Artificial neural networks for the definition of kinetic subpopulations in electroejaculated and epididymal spermatozoa in the domestic cat

Abstract

This study was designed for the identification of different sperm kinetic subpopulations in feline semen using artificial neural networks (ANNs) and for the evaluation of the effect of ejaculation on motility patterns of these subpopulations. Seven tomcats presented for routine orchiectomy were electroejaculated, and after 5 days, orchiectomized and epididymal tail sperms were collected. Sperm motility characteristics were evaluated using a computer-assisted sperm analyzer that provided individual kinetic characteristics of each spermatozoon. A total of 23 400 spermatozoa for electroejaculated and 9200 for epididymal tail samples were evaluated using a multivariate approach, comprising principal component analysis and ANN classification. The multivariate approach allowed the identification and characterization of three different and well-defined sperm subpopulations. There were significant differences before (epididymal tail spermatozoa) and after (electroejaculated sperm) ejaculation in sperm kinetic subpopulation characteristics. In both epididymal and ejaculated samples, the majority of subpopulation was characterized by high velocity and progressiveness; however, the electroejaculated samples showed significantly higher values, suggesting that the microenvironment of the epididymal tail could affect the sperm motility or, alternatively, seminal plasma could increase the kinetic characteristics of the spermatozoa, indicating that only after ejaculation, the spermatozoa express their motility potential. Nevertheless, further studies are required to clarify the functional significance of each kinetic subpopulation.