Back

Transposable elements contribute to regulatory hubs stress-related long noncoding RNAs in Maize

Lv, Y.; Hu, F.; Zhou, Y.; Wu, F.; Zhou, L.; Gaut, B. S.

2019-10-11 plant biology
10.1101/801704 bioRxiv
Show abstract

Several studies have mined short-read RNA sequencing datasets to identify lncRNAs, and others have focused on the function of individual lncRNA in abiotic stress response. However, our understanding of the complement, function and origin of long-non-coding RNA (lncRNAs) response to abiotic stress, especially transposon derived lncRNA (TE-lncRNA), is still in its infancy. To discover and study lncRNAs in maize (Zea mays ssp. mays), we utilized a dataset of 127 RNA sequencing samples that included PacBio fl-cDNA and total RNA-Seq datasets. Overall, we identified 23,309 candidate lncRNAs, 60% of which were identified in polyadenylated (polyA+) samples. The majority (65%) of the 23,309 lncRNAs had sequence similarity to transposable elements (TEs). Most had similarity to long-terminal-repeat retrotransposons from the Copia and Gypsy superfamilies, representing the high proportion of these elements in the genome, but class II, DNA transposons were enriched for lncRNAs relative to their genomic representation by 2-fold. By assessing the fraction of lncRNAs that respond to abiotic stresses like heat, cold, salt and drought, we identified 1,077 differentially expressed lncRNA transcripts. Their expression was correlated (r2=0.48) with their nearest gene, suggesting that lncRNAs are subject to some of the cis regulatory features as neighboring genes. By inferring co-expression networks across our large dataset, we found that 39 lncRNAs act as major hubs in co-expression networks, of which 18 appeared to be derived from TEs. These results suggest that lncRNAs, especially TE-lncRNAs, may play key regulatory roles in moderating abiotic responses.

Matching journals

The top 8 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.