Key points are not available for this paper at this time.
Artificial insemination (AI) plays a critical role in livestock reproduction, with semen quality being essential. In swine, AI primarily uses cool-stored semen adhering to industry standards assessed through routine analysis, yet fertility inconsistencies highlight the need for enhanced semen evaluation. Over 10-day storage at 17 °C, boar semen samples were analyzed for motility, morphology, sperm membrane integrity, apoptosis, and oxidative stress indicators. Additionally, machine learning tools were employed to explore the potential of Raman and near-infrared (NIR) spectroscopy in enhancing semen sample evaluation. Sperm motility and morphology gradually decreased during storage, with distinct groups categorized as “Good” or “Poor” survival semen according to motility on Day 7 of storage. Initially similar on Day 0 of semen collection, “Poor” samples revealed significantly lower total motility (21.69 ± 4.64% vs. 80.19 ± 1.42%), progressive motility (4.74 ± 1.71% vs. 39.73 ± 2.57%), and normal morphology (66.43 ± 2.60% vs. 87.91 ± 1.92%) than their “Good” counterparts by Day 7, using a computer-assisted sperm analyzer. Furthermore, “Poor” samples had higher levels of apoptotic cells, membrane damage, and intracellular reactive oxygen species on Day 0. Conversely, “Good” samples maintained higher total antioxidant capacity. Raman spectroscopy outperformed NIR, providing distinctive spectral profiles aligned with semen biochemical changes and enabling the prediction of semen survival during storage. Overall, the spectral profiles coupled with machine learning tools might assist in enhancing semen evaluation and prognosis.
Kameni et al. (Thu,) studied this question.