A mobile-phone-based diagnostic tool, which most of the population can easily access, could be a game changer in increasing the number of examinations of people with dental caries. This study aimed to apply a deep learning algorithm in diagnosing the stages of smooth surface caries via smartphone images. Materials and methods: A training dataset consisting of 1902 photos of the smooth surface of teeth taken with an iPhone 7 from 695 people was used. Four deep learning models, consisting of Faster Region-Based Convolutional Neural Networks (Faster R-CNNs), You Only Look Once version 3 (YOLOv3), RetinaNet, and Single-Shot Multi-Box Detector (SSD), were tested to detect initial caries lesions and cavities. The reference standard was the diagnosis of a dentist based on image examination according to the International Caries Classification and Management System (ICCMS) classification. Results: For cavitated caries, YOLOv3 and Faster R-CNN showed the highest sensitivity among the four tested models, at 87.4% and 71.4%, respectively. The sensitivity levels of these two models were only 36.9 % and 26% for visually non-cavitated (VNC). The specificity of the four models reached above 86% for cavitated caries and above 71% for VNC. Conclusion: The clinical application of YOLOv3 and Faster R-CNN models for diagnosing dental caries via smartphone images was promising. The current study provides a preliminary insight into the potential translation of AI from the laboratory to clinical practice.

Nesprin-1-giant and nesprin-2-giant regulate nuclear positioning by the interaction of their C-terminal KASH domains with nuclear membrane SUN proteins and their N-terminal calponin-homology domains with cytoskeletal actin. A number of short isoforms lacking the actin-binding domains are produced by internal promotion. We have evaluated the significance of these shorter isoforms using quantitative RT-PCR and western blotting with site-specific monoclonal antibodies. Within a complete map of nesprin isoforms, we describe two novel nesprin-2 epsilon isoforms for the first time. Epsilon isoforms are similar in size and structure to nesprin-1-alpha. Expression of nesprin isoforms was highly tissue-dependent. Nesprin-2-epsilon-1 was found in early embryonic cells, while nesprin-2-epsilon-2 was present in heart and other adult tissues, but not skeletal muscle. Some cell lines lack shorter isoforms and express only one of the two nesprin genes, suggesting that either of the giant nesprins is sufficient for basic cell functions. For the first time, localisation of endogenous nesprin away from the nuclear membrane was shown in cells where removal of the KASH domain by alternative splicing occurs. By distinguishing between degradation products and true isoforms on western blots, it was found that previously-described beta and gamma isoforms are expressed either at only low levels or with a limited tissue distribution. Two of the shortest alpha isoforms, nesprin-1-alpha-2 and nesprin-2-alpha-1, were found almost exclusively in cardiac and skeletal muscle and a highly conserved and alternatively-spliced exon, available in both nesprin genes, was always included in these tissues. These “muscle-specific” isoforms are thought to form a complex with emerin and lamin A/C at the inner nuclear membrane and mutations in all three proteins cause Emery-Dreifuss muscular dystrophy and/or inherited dilated cardiomyopathy, disorders in which only skeletal muscle and/or heart are affected.

The LINC (linker of nucleoskeleton and cytoskeleton) complex is a proposed mechanical link tethering the nucleo- and cyto-skeleton via the NE (nuclear envelope). The LINC components emerin, lamin A/C, SUN1, SUN2, nesprin-1 and nesprin-2 interact with each other at the NE and also with other binding partners including actin filaments and B-type lamins. Besides the mechanostructural functions, the LINC complex is also involved in signalling pathways and gene regulation. Emerin was the first LINC component associated with a human disease, namely EDMD (Emery–Dreifuss muscular dystrophy). Later on, other components of the LINC complex, such as lamins A/C and small isoforms of nesprin-1 and nesprin-2, were found to be associated with EDMD, reflecting a genetic heterogeneity that has not been resolved so far. Only approximately 46% of the EDMD patients can be linked to genes of LINC and non-LINC components, pointing to further genes involved in the pathology of EDMD. Obvious candidates are the LINC proteins SUN1 and SUN2. Recently, screening of binding partners of LINC components as candidates identified LUMA (TMEM43), encoding a binding partner of emerin and lamins, as a gene involved in atypical EDMD. Nevertheless, such mutations contribute only to a very small fraction of EDMD patients. EDMD-causing mutations in STA/EMD (encoding emerin) that disrupt emerin binding to Btf (Bcl-2-associated transcription factor), GCL (germ cell-less) and BAF (barrier to autointegration factor) provide the first glimpses into LINC being involved in gene regulation and thus opening new avenues for functional studies. Thus the association of LINC with human disease provides tools for understanding its functions within the cell.