Search

The COVID-19 pandemic has frequently produced more highly transmissible SARS-CoV-2 variants, such as Omicron, which has produced sublineages. It is a challenge to tell apart high-risk Omicron sublineages and other lineages of SARS-CoV-2 variants. We aimed to build a fine-grained deep learning (DL) model to assess SARS-CoV-2 transmissibility, updating our former coarse-grained model, with the training/validating data of early-stage SARS-CoV-2 variants and based on sequential Spike samples. Sequential amino acid (AA) frequency was decomposed into serially and slidingly windowed fragments in Spike. Unsupervised machine learning approaches were performed to observe the distribution in sequential AA frequency and then a supervised Convolutional Neural Network (CNN) was built with three adaptation labels to predict the human adaptation of Omicron variants in sublineages. Results indicated clear inter-lineage separation and intra-lineage clustering for SARS-CoV-2 variants in the decomposed sequential AAs. Accurate classification by the predictor was validated for the variants with different adaptations. Higher adaptation for the BA.2 sublineage and middle-level adaptation for the BA.1/BA.1.1 sublineages were predicted for Omicron variants. Summarily, the Omicron BA.2 sublineage is more adaptive than BA.1/BA.1.1 and has spread more rapidly, particularly in Europe. The fine-grained adaptation DL model works well for the timely assessment of the transmissibility of SARS-CoV-2 variants, facilitating the control of emerging SARS-CoV-2 variants. [ABSTRACT FROM AUTHOR]

With recent advancements in machine learning, especially in deep learning, the prediction of eye diseases based on fundus photography using deep convolutional neural networks (DCNNs) has attracted great attention. However, studies focusing on identifying the right disease among several candidates, which is a better approximation of clinical diagnosis in practice comparing with the case that aims to distinguish one particular eye disease from normal controls, are limited. The performance of existing algorithms for multi-class classification of fundus images is at most mediocre. Moreover, in many studies consisting of different eye diseases, labeled images are quite limited mainly due to privacy concern of patients. In this case, it is infeasible to train huge DCNNs, which usually have millions of parameters. To address these challenges, we propose to utilize a lightweight deep learning architecture called MobileNetV2 and transfer learning to distinguish four common eye diseases, including Glaucoma, Maculopathy, Pathological Myopia, and Retinitis Pigmentosa, from normal controls using a small training data. We also apply a visualization approach to highlight the loci that are most related to the disease labels to make the model more explainable. The highlighted area chosen by the algorithm itself may give some hints for further fundus image studies. Our experimental results show that our system achieves an average accuracy of 96.2%, sensitivity of 90.4%, and specificity of 97.6% on the test data via five independent runs, and outperforms two other deep learning-based algorithms both in terms of accuracy and efficiency. [ABSTRACT FROM AUTHOR]