Biological sex is one of the more critically important physiological parameters needed for managing threatened animal species because it is crucial for informing several of the management decisions surrounding conservation breeding programs. Near-infrared spectroscopy (NIRS) is a non-invasive technology that has been recently applied in the field of wildlife science to evaluate various aspects of animal physiology and may have potential as an in vivo technique for determining biological sex in live amphibian species. This study investigated whether NIRS could be used as a rapid and non-invasive method for discriminating biological sex in the endangered Houston toad (Anaxyrus houstonensis). NIR spectra (N = 396) were collected from live A. houstonensis individuals (N = 132), and distinct spectral patterns between males and females were identified using chemometrics. Linear discriminant analysis (PCA-LDA) classified the spectra from each biological sex with accuracy ≥ 98% in the calibration and internal validation datasets and 94% in the external validation process. Through the use of NIRS, we have determined that unique spectral signatures can be holistically captured in the skin of male and female anurans, bringing to light the possibility of further application of this technique for juveniles and sexually monomorphic species, whose sex designation is important for breeding-related decisions.

More than 50% of caudates are threatened with extinction and are in need of ex-situ breeding programs to support conservation efforts and species recovery. Unfortunately, many salamander populations under human care can experience reproductive failure, primarily due to missing environmental cues necessary for breeding. Assisted reproductive technologies (ARTs) are a useful suite of techniques for overcoming or bypassing these missing environmental cues to promote breeding. Exogenous hormones are used to stimulate natural breeding behaviors or gamete expression for in-vitro fertilization or biobanking and are typically administered intramuscularly in caudates. While effective, intramuscular injection is risky to perform in smaller-bodied animals, resulting in health and welfare risks. This research investigated the spermiation response to hormone administration through a non-invasive oral bioencapsulation route using the tiger salamander (Ambystoma tigrinum) as a model species. Male salamanders were randomly rotated six weeks apart through four treatments (n = 11 males/treatment) in which animals received a resolving dose of gonadotropin-releasing hormone (GnRH) as follows: (1) Prime-Only (0.0 μg/g); (2) Low (0.25 μg/g); (3) Medium (1.0 μg/g); and (4) High (2.0 μg/g). All males were given a GnRH priming dose (0.25 μg/g) 24 hours prior to the resolving dose. Exogenous hormone was delivered inside of a cricket (Gryllodes sigillatus) that was presented as a food item by tweezers. Sperm samples were collected at 1, 3, 6, 9, 12, and 24 hours after the resolving dose and analyzed for quantity and quality. For all treatments, sperm concentration was produced in an episodic pattern over time. The Prime-Only treatment had a lower (p < 0.05) percent of sperm exhibiting normal morphology compared to treatments utilizing a resolving dose of GnRH. Overall, oral administration of GnRH is a feasible route of inducing spermiation in salamanders, yielding sperm of sufficient quantity and quality for in-vitro fertilization and biobanking efforts.

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.

Near infrared spectroscopy coupled with predictive modeling is a growing field of study for addressing questions in wildlife science aimed at improving management strategies and conservation outcomes for managed and threatened fauna. To date, the majority of spectroscopic studies in wildlife and fisheries applied chemometrics and predictive modeling with a single-algorithm approach. By contrast, multi-model approaches are used routinely for analyzing spectroscopic datasets across many major industries (e.g., medicine, agriculture) to maximize predictive outcomes for real-world applications. In this study, we conducted a benchmark modeling exercise to compare the performance of several machine learning algorithms in a multi-class problem utilizing a multivariate spectroscopic dataset obtained from live animals. Spectra obtained from live individuals representing eleven amphibian species were classified according to taxonomic designation. Seven modeling techniques were applied to generate prediction models, which varied significantly (p < 0.05) with regard to mean classification accuracy (e.g., support vector machine: 95.8 ± 0.8% vs. K-nearest neighbors: 89.3 ± 1.0%). Through the use of a multi-algorithm approach, candidate algorithms can be identified and applied to more effectively model complex spectroscopic data collected for wildlife sciences. Other key considerations in the predictive modeling workflow that serve to optimize spectroscopic model performance (e.g., variable selection and cross-validation procedures) are also discussed.