Infection with COVID-19 promotes the progression of pancreatic cancer through the PI3K-AKT signaling pathway.

OBJECTIVE: To investigate the effect of COVID-19 infection on pancreatic cancer. METHODS: Based on the mRNA-Seq data of COVID-19 patients and pancreatic cancer (PC) patients in the GEO database, we used a support vector machine (SVM), LASSO-Cox regression analysis and random forest tree (RF) to screen the common signature genes of the two diseases and further investigate their effects and functional characteristics on PC, respectively. The above procedures were performed in R software. RESULTS: The proteins COL10A1/FAP/FN1 were found to be common signature genes for COVID-19 and PC, were significantly up-regulated in both diseases and showed good diagnostic efficacy for PC. The risk model based on COL10A1/FAP/FN1 showed good PC risk prediction ability and clinical application potential. Tumor typing based on COL10A1/FAP/FN1 expression levels effectively classified PC into different subtypes and showed significant differences between the two subtypes in terms of survival prognosis, immune levels, immune checkpoint expression levels, mutation status of common tumor mutation sites, and drug sensitivity analysis. While pathway analysis also revealed that FN1 as an extracellular matrix component may be involved in the biological process of PC by regulating the PI3K-AKT signaling axis. CONCLUSION: The upregulated expression of COL10A1/FAP/FN1, the characteristic genes of COVID-19, are potential diagnostic targets for PC, and the upregulated expression of FN1 may promote the progression of PC by activating the PI3K-AKT signaling pathway. The COL10A1/FAP/FN1-based typing provides a new typing approach for PC, and also provides a good reference and idea for the refinement of PC treatment and subsequent clinical research.

A Deep Learning Based Approach for Grading of Diabetic Retinopathy Using Large Fundus Image Dataset.

Diabetic Retinopathy affects one-third of all diabetic patients and may cause vision impairment. It has four stages of progression, i.e., mild non-proliferative, moderate non-proliferative, severe non-proliferative and proliferative Diabetic Retinopathy. The disease has no noticeable symptoms at early stages and may lead to chronic destruction, thus causing permanent blindness if not detected at an early stage. The proposed research provides deep learning frameworks for autonomous detection of Diabetic Retinopathy at an early stage using fundus images. The first framework consists of cascaded neural networks, spanned in three layers where each layer classifies data into two classes, one is the desired stage and the other output is passed to another classifier until the input image is classified as one of the stages. The second framework takes normalized, HSV and RGB fundus images as input to three Convolutional Neural Networks, and the resultant probabilistic vectors are averaged together to obtain the final output of the input image. Third framework used the Long Short Term Memory Module in CNN to emphasize the network in remembering information over a long time span. Proposed frameworks were tested and compared on the large-scale Kaggle fundus image dataset EYEPAC. The evaluations have shown that the second framework outperformed others and achieved an accuracy of 78.06% and 83.78% without and with augmentation, respectively.

Iterative principal component analysis method for improvised classification of breast cancer disease using blood sample analysis.

Breast cancer is the most common cancer in women occurring worldwide. Some of the procedures used to diagnose breast cancer are mammogram, breast ultrasound, biopsy, breast magnetic resonance imaging, and blood tests such as complete blood count. Detecting breast cancer at an early stage plays an important role in diagnostic and curative procedures. This paper aims to develop a predictive model for detecting the breast cancer using blood samples data containing age, body mass index (BMI), glucose, insulin, homeostasis model assessment (HOMA), leptin, adiponectin, resistin, and chemokine monocyte chemoattractant protein 1 (MCP-1).The two main challenges encountered in this process are identification of biomarkers and the precision of disease prediction accuracy. The proposed methodology employs principal component analysis in a peculiar approach followed by random forest tree prediction model to discriminate between healthy and breast cancer patients. This approach extracts high communalities, a linear combination of input attributes in a systematic procedure as principal axis elements. The iteratively extracted principal axis elements combined with minimum number of input attributes are able to predict the disease with higher accuracy of classification with increased sensitivity and specificity score. The results proved that the proposed approach generates a higher predictor performance than the previous reported results by opting relevant extracted principal axis elements and attributes that commend the classifier with increased performance measures.

Significant changes in the chemical compositions and sources of PM2.5 in Wuhan since the city lockdown as COVID-19.

Wuhan was the first city to adopt the lockdown measures to prevent COVID-19 spreading, which improved the air quality accordingly. This study investigated the variations in chemical compositions, source contributions, and regional transport of fine particles (PM2.5) during January 23-February 22 of 2020, compared with the same period in 2019. The average mass concentration of PM2.5 decreased from 72.9 µg m-3 (2019) to 45.9 µg m-3 (2020), by 27.0 µg m-3. It was predominantly contributed by the emission reduction (92.0%), retrieved from a random forest tree approach. The main chemical species of PM2.5 all decreased with the reductions ranging from 0.85 µg m-3 (chloride) to 9.86 µg m-3 (nitrate) (p < 0.01). Positive matrix factorization model indicated that the mass contributions of seven PM2.5 sources all decreased. However, their contribution percentages varied from -11.0% (industrial processes) to 8.70% (secondary inorganic aerosol). Source contributions of PM2.5 transported from potential geographical regions showed reductions with mean values ranging from 0.22 to 4.36 µg m-3. However, increased contributions of firework burning, secondary inorganic aerosol, road dust, and vehicle emissions from transboundary transport were observed. This study highlighted the complex and nonlinear response of chemical compositions and sources of PM2.5 to air pollution control measures, suggesting the importance of regional-joint control.

Association of suboptimal health status with intestinal microbiota in Chinese youths.

Suboptimal health status (SHS), a physical state between health and disease, is a subclinical and reversible stage of chronic disease. Previous studies have shown alterations in the intestinal microbiota in patients with some chronic diseases. This study aimed to investigate the association between SHS and intestinal microbiota in a case-control study with 50 SHS individuals and 50 matched healthy controls. Intestinal microbiota was analysed by MiSeq 250PE. Alpha diversity of intestinal microbiota in SHS individuals was higher compared with that of healthy controls (Simpson index, W = 2238, P = .048). Beta diversity was different between SHS and healthy controls (P = .018). At the phylum level, the relative abundance of Verrucomicrobia was higher in the SHS group than that in the controls (W = 2201, P = .049). Compared with that of the control group, nine genera were significantly higher and five genera were lower in abundance in the SHS group (all P < .05). The intestinal microbiota, analysed by a random forest model, was able to distinguish individuals with SHS from the controls, with an area under the curve of 0.79 (95% confidence interval: 0.77-0.81). We demonstrated that the alteration of intestinal microbiota occurs with SHS, an early stage of disease, which might shed light on the importance of intestinal microbiota in the primary prevention of noncommunicable chronic diseases.

Foot placement during error and pedal applications in naturalistic driving.

Data from a naturalistic driving study was used to examine foot placement during routine foot pedal movements and possible pedal misapplications. The study included four weeks of observations from 30 drivers, where pedal responses were recorded and categorized. The foot movements associated with pedal misapplications and errors were the focus of the analyses. A random forest algorithm was used to predict the pedal application types based the video observations, foot placements, drivers' characteristics, drivers' cognitive function levels and anthropometric measurements. A repeated multinomial logit model was then used to estimate the likelihood of the foot placement given various driver characteristics and driving scenarios. The findings showed that prior foot location, the drivers' seat position, and the drive sequence were all associated with incorrect foot placement during an event. The study showed that there is a potential to develop a driver assistance system that can reduce the likelihood of a pedal error.