Comprehensive Summary
This study presented the novel deep learning algorithm RetroNet. RetroNet is presented as a method for detection for low-mosaicism somatic mobile element insertions (MEIs). In particular, RetroNet was trained across a diverse dataset and was found to be able to accurately detect somatic MEIs with low mosaicism with as few as two reads. RetroNet was found to outperform both manual examinations alongside other previously established methods, with RetroNet demonstrating high precision (0.885) and recall (0.579) on cancer cell lines. It was also found as an effective tool for identification of degraded DNA. Overall, this study presents the novel AI-based algorithm RetroNet which demonstrated a high degree of accuracy in identification of low mosaic somatic MEIs, indicating an effective and new approach to screening somatic DNA for MEIs.
Outcomes and Implications
In this study, RetroNet is presented as a more accurate and efficient alternative to determining low-mosaicism somatic mobile element insertions (MEIs). Especially because this deep learning algorithm is able to identify both degraded DNA and low-mosaicism somatic MEIs, this method may help health-care providers more precisely recognize these MEIs. Not only can this allow for earlier treatment and gene therapy, but also this method can help geneticists expand current treatment methods to better address MEIs. Overall, RetroNet may help health-care professionals diagnose MEIs and may also allow researchers to gain a deeper understanding of the genetic implications of somatic MEIs.