Project description:<p>High throughput RNA Sequencing has revealed that the human genome is widely transcribed. However, the extent of natural antisense transcription, the molecular mechanisms by which natural antisense transcripts (NATs) might affect their cognate sense genes, and the role of NATs in cancer are less well understood. Here, we use strand-specific paired-end RNA sequencing (ssRNASeq) on a cohort of 376 cancer patients covering 9 tissue types to comprehensively characterize the landscape of antisense expression. Our results reveal that greater than 60% of annotated transcripts have measureable antisense expression and the expression of sense and antisense transcript pairs is in general positively correlated. Furthermore, by studying the expression of sense/antisense pairs across tissues we identify lineage-specific, ubiquitous and cancer-specific antisense loci. Our results raise the possibility that NATs participate in the regulation of well-known tumor suppressors and oncogenes. Finally, this study provides a catalogue of cancer related genes with significant antisense transcription (oncoNAT). This resource will allow researchers to investigate the molecular mechanisms of sense/antisense regulation and further advance our understanding of their role in cancer.</p>
Project description:Bioinformatics powered correlative analysis of epigenomic patterns is an effective method to help derive biological hypotheses that can be tested genetically or biochemically. To accommodate the variety and complexity of epigenomic and transcriptomic patterns, ANchored COrrelative Patterns (ANCORP) was developed as a platform to integrate and intuitively visualize a large number of genome-wide profiles. With global profiles of 9 histone modifications mapped by ChIP-seq and a strand-specific RNA-seq dataset, we have applied the ANCORP-genetics pipeline for hypothesis building and testing in order to understand how global transcription may be regulated by epigenetic pathways such as histone modifications. It was found that intragenic antisense RNAs were depleted from genes with strong gene-body H3K36me2 mark and cytosine methylation enrichments but were significantly overrepresented in H3K4me3/H3K27me3 bivalent genes. Moreover, gene body H3K36me2 and DNA methylation anti-correlated with multiple active chromatin marks including H3K4me2/3, H3K9Ac and H3K18Ac. These observations lead us to hypothesize that H3K36me2 and DNA methylation may synergistically repress active chromatin marks in gene bodies and subsequently inhibit transcription of the antisense strand. Mutant analyses revealed that Polymerase Associated Factors (PAF) may be universally required for modulating NAT abundance whereas the role for the 5mC and H3K36me marks are more locus specific. H3K36me and PAF may either repress or permit the accumulation NATs depending on the chromatin state context in a particular transcription unit. Interestingly, the activation of antisense RNA in sdg8-2 or elf8-1 mutants does not associate with any increase of histone marks in gene bodies that are known to correlate with gene activation. Our results suggest that ANCORP-genetics is an effective approach to uncover epigenetic regulatory mechanisms by leveraging on the rapid advances in sequencing technologies and the resultant wealth of genome-wide information.
Project description:MicroRNAs (miRNAs) are 21-24 nucleotide (nt) small non-coding RNAs that regulate a wide variety of biological processes at the posttranscriptional level. MiRNA expression often exhibits spatial and temporal specificity. However, genome-wide miRNA expression patterns in different Arabidopsis organs during plant development have not yet been fully investigated. In this study, we sequenced 59 small RNA libraries generated from different tissue types at different developmental stages of Arabidopsis. We then re-annotated Arabidopsis miRNAs based on the most recent criteria. Global analysis of miRNA expression patterns showed that most miRNAs are ubiquitously expressed in different organs or tissues. But a small set of miRNAs, either previously annotated or newly identified, show highly specific expression patterns. In addition, the expression of some miRNA members belonging to the same family is strictly regulated spatially and temporally. Unexpectedly, we found that quite a few miRNAs are produced from different arms of their hairpin precursors at different developmental stages, suggesting that arm switching could be a general and important mechanism in developmental regulation.