Fibrosis is a common pathology in cardiovascular disease1. In the heart, fibrosis causes mechanical and electrical dysfunction1,2 and in the kidney, it predicts the onset of renal failure3. Transforming growth factor β1 (TGFβ1) is the principal pro-fibrotic factor4,5, but its inhibition is associated with side effects due to its pleiotropic roles6,7. We hypothesized that downstream effectors of TGFβ1 in fibroblasts could be attractive therapeutic targets and lack upstream toxicity. Here we show, using integrated imaging–genomics analyses of primary human fibroblasts, that upregulation of interleukin-11 (IL-11) is the dominant transcriptional response to TGFβ1 exposure and required for its pro-fibrotic effect. IL-11 and its receptor (IL11RA) are expressed specifically in fibroblasts, in which they drive non-canonical, ERK-dependent autocrine signalling that is required for fibrogenic protein synthesis. In mice, fibroblast-specific Il11 transgene expression or Il-11 injection causes heart and kidney fibrosis and organ failure, whereas genetic deletion of Il11ra1 protects against disease. Therefore, inhibition of IL-11 prevents fibroblast activation across organs and species in response to a range of important pro-fibrotic stimuli. These results reveal a central role of IL-11 in fibrosis and we propose that inhibition of IL-11 is a potential therapeutic strategy to treat fibrotic diseases.

Transforming growth factor beta-1 (TGFβ1) is a major driver of vascular smooth muscle cell (VSMC) phenotypic switching, an important pathobiology in arterial disease. We performed RNA-sequencing of TGFβ1-stimulated human aortic or arterial VSMCs which revealed large and consistent upregulation of Interleukin 11 (IL11). IL11 has an unknown function in VSMCs, which highly express the IL11 receptor alpha, suggestive of an autocrine loop. In vitro, IL11 activated ERK signaling, but inhibited STAT3 activity, and caused VSMC phenotypic switching to a similar extent as TGFβ1 or angiotensin II (ANGII) stimulation. Genetic or therapeutic inhibition of IL11 signaling reduced TGFβ1- or ANGII-induced VSMC phenotypic switching, placing IL11 activity downstream of these factors. Aortas of mice with Myh11-driven IL11 expression were remodeled and had reduced contractile but increased matrix and inflammatory genes expression. In two models of arterial pressure loading, IL11 was upregulated in the aorta and neutralizing IL11 antibodies reduced remodeling along with matrix and pro-inflammatory gene expression. These data show that IL11 plays an important role in VSMC phenotype switching, vascular inflammation and aortic pathobiology.

Big data analytics has shown tremendous success in several fields such as businesses, agriculture, health, and meteorology, and education is no exception. Concerning its role in education, it is used to boost students' learning process by predicting their performance in advance and adapting the relevant instructional design strategies. This study primarily intends to develop a system that can predict students' performance and help teachers to timely introduce corrective interventions to uplift the performance of low-performing students. As a secondary part of this research, it also explores the potential of collaborative learning as an intervention to act in combination with the prediction system to improve the performance of students. To support such changes, a visualization system is also developed to track and monitor the performance of students, groups, and overall class to help teachers in the regrouping of students concerning their performance. Several well-known machine learning models are applied to predict students performance. Results suggest that experimental groups performed better after treatment than before treatment. The students who took part in each class activity, prepared and submitted their tasks perform much better than other students. Overall, the study found that collaborative learning methods play a significant role to enhance the learning capability of the students.

Sentiment analysis is a widely researched area due to its various applications in customer services, brand monitoring, and market research. Automatic sentiment classification is an important but challenging task. Contrary to the English language, sentiment analysis for low-resource languages like Urdu is an under-explored research area. Most of the work on sentiment analysis in the Urdu language is domain-dependent where models are mostly trained and tested on the same dataset on limited domains. However, sentiments in different domains are expressed differently, and manually annotating the datasets for all possible domains is unfeasible. Training a sentiment classifier using annotated data on one domain and testing it on another domain results in poor performance as the terms appearing in the source domain (training data) might not appear in the target (testing data) domain. In this paper, we present a baseline method for cross-domain sentiment analysis in the Urdu language using two different domains. Feature extraction is performed using n-grams and word embedding techniques. Sentiment classification is performed using machine learning and deep learning classifiers. The proposed method achieves an accuracy, precision, recall, and F1 scores of 0.77, 0.83, 0.68, and 0.75, respectively.INDEX TERMS Cross-domain sentiment analysis; deep learning; Urdu language processing; feature engineering

No abstract