Sperm motility assessed by deep convolutional neural networks into WHO categories.

Semen analysis is central in infertility investigation. Manual assessment of sperm motility according to the WHO recommendations is the golden standard, and extensive training is a requirement for accurate and reproducible results. Deep convolutional neural networks (DCNN) are especially suitable for image classification. In this study, we evaluated the performance of the DCNN ResNet-50 in predicting the proportion of sperm in the WHO motility categories. Two models were evaluated using tenfold cross-validation with 65 video recordings of wet semen preparations from an external quality assessment programme for semen analysis. The corresponding manually assessed data was obtained from several of the reference laboratories, and the mean values were used for training of the DCNN models. One model was trained to predict the three categories progressive motility, non-progressive motility, and immotile spermatozoa. Another model was used in predicting four categories, where progressive motility was differentiated into rapid and slow. The resulting average mean absolute error (MAE) was 0.05 and 0.07, and the average ZeroR baseline was 0.09 and 0.10 for the three-category and the four-category model, respectively. Manual and DCNN-predicted motility was compared by Pearson's correlation coefficient and by difference plots. The strongest correlation between the mean manually assessed values and DCNN-predicted motility was observed for % progressively motile spermatozoa (Pearson's r = 0.88, p < 0.001) and % immotile spermatozoa (r = 0.89, p < 0.001). For rapid progressive motility, the correlation was moderate (Pearson's r = 0.673, p < 0.001). The median difference between manual and predicted progressive motility was 0 and 2 for immotile spermatozoa. The largest bias was observed at high and low percentages of progressive and immotile spermatozoa. The DCNN-predicted value was within the range of the interlaboratory variation of the results for most of the samples. In conclusion, DCNN models were able to predict the proportion of spermatozoa into the WHO motility categories with significantly lower error than the baseline. The best correlation between the manual and the DCNN-predicted motility values was found for the categories progressive and immotile. Of note, there was considerable variation between the mean motility values obtained for each category by the reference laboratories, especially for rapid progressive motility, which impacts the training of the DCNN models.

VISEM-Tracking, a human spermatozoa tracking dataset.

A manual assessment of sperm motility requires microscopy observation, which is challenging due to the fast-moving spermatozoa in the field of view. To obtain correct results, manual evaluation requires extensive training. Therefore, computer-aided sperm analysis (CASA) has become increasingly used in clinics. Despite this, more data is needed to train supervised machine learning approaches in order to improve accuracy and reliability in the assessment of sperm motility and kinematics. In this regard, we provide a dataset called VISEM-Tracking with 20 video recordings of 30 seconds (comprising 29,196 frames) of wet semen preparations with manually annotated bounding-box coordinates and a set of sperm characteristics analyzed by experts in the domain. In addition to the annotated data, we provide unlabeled video clips for easy-to-use access and analysis of the data via methods such as self- or unsupervised learning. As part of this paper, we present baseline sperm detection performances using the YOLOv5 deep learning (DL) model trained on the VISEM-Tracking dataset. As a result, we show that the dataset can be used to train complex DL models to analyze spermatozoa.

Levels of L-carnitine in human seminal plasma are associated with sperm fatty acid composition.

The fatty acid composition of spermatozoa has been shown to be important for their function, and L-carnitine is crucial for fatty acid metabolism. Its levels in the seminal plasma positively correlate with semen quality, whereas high body mass index (BMI) is associated with both reduced semen quality and altered sperm fatty acid composition. Here, we examined the associations between free seminal L-carnitine levels and sperm fatty acid composition as well as BMI. Semen samples were collected and analyzed from 128 men with unknown fertility status and with BMI ranging from 19 kg m-2 to 63 kg m-2. Sperm fatty acid composition was assessed by gas chromatography, while free seminal L-carnitine analysis was performed using high-performance liquid chromatography. Multiple linear regression analysis showed a positive correlation of free seminal L-carnitine levels with the amount of sperm palmitic acid (ß = 0.21; P = 0.014), docosahexaenoic acid (DHA; ß = 0.23; P = 0.007), and total n-3 polyunsaturated fatty acids (ß = 0.23; P = 0.008) and a negative correlation of free seminal L-carnitine levels with lignoceric acid (ß = -0.29; P = 0.001) and total n-6 polyunsaturated fatty acids (ß = -0.24; P = 0.012) when adjusted for covariates. There was no relationship between free seminal L-carnitine levels and BMI. Since free seminal L-carnitine levels are associated with semen quality, the absence of a correlation with BMI suggests that reduced semen quality in obese men is independent of seminal L-carnitine.

Machine Learning-Based Analysis of Sperm Videos and Participant Data for Male Fertility Prediction.

Methods for automatic analysis of clinical data are usually targeted towards a specific modality and do not make use of all relevant data available. In the field of male human reproduction, clinical and biological data are not used to its fullest potential. Manual evaluation of a semen sample using a microscope is time-consuming and requires extensive training. Furthermore, the validity of manual semen analysis has been questioned due to limited reproducibility, and often high inter-personnel variation. The existing computer-aided sperm analyzer systems are not recommended for routine clinical use due to methodological challenges caused by the consistency of the semen sample. Thus, there is a need for an improved methodology. We use modern and classical machine learning techniques together with a dataset consisting of 85 videos of human semen samples and related participant data to automatically predict sperm motility. Used techniques include simple linear regression and more sophisticated methods using convolutional neural networks. Our results indicate that sperm motility prediction based on deep learning using sperm motility videos is rapid to perform and consistent. Adding participant data did not improve the algorithms performance. In conclusion, machine learning-based automatic analysis may become a valuable tool in male infertility investigation and research.

Body Mass Index Is Associated with Impaired Semen Characteristics and Reduced Levels of Anti-Müllerian Hormone across a Wide Weight Range.

There is still controversy as to how body mass index (BMI) affects male reproduction. We investigated how BMI is associated with semen quality and reproductive hormones in 166 men, including 38 severely obese men. Standard semen analysis and sperm DNA integrity analysis were performed, and blood samples were analysed for reproductive hormones. Adjusted for age and time of abstinence, BMI was negatively associated with sperm concentration (B = -0.088, P = 0.009), total sperm count (B = -0.223, P = 0.001), progressive sperm motility (B = -0.675, P = 0.007), normal sperm morphology (B = -0.078, P = 0.001), and percentage of vital spermatozoa (B = -0.006, P = 0.027). A negative relationship was observed between BMI and total testosterone (B = -0.378, P < 0.001), sex hormone binding globulin (B = -0.572, P < 0.001), inhibin B (B = -3.120, P < 0.001) and anti-Müllerian hormone (AMH) (B = -0.009, P < 0.001). Our findings suggest that high BMI is negatively associated with semen characteristics and serum levels of AMH.

Optic atrophy 1 is an A-kinase anchoring protein on lipid droplets that mediates adrenergic control of lipolysis.

Adrenergic stimulation of adipocytes yields a cAMP signal that activates protein kinase A (PKA). PKA phosphorylates perilipin, a protein localized on the surface of lipid droplets that serves as a gatekeeper to regulate access of lipases converting stored triglycerides to free fatty acids and glycerol in a phosphorylation-dependent manner. Here, we report a new function for optic atrophy 1 (OPA1), a protein known to regulate mitochondrial dynamics, as a dual-specificity A-kinase anchoring protein associated with lipid droplets. By a variety of protein interaction assays, immunoprecipitation and immunolocalization experiments, we show that OPA1 organizes a supramolecular complex containing both PKA and perilipin. Furthermore, by a combination of siRNA-mediated knockdown, reconstitution experiments using full-length OPA1 with or without the ability to bind PKA or truncated OPA1 fused to a lipid droplet targeting domain and cellular delivery of PKA anchoring disruptor peptides, we demonstrate that OPA1 targeting of PKA to lipid droplets is necessary for hormonal control of perilipin phosphorylation and lipolysis.

Dissociating the centrosomal matrix protein AKAP450 from centrioles impairs centriole duplication and cell cycle progression.

Centrosomes provide docking sites for regulatory molecules involved in the control of the cell division cycle. The centrosomal matrix contains several proteins, which anchor kinases and phosphatases. The large A-Kinase Anchoring Protein AKAP450 is acting as a scaffolding protein for other components of the cell signaling machinery. We selectively perturbed the centrosome by modifying the cellular localization of AKAP450. We report that the expression in HeLa cells of the C terminus of AKAP450, which contains the centrosome-targeting domain of AKAP450 but not its coiled-coil domains or binding sites for signaling molecules, leads to the displacement of the endogenous centrosomal AKAP450 without removing centriolar or pericentrosomal components such as centrin, gamma-tubulin, or pericentrin. The centrosomal protein kinase A type II alpha was delocalized. We further show that this expression impairs cytokinesis and increases ploidy in HeLa cells, whereas it arrests diploid RPE1 fibroblasts in G1, thus further establishing a role of the centrosome in the regulation of the cell division cycle. Moreover, centriole duplication is interrupted. Our data show that the association between centrioles and the centrosomal matrix protein AKAP450 is critical for the integrity of the centrosome and for its reproduction.