No abstract

MicroRNAs (miRNAs) modulate diverse biological and pathological processes via post-transcriptional gene silencing. High-throughput small RNA sequencing (sRNA-seq) has been widely adopted to investigate the functions and regulatory mechanisms of miRNAs. However, accurate quantification of miRNAs has been limited owing to the severe ligation bias in conventional sRNA-seq methods. Here, we quantify miRNAs and their variants (known as isomiRs) by an improved sRNA-seq protocol, termed AQ-seq (accurate quantification by sequencing), that utilizes adapters with terminal degenerate sequences and a high concentration of polyethylene glycol (PEG), which minimize the ligation bias during library preparation. Measurement using AQ-seq allows us to correct the previously misannotated 5′ end usage and strand preference in public databases. Importantly, the analysis of 5′ terminal heterogeneity reveals widespread alternative processing events which have been underestimated. We also identify highly uridylated miRNAs originating from the 3p strands, indicating regulations mediated by terminal uridylyl transferases at the pre-miRNA stage. Taken together, our study reveals the complexity of the miRNA isoform landscape, allowing us to refine miRNA annotation and to advance our understanding of miRNA regulation. Furthermore, AQ-seq can be adopted to improve other ligation-based sequencing methods including crosslinking-immunoprecipitation-sequencing (CLIP-seq) and ribosome profiling (Ribo-seq).

A progressive loss of cartilage matrix leads to the development of osteoarthritis (OA). Matrix homeostasis is disturbed in OA cartilage as the result of reduced production of cartilage-specific matrix and increased secretion of catabolic mediators by chondrocytes. Chondrocyte senescence is a crucial cellular event contributing to such imbalance in matrix metabolism during OA development. Here, we identify miR-204 as a markedly up-regulated microRNA in OA cartilage. miR-204 is induced by transcription factors GATA4 and NF-κB in response to senescence signals. Up-regulated miR-204 simultaneously targets multiple components of the sulfated proteoglycan (PG) biosynthesis pathway, effectively shutting down PG anabolism. Ectopic expression of miR-204 in joints triggers spontaneous cartilage loss and OA development, whereas miR-204 inhibition ameliorates experimental OA, with concomitant recovery of PG synthesis and suppression of inflammatory senescence-associated secretory phenotype (SASP) factors in cartilage. Collectively, we unravel a stress-activated senescence pathway that underlies disrupted matrix homeostasis in OA cartilage.

Strand selection is a critical step in microRNA (miRNA) biogenesis.Although the dominant strand may alter depending on cellular contexts, the molecular mechanism and physiological significance of such alternative strand selection (or "arm switching") remain elusive. Here we find mir-324 as one of the strongly regulated miRNAs by arm switching, and identify terminal uridylyl transferases TUT4 and TUT7 as the key regulators. Uridylation of pre-mir-324 by TUT4/7 re-positions DICER on the pre-miRNA and shifts the cleavage site. This alternative processing produces a duplex with a different terminus, from which the 3 ′ strand (3p) is selected instead of the 5 ′ strand (5p). In glioblastoma, the TUT4/7 and 3p levels are upregulated while the 5p level is reduced. Manipulation of the strand ratio is sufficient to impair glioblastoma cell proliferation. This study uncovers a role of uridylation as a molecular switch in alternative strand selection and implicates its therapeutic potential.

No abstract

The purpose of sentiment classification is to determine whether a particular document has a positive or negative nuance. Sentiment classification is extensively used in many business domains to improve products or services by understanding the opinions of customers regarding these products. Deep learning achieves state-of-the-art results in various challenging domains. With the success of deep learning, many studies have proposed deep-learning-based sentiment classification models and achieved better performances compared with conventional machine learning models. However, one practical issue occurring in deep-learning-based sentiment classification is that the best model structure depends on the characteristics of the dataset on which the deep learning model is trained; moreover, it is manually determined based on the domain knowledge of an expert or selected from a grid search of possible candidates. Herein, we present a comparative study of different deep-learning-based sentiment classification model structures to derive meaningful implications for building sentiment classification models. Specifically, eight deeplearning models, three based on convolutional neural networks and five based on recurrent neural networks, with two types of input structures, i.e., word level and character level, are compared for 13 review datasets, and the classification performances are discussed under different perspectives. INDEX TERMS Sentiment classification, deep learning, convolutional neural network, recurrent neural network, word embedding, character embedding.

No abstract

No abstract