Spermatozoa were isolated from the seminal fluid using washing (wash), post-capacitation (POS), and swim-up (SU) techniques, followed by analysis through Surface-Enhanced Raman Spectroscopy (SERS). Density gradient and swim-up methods were applied to 35 semen samples to validate sperm quality. The resulting spectra showed notable variations at 408 cm⁻¹ (S–S stretch attributed to lysozyme) and 728 cm⁻¹ (associated with DNA alterations and methylation). These spectral markers were incorporated into partial least squares discriminant analysis (PLS-DA) models to distinguish among sperm populations prepared by different methods. One PLS-DA model differentiated wash from POS and SU, attaining 86% sensitivity and 91% accuracy. Another model distinguished between POS and SU, achieving 77% sensitivity and 74% accuracy. The combined use of SERS and multivariate analysis offers a promising alternative for assessing sperm quality, supported by motility assessments in 35 validated samples. This approach could enhance both the accuracy and efficiency of reproductive diagnostics.

Spermatozoa were isolated from seminal fluid using washing (wash), post-capacitation (POS), and swim-up (SU) techniques, followed by analysis through Surface-Enhanced Raman Spectroscopy (SERS). Density gradient and swim-up methods were applied to 35 semen samples to validate sperm quality. The resulting spectra showed notable variations at 408 cm⁻¹ (S–S stretch attributed to lysozyme) and 728 cm⁻¹ (associated with DNA al-terations and methylation). These spectral markers were incorporated into partial least squares discriminant analysis (PLS-DA) models to distinguish among sperm populations prepared by different methods. One PLS-DA model differentiated wash from POS and SU, attaining 86% sensitivity and 91% accuracy. Another model distinguished between POS and SU, achieving 77% sensitivity and 74% accuracy. The combined use of SERS and multivariate analysis offers a promising alternative for assessing sperm quality, sup-ported by motility assessments in 35 validated samples. This approach could enhance both the accuracy and efficiency of reproductive diagnostics.