified using an additional matrix, confirming generalizability for multicomponent analysis. Sustainability evaluation using Multicolor Assessment, carbon footprint, and the Need-Quality-Sustainability index demonstrated reduced solvent consumption and environmental impact. Overall, the proposed framework provides a versatile and sustainable bioanalytical platform for multianalyte determination in complex biological systems.