Non-Invasive Label-Free Optical Biosensor for Accurate Peptide Detection Using Refractive Index Measurement and Machine Learning

Peptide diagnostics serve an important role for initial disease recognition, pharmaceutical evaluation, and environmental monitoring. Conventional methods for diagnosis typically involve labelling concepts that reduce sensitivity, increase test complexities, and limitations in real-time analysis. In the proposed work, we have introduced Corner-Triangle, Floral Geometry Refractive Index Biosensor (CTFGRIB) for monitoring peptide concentrations by names, Glycylleucine (Gly-Leu), Triglycine (Tri), Glycine (Gly), Glycytyrosine (Gly-Tyr), Diglycine (Dig), and Glycylaspartate (Gly-Asp) with a combination of machine learning evaluation. A periodical arrangement of corner-triangle patterns surrounded by a floral layout, as a distinctive geometry, provides a number of synergistic benefits that directly boost biosensing capabilities. The parametric assessments involve outstanding performance parameters with the favourable values of sensitivity being 1023.25 nm/RIU, and favourable values of detection limit are 0.0733 RIU for the Gly-Leu peptide cell. The favourable quality factor value of 24.0368, and the figure of merit value of 10.9508 RIU-1 have been achieved for the Gly-Leu peptide cell. The favourable transmittance rate of 33.6%, 33.3%, 33.0%, 33.0%, 32.9%, and 32.9% have been observed for Gly-Leu, Tri, Gly, Gly Tyr, Dig, and Gly-Asp, respectively. The optimised R squared value of 0.997604 and the MSE value of 9.607930 × 10-05 have been achieved from the machine learning method.