XIBBIT: A biometric recognition tool for efficient Xenopus laevis identification and colony management

Abstract

The African clawed frog Xenopus laevis is a widely utilized model organism in biomedical research; however, significant challenges in experimental reproducibility and colony management remain. A major obstacle lies in the reliable identification of individual animals, since frogs are generally housed in large groups and are difficult to distinguish due to their high morphological similarity. Conventional methods, including toe clipping and microchipping, are invasive and cause distress, emphasizing the need for non-invasive methods for accurate documentation and welfare monitoring. In this study, we introduce XIBBIT (Xenopus Image-Based Biometric-pattern Identification Tool), a web-based application integrating computer vision and machine learning to identify individual Xenopus laevis based on their dorsal patterning. By exploiting these natural biometric signatures, the platform achieves reliable identification with up to 95.7% accuracy within three image captures under real life conditions. In addition to identification, XIBBIT provides a centralized colony management system. It archives individual data, including health records and experimental histories, with customizable fields. To demonstrate XIBBITs capabilities, we used the application to track egg quality across repeated egg-laying events, revealing that egg quality is a repeatable, individual-specific trait in Xenopus laevis. Furthermore, we find seasonal effects on egg laying performance with the lowest performance during late-spring and summer months. Ultimately, XIBBIT provides an effective, time-efficient, and non-invasive solution to the problem of individual Xenopus laevis identification, facilitating both experimental reproducibility and high animal welfare standards.