Long noncoding RNAs (lncRNAs) constitute the majority of transcripts in the mammalian genomes, and yet, their functions remain largely unknown. As part of the FANTOM6 project, we systematically knocked down the expression of 285 lncRNAs in human dermal fibroblasts and quantified cellular growth, morphological changes, and transcriptomic responses using Capped Analysis of Gene Expression (CAGE). Antisense oligonucleotides targeting the same lncRNAs exhibited global concordance, and the molecular phenotype, measured by CAGE, recapitulated the observed cellular phenotypes while providing additional insights on the affected genes and pathways. Here, we disseminate the largest-todate lncRNA knockdown data set with molecular phenotyping (over 1000 CAGE deep-sequencing libraries) for further exploration and highlight functional roles for ZNF213-AS1 and lnc-KHDC3L-2.

Long non-coding RNAs (lncRNAs) constitute the majority of transcripts in the mammalian genomes and yet, their functions remain largely unknown. We systematically knockdown 285 lncRNAs expression in human dermal fibroblasts and quantified cellular growth, morphological changes, and transcriptomic responses using Capped Analysis of Gene Expression (CAGE). Antisense oligonucleotides targeting the same lncRNA exhibited global concordance, and the molecular phenotype, measured by CAGE, recapitulated the observed cellular phenotypes while providing additional insights on the affected genes and pathways. Here, we disseminate the largest to-date lncRNA knockdown dataset with molecular phenotyping (over 1,000 CAGE deep-sequencing libraries) for further exploration and highlight functional roles for ZNF213-AS1 and lnc-KHDC3L-2.

Long non-coding RNAs or lncRNAs are a class of non-protein-coding RNAs that are >200 nucleotides in length. Almost 50% of lncRNAs during zebrafish development are transcribed in an antisense direction to a protein-coding gene. However, the role of these Natural Antisense Transcripts or NATs during development remains enigmatic. To understand NATs in early vertebrate development, we took a computational biology approach and analyzed existing as well as novel datasets. Our analysis indicates that zebrafish NATs can be divided into two major classes based on their co-expression patterns with respect to the overlapping protein-coding genes. Group-1 NATs have characteristics similar to maternally deposited RNAs in that their levels decrease as development progresses. Group-1 NAT levels are negatively correlated with that of overlapping sense-strand protein-coding genes.Group-2 NATs, on the other hand, are co-expressed with overlapping protein-coding genes. In contrast to group-1, which is enriched in genes involved in developmental pathways, group-2 proteincoding genes are enriched in house-keeping functions. Group-1 NATs also show larger overlap and higher complementarity with the sense-strand mRNAs as compared to other NATs. In addition, our transcriptomics data, quantifying RNA levels from cytoplasmic and nuclear compartments, indicates that group-1 NATs are more populated in the cytosol. Based on their expression pattern, cytosolic nature and their higher complementarity to the overlapping developmental mRNAs, we speculate that group-1 NATs function post-transcriptionally to silence spurious expression of developmental genes.

The authors would like to correct Figure 3, panel J, in which the rightmost upper image of SA-β-gal stained 293T cells following short hairpin RNA (shRNA)-mediated knockdown of RRAS2 with sh769 (RRAS2-KD-sh769) was inadvertently, and due to a labeling error, taken from the same original source image presented in the middle upper panel, which shows increased SA-β-gal activity following RRAS2 knockdown by a different shRNA (sh646). This correction does not affect any of the conclusions of the article. The corrected image representative of RRAS2-KD-sh769 is provided below, and Figure 3 has been updated in the article online. Additionally, the authors have provided a revised Supplemental Figure S7 file in which the redundant successive Supplemental figure files have been removed. This can be found in the Revised Supplemental Material online.

Long non-coding RNAs or lncRNAs are a broad class of non-protein coding RNAs that are >200nucleotides in length. A number of lncRNAs are shown to play an important role in gene expression regulation. LncRNAs antisense to a protein-coding gene can act either as positive or negative regulators of overlapping protein-coding mRNAs. Almost 50% of lncRNAs present during development of vertebrates such as zebrafish are of antisense lncRNA class. However, their role in gene expression regulation during development remains enigmatic. To understand the role of antisense lncRNAs in early vertebrate development, we took a computational biology approach to analyze existing as well as novel dataset. Our analysis of RNA sequencing data from zebrafish development indicates that antisense RNAs can be divided into two major classes based on their positive or negative co-expression patterns to the sense protein-coding genes. The ones with negative co-expression patterns or group-1 are maternal antisense lncRNAs that overlap mainly developmental genes. Group-2 with positive expression pattern overlap mainly house-keeping genes.Group-1 antisense lncRNAs are longer and more stable as compared to antisense lncRNAs in group-2. In addition, to answer if antisense RNAs in the two groups are differently localized in cell compartments, we deep-sequenced RNA from cytoplasmic and nuclear compartments during early developmental stages. The analysis of these compartment specific datasets revealed group-1 lncRNAs are cytosolic. Based on the cytosolic nature of group-1 RNAs and their higher complementarity to the overlapping developmental mRNAs, we speculate that the group-1 RNAs might function similar to microRNAs in silencing spurious expression of developmental genes. Group-1 and group-2 RNAs are also distinct in terms of their genomic configuration, conservation, length and transcriptional regulation. These results are not only important in understanding the role of antisense RNAs in development but also for predicting the nature of association between antisense lncRNA and overlapping protein-coding genes. ACCESSION NUMBERSThe RNA-seq and CAGE-seq data generated in this study is deposited in GEO database under accession numbers GSE143208 and GSE144040 respectively. SUPPLEMENTARY DATASupplementary Data are available online. ACKNOWLEDGEMENTWe thank Prof Ferenc Mueller and his lab for useful discussions and help with zebrafish methods.We are grateful to all members of ZENCODE ITN for critical comments on the work. We also thank