Fig. 1

Flowchart of identification of lncRNAs responsive to Pi starvation in Arabidopsis. a 1. Plant treatment and poly(A) + and poly(A)– RNA extractions and purifications. 2. Construction of strand-specific cDNA libraries and sequencing. 3. RNA-Seq data mapping and assembly. 4. Novel lncRNAs were obtained after three filter steps, including overlap with annotation, calculation of length, and coding potential. 5. Characterization of novel lncRNAs at different levels, such as transcript length, exon number, polyadenylation, expression, epigenetic signature, and transcriptional and post-transcriptional regulation. b Assembled ratio of protein-coding transcripts and lncRNA transcripts. Approximately 90 % of protein-coding transcripts and over 80 % of TAIR10 lncRNAs could be completely assembled based on our RNA-Seq data. c Genomic positions of TAIR10 and novel lncRNAs. The majority of TAIR10 and novel lncRNAs were antisense to coding transcripts, and lncRNAs overlapped with TEs or pseudogenes accounted for a small fraction. Other lncRNAs with no overlap with any annotated coding transcripts or lncRNA were defined as intergenic or cis-lncRNAs according to the distance between lncRNAs and adjacent genes