Exosomes are cell-derived vesicles and are abundant in biological fluids; they contain RNA molecules which may serve as potential diagnostic biomarkers in 'precision medicine'. To promote the clinical application of exosomal RNA (exoRNA), many isolation methods must be compared and validated. Exosomes in cell culture medium (CCM) and serum may be isolated using ultracentrifugation (UC), ExoQuick or Total Exosome Isolation Reagent (TEI), and exoRNA may be extracted using TRIzol-LS, SeraMir, Total Exosome RNA Isolation (TER), HiPure Liquid RNA/miRNA kit (HLR), miRNeasy or exoRNeasy. ExoRNA was assessed using NanoDrop, Bioanalyzer 2100, quantitative polymerase chain reaction and high-throughput sequencing. UC showed the lowest recovery of particles, but the highest protein purity for exosome isolation. For isolation of exoRNA, we found that combinations of the TEI and TER methods resulted in high extraction efficiency and purity of small RNA obtained using CCM. High yield and a narrow size distribution pattern of small RNA were shown in exoRNA isolated by exoRNeasy from serum. In RNA profile analysis, the small RNA constituent ratio, miRNA content and amount varied as a result of methodological differences. This study showed that different methods may introduce variations in the concentration, purity and size of exosomes and exoRNA. Herein we discuss the advantages and disadvantages of each method and their application to different materials, therefore providing a reference according to research design.

DNA hypomethylation and/or hypermethylation are presumed to be early events in carcinogenesis, and one or more DNA methyltransferases (DNMTs) have been suggested to play roles in carcinogenesis of gastric cancer (GC). However, there have been no systematic studies regarding the association between DNMT gene polymorphisms and GC risk. Here, we examined the associations of 16 single nucleotide polymorphisms (SNPs) from DNMT1 (rs2114724, rs2228611, rs2228612, rs8101866, rs16999593), DNMT2 (rs11695471, rs11254413), DNMT3A (rs1550117, rs11887120, rs13420827, rs13428812, rs6733301), DNMT3B (rs2424908, rs2424913, rs6087990) and DNMT3L (rs113593938) with GC in the Southern Chinese population. We assessed the associations of these 16 SNPs with GC in a case-control study that consisted of 242 GC cases and 294 controls, using the Sequenom MALDI-TOF-MS platform. Association analyses based on the χ2 test and binary logistic regression were performed to determine the odds ratio (OR) and 95% confidence interval (95%CI) for each SNP. We found that rs16999593 in DNMT1, rs11254413 in DNMT2 and rs13420827 in DNMT3A were significantly associated with GC susceptibility (OR 1.45, 0.15, 0.66, respectively; 95% CI 1.00–2.11, p = 0.047; 0.08–0.27, p < 0.01; 0.45–0.97, p = 0.034, respectively, overdominant model). These results suggested that DNMT1, DNMT2 and DNMT3A may play important roles in GC carcinogenesis. However, further studies are required to elucidate the mechanism.

Recently, a genome-wide association study of gastric cancer (GC) reported the significant association of seven genetic variants (rs4072037 and rs4460629 on 1q22; rs753724, rs11187842, rs3765524, rs2274223, and rs3781264 on 10q23) with GC in a Chinese population. These findings were confirmed in a subsequent independent study. However, it remains unknown whether these loci are associated with an increased risk of colorectal cancer (CRC). This study was to test whether the seven single nucleotide polymorphisms (SNPs) associated with GC were also associated with CRC in a Chinese population. The seven SNPs were genotyped using MassARRAY system. Allelic, genotypic, and haplotypic associations of the SNPs with CRC were investigated using χ(2) tests and logistic regression analysis. The SNPs rs3765524 and rs2274223, located on 10q23, were found to have significant protective effects against CRC, with equal odds ratios per allele. The two SNPs located on 1q22 (rs4072037 and rs4460629) showed a weak association with CRC. No significant association was identified with CRC for the remaining three SNPs located on 10q23 (rs753724, rs11187842, and rs3781264). These results suggest that rs3765524 and rs2274223 on 10q23 are associated with a protective effect against CRC in a Chinese population.

The missing values, widely existed in multivariate time series data, hinder the effective data analysis. Existing time series imputation methods do not make full use of the label information in real-life time series data. In this paper, we propose a novel semi-supervised generative adversarial network model, named SSGAN, for missing value imputation in multivariate time series data. It consists of three players, i.e., a generator, a discriminator, and a classifier. The classifier predicts labels of time series data, and thus it drives the generator to estimate the missing values (or components), conditioned on observed components and data labels at the same time. We introduce a temporal reminder matrix to help the discriminator better distinguish the observed components from the imputed ones. Moreover, we theoretically prove that, SSGAN using the temporal reminder matrix and the classifier does learn to estimate missing values converging to the true data distribution when the Nash equilibrium is achieved. Extensive experiments on three public real-world datasets demonstrate that, SSGAN yields a more than 15% gain in performance, compared with the state-of-the-art methods.

Micropeptides (≤100 amino acids) are essential regulators of physiological and pathological processes, which can be encoded by small open reading frames (smORFs) derived from long non-coding RNAs (lncRNAs). Recently, lncRNA-encoded micropeptides have been shown to have essential roles in tumorigenesis. Since translated smORF identification remains technically challenging, little is known of their pathological functions in cancer. Therefore, we created classifiers to identify translated smORFs derived from lncRNAs based on ribosome-protected fragment sequencing and machine learning methods. In total, 537 putative translated smORFs were identified and the coding potential of five smORFs was experimentally validated via green fluorescent protein-tagged protein generation and mass spectrometry. After analyzing 11 lncRNA expression profiles of seven cancer types, we identified one validated translated lncRNA, ZFAS1, which was significantly up-regulated in hepatocellular carcinoma (HCC). Functional studies revealed that ZFAS1 can promote cancer cell migration by elevating intracellular reactive oxygen species production by inhibiting nicotinamide adenine dinucleotide dehydrogenase expression, indicating that translated ZFAS1 may be an essential oncogene in the progression of HCC. In this study, we systematically identified translated smORFs derived from lncRNAs and explored their potential pathological functions in cancer to improve our comprehensive understanding of the building blocks of living systems