Regulatory roles of microRNA163 in responses to stresses in Arabidopsis.

MicroRNAs (miRNAs) are small regulatory RNAs that participate in various biological processes by silencing target genes. In Arabidopsis, microRNA163 (miR163) was found to be involved in seed germination, root development, and biotic resistance. However, the regulatory roles of miR163 remain unclear. In the current study, the mir163 mutant was investigated to comprehensively understand and characterize its functions in Arabidopsis. RNA-sequencing and Gene Ontology enrichment analyses revealed that miR163 might be involved in "response to stimulus" and "metabolic process". Interestingly, "response to stress", including heat, cold, and oxidative stress, was enriched under the subcategory of "response to stimulus". We observed that miR163 and PXMT were repressed and induced under heat stress, respectively. Furthermore, the study detected significant differences in seed germination rate, hypocotyl length, and survival rate, indicating a variation in the thermotolerance between WT and mir163 mutant. The results revealed that the mir163 mutant had a lesser degree of germination inhibition by heat treatment than WT. In addition, the mir163 mutant showed a better survival rate and longer hypocotyl length under heat treatment than the WT. The metabolomes of WT and mir163 mutant were further analyzed. The contents of benzene derivatives and flavonoids were affected by miR163, which could enhance plants' defense abilities. In conclusion, miR163/targets regulated the expression of stress-responsive genes and the accumulation of defense-related metabolites to alter stress tolerance.

Discovering unknown human and mouse transcription factor binding sites and their characteristics from ChIP-seq data.

Transcription factor binding sites (TFBSs) are essential for gene regulation, but the number of known TFBSs remains limited. We aimed to discover and characterize unknown TFBSs by developing a computational pipeline for analyzing ChIP-seq (chromatin immunoprecipitation followed by sequencing) data. Applying it to the latest ENCODE ChIP-seq data for human and mouse, we found that using the irreproducible discovery rate as a quality-control criterion resulted in many experiments being unnecessarily discarded. By contrast, the number of motif occurrences in ChIP-seq peak regions provides a highly effective criterion, which is reliable even if supported by only one experimental replicate. In total, we obtained 2,058 motifs from 1,089 experiments for 354 human TFs and 163 motifs from 101 experiments for 34 mouse TFs. Among these motifs, 487 have not previously been reported. Mapping the canonical motifs to the human genome reveals a high TFBS density ±2 kb around transcription start sites (TSSs) with a peak at -50 bp. On average, a promoter contains 5.7 TFBSs. However, 70% of TFBSs are in introns (41%) and intergenic regions (29%), whereas only 12% are in promoters (-1 kb to +100 bp from TSSs). Notably, some TFs (e.g., CTCF, JUN, JUNB, and NFE2) have motifs enriched in intergenic regions, including enhancers. We inferred 142 cobinding TF pairs and 186 (including 115 completely) tethered binding TF pairs, indicating frequent interactions between TFs and a higher frequency of tethered binding than cobinding. This study provides a large number of previously undocumented motifs and insights into the biological and genomic features of TFBSs.

miRgo: integrating various off-the-shelf tools for identification of microRNA-target interactions by heterogeneous features and a novel evaluation indicator.

MicroRNAs (miRNAs) are short non-coding RNAs that regulate gene expression and biological processes through binding to messenger RNAs. Predicting the relationship between miRNAs and their targets is crucial for research and clinical applications. Many tools have been developed to predict miRNA-target interactions, but variable results among the different prediction tools have caused confusion for users. To solve this problem, we developed miRgo, an application that integrates many of these tools. To train the prediction model, extreme values and median values from four different data combinations, which were obtained via an energy distribution function, were used to find the most representative dataset. Support vector machines were used to integrate 11 prediction tools, and numerous feature types used in these tools were classified into six categories-binding energy, scoring function, evolution evidence, binding type, sequence property, and structure-to simplify feature selection. In addition, a novel evaluation indicator, the Chu-Hsieh-Liang (CHL) index, was developed to improve the prediction power in positive data for feature selection. miRgo achieved better results than all other prediction tools in evaluation by an independent testing set and by its subset of functionally important genes. The tool is available at http://predictor.nchu.edu.tw/miRgo.