JustDeepIt
Image analysis based on deep learning is becoming mainstream and increasingly accessible for solving various scientific problems in diverse fields including plant science. Python programming language and its libraries including PyTorch, MMDetection, and Detectron2 have made deep learning much easier and more accessible to researchers. However, deep learning technologies remain challenging for programming beginners because they require computer programming skills and a basic familiarity with character user interfaces (CUIs). JustDeepIt aims to simplify object detection, instance segmentation, and salient object detection using deep learning by providing graphical user interfaces (GUIs). Additionally, to ensure flexibility and extensibility, JustDeepIt also provides CUIs. It may be applicable for image analysis in various disciplines beyond plant science. (Sun et al., Front. Plant Sci., 2022)
CircSeqAlignTk
CircSeqAlignTk is designed for end-to-end RNA-Seq data analysis of circular genome sequences, from alignment to visualization. It mainly targets viroids which are composed of 246-401 nt circular RNAs. In addition, CircSeqAlignTk implements a tidy interface to generate synthetic sequencing data that mimic real RNA-Seq data, allowing developers to evaluate the performance of alignment tools and workflows. (Sun et al., F1000Res. [preprint], 2022)
Dragonfly Identification
This application predicts the names of approximately 200 species of dragonflies and damselflies that inhabit Japan. In addition to using deep learning for image recognition, the application also incorporates habitat data to improve its predictions. (Sun et al., Front. Ecol. Evol., 2021)