Systematic Method for Identifying Kinematic Sperm Subpopulations v2

The identification of kinematic sperm subpopulations typically relies on motility parameter values obtained using a Computer-Assisted Sperm Analysis (CASA) system. These parameters are calculated individually for each sperm cell. The number of motility parameters provided depends on the specific CASA system manufacturer. However, commonly evaluated parameters include curvilinear velocity (VCL), average path velocity (VAP), straight-line velocity (VSL), linearity (LIN), straightness (STR), beat cross-frequency (BCF), wobble (WOB), and amplitude of lateral head displacement (ALH). Various statistical approaches utilize the motility parameter values generated by a CASA system to identify kinematic sperm subpopulations (Martínez-Pastor et al., 2011; Ramón and Martínez-Pastor, 2018). This identification process is typically carried out in two stages. In the first stage, the dataset’s dimensionality is reduced using principal component analysis (PCA). In PCA, components with eigenvalues greater than 1 are retained. In the second stage, these principal components are used as input for clustering algorithms, such as K-means or hierarchical agglomerative clustering. When employing hierarchical agglomerative clustering, the objective is to group data points based on their proximity to one another while maximizing the separation between groups. This is typically achieved using various distance metrics, with Euclidean distance being the most commonly applied. Selecting the appropriate number of clusters is critical and often involves the use of an objective criterion, such as inertia gain. Graphical analysis of inertia gain can be particularly helpful in this context. In this document, we present the strategy and procedure for identifying kinematic sperm subpopulations in boar sperm data subjected to different pH conditions (Rivas et al., 2022). The proposed workflow enables the processing of motility parameter datasets obtained via a CASA system to identify kinematic subpopulations in two systematic steps. The final output is a dataset containing the original input data alongside an additional column indicating the cluster assignment for each data point. This resulting dataset facilitates the determination of the proportion of sperm within each subpopulation and allows for assessing the effects of different treatments on these subpopulations.