HostNet: improved sequence representation in deep neural networks for virus-host prediction.

BACKGROUND: The escalation of viruses over the past decade has highlighted the need to determine their respective hosts, particularly for emerging ones that pose a potential menace to the welfare of both human and animal life. Yet, the traditional means of ascertaining the host range of viruses, which involves field surveillance and laboratory experiments, is a laborious and demanding undertaking. A computational tool with the capability to reliably predict host ranges for novel viruses can provide timely responses in the prevention and control of emerging infectious diseases. The intricate nature of viral-host prediction involves issues such as data imbalance and deficiency. Therefore, developing highly accurate computational tools capable of predicting virus-host associations is a challenging and pressing demand. RESULTS: To overcome the challenges of virus-host prediction, we present HostNet, a deep learning framework that utilizes a Transformer-CNN-BiGRU architecture and two enhanced sequence representation modules. The first module, k-mer to vector, pre-trains a background vector representation of k-mers from a broad range of virus sequences to address the issue of data deficiency. The second module, an adaptive sliding window, truncates virus sequences of various lengths to create a uniform number of informative and distinct samples for each sequence to address the issue of data imbalance. We assess HostNet's performance on a benchmark dataset of "Rabies lyssavirus" and an in-house dataset of "Flavivirus". Our results show that HostNet surpasses the state-of-the-art deep learning-based method in host-prediction accuracies and F1 score. The enhanced sequence representation modules, significantly improve HostNet's training generalization, performance in challenging classes, and stability. CONCLUSION: HostNet is a promising framework for predicting virus hosts from genomic sequences, addressing challenges posed by sparse and varying-length virus sequence data. Our results demonstrate its potential as a valuable tool for virus-host prediction in various biological contexts. Virus-host prediction based on genomic sequences using deep neural networks is a promising approach to identifying their potential hosts accurately and efficiently, with significant impacts on public health, disease prevention, and vaccine development.

Evolving landscape of research on cancer-related cognitive impairment: A bibliometric analysis.

Objective: This study describes the state of the art in the field of cancer-related cognitive impairment (CRCI) to facilitate research opportunities in future CRCI research. Methods: Five databases were searched: PubMed, Web of Science, Cochrane Library, Cumulative Index to Nursing and Allied Health (CINAHL), and PsycINFO, from inception to August 20, 2022. Python, VOSviewer, and CiteSpace software were used for data preprocessing and analysis. Results: The published articles were predominantly from the United States, followed by China and Canada. Breast cancer and brain tumors were the dominant cancer types. The study population consisted mainly of adult cancer survivors. Prospective and multicenter studies were the most frequently used study designs. Keyword co-occurrence and mutation analysis indicated major themes: drug therapy was the most common treatment cluster, and adverse effects were another major cluster. The etiology of CRCI was a research hotspot and included the exploration of chemotherapy-associated and psychosocial factors by using measurement tools, such as neuropsychological tests and treatment outcomes. Conclusions: This study's findings highlight CRCI as a major research area, on the basis of the significantly increasing number of annual publications. Keyword co-occurrence analysis provided a quantitative visualization of the current research status for CRCI, but this method cannot provide in-depth qualitative insights explaining the potential emerging trends in this field.

A global dataset of sequence, diversity and biosafety recommendation of arbovirus and arthropod-specific virus.

Arthropod-borne virus (arbovirus) and arthropod-specific virus (ASV) are viruses circulating amongst hematophagous arthropods that are broadly transmitted in ecological systems. Arbovirus may replicate in both vertebrates and invertebrates and some are known to be pathogenic to animals or humans. ASV only replicate in invertebrate arthropods yet they are basal to many types of arboviruses. We built a comprehensive dataset of arbovirus and ASV by curating globally available data from the Arbovirus Catalog, the arbovirus list in Section VIII-F of the Biosafety in Microbiological and Biomedical Laboratories 6th edition, Virus Metadata Resource of International Committee on Taxonomy of Viruses, and GenBank. Revealing the diversity, distribution and biosafety recommendation of arbovirus and ASV at a global scale is essential to the understanding of potential interactions, evolution, and risks associated with these viruses. Moreover, the genomic sequences associated with the dataset will enable the investigation of genetic patterns distinguishing the two groups, as well as aid in predicting the vector/host relationships of the newly discovered viruses.

[Accuracy verification of diet recording tool based on WeChat applet].

OBJECTIVE: To explore the accuracy of a dietary recording tool based on the mobile phone WeChat applet-"Zhishi AI Dietitian" applied to dietary records. METHODS: The research subjects were 109 full-time undergraduates from Zhejiang University. Respondents completed one round of dietary records of "Zhishi AI Dietitian" for three non-consecutive days and one round of non-consecutive three-day 24-hour dietary review method records. The two method must overlap for one day. The energy, nutrients and various food intake data obtained from the Zhishi AI nutritionist survey were sorted and compared with the corresponding survey result of the 24-hour dietary review method. Pearson correlation coefficient or Spearman correlation coefficient was used for correlation analysis, intra-group correlation coefficient was used for reliability analysis, and Bland-Atlman scatter plot was used for consistency analysis. RESULTS: In terms of reliability, the two method had certain reliability in assessing intake of various foods, energy and nutrients. After energy correction, the reliability of nutrient intake was enhanced. In terms of correlation, the correlation coefficients of food groups ranged from 0.34 to 0.79(mean 0.60), and the energy and nutrient correlation coefficients ranged from 0.34 to 0.72(mean 0.55). In terms of consistency, the proportion of research subjects outside the 95% consistency interval is less than 10%, indicating that the two have good consistency. CONCLUSION: Zhishi AI Dietitian applied to college students' dietary records has good accuracy.

Comparison and development of advanced machine learning tools to predict nonalcoholic fatty liver disease: An extended study.

BACKGROUND: Nonalcoholic fatty liver disease (NAFLD) is a public health challenge and significant cause of morbidity and mortality worldwide. Early identification is crucial for disease intervention. We recently proposed a nomogram-based NAFLD prediction model from a large population cohort. We aimed to explore machine learning tools in predicting NAFLD. METHODS: A retrospective cross-sectional study was performed on 15 315 Chinese subjects (10 373 training and 4942 testing sets). Selected clinical and biochemical factors were evaluated by different types of machine learning algorithms to develop and validate seven predictive models. Nine evaluation indicators including area under the receiver operating characteristic curve (AUROC), area under the precision-recall curve (AUPRC), accuracy, positive predictive value, sensitivity, F1 score, Matthews correlation coefficient (MCC), specificity and negative prognostic value were applied to compare the performance among the models. The selected clinical and biochemical factors were ranked according to the importance in prediction ability. RESULTS: Totally 4018/10 373 (38.74%) and 1860/4942 (37.64%) subjects had ultrasound-proven NAFLD in the training and testing sets, respectively. Seven machine learning based models were developed and demonstrated good performance in predicting NAFLD. Among these models, the XGBoost model revealed the highest AUROC (0.873), AUPRC (0.810), accuracy (0.795), positive predictive value (0.806), F1 score (0.695), MCC (0.557), specificity (0.909), demonstrating the best prediction ability among the built models. Body mass index was the most valuable indicator to predict NAFLD according to the feature ranking scores. CONCLUSIONS: The XGBoost model has the best overall prediction ability for diagnosing NAFLD. The novel machine learning tools provide considerable beneficial potential in NAFLD screening.

Promoting healthy lifestyle in Chinese college students: evaluation of a social media-based intervention applying the RE-AIM framework.

BACKGROUND: A health program aiming at college students is pressingly needed to improve their lifestyle and prevent diseases. However, a health intervention often requires health facilities and the many efforts of health workers. This project attempts to evolve traditional health intervention by using integrated methods based on social media and multiple mobile tools. METHODS: A total of 110 undergraduates from Zhejiang University were recruited. In all, 87 participants volunteered to enroll in the intervention group, whereas 23 stayed in a control group. Fifteen staff (dietitians, health assistants and a sports coach) used the WeChat app and its plugin Zhishi mini-program for health education, diet and physical activity (PA) supervision during 21 days. Pre-to-post changes of eating habits, physical fitness tests and anthropometry data were measured. The RE-AIM framework was employed to evaluate the intervention, dimensions of which were Reach, Effectiveness, Adoption, Implementation, and Maintenance. RESULTS: The recruitment rate of students was 79.1%. The intervention group showed significant progress in terms of healthy food intake (all P < 0.05), and an improvement in PA level (P = 0.004) over 21 days. About 60.9% subjects were satisfied with the whole program and 64.4% would like to join the program again. CONCLUSIONS: This intervention showed a great improvement in healthy behavior with great feasibility for further dissemination.