Human granulocyte transcriptome annotation and analysis reveals increased expression variability for lncRNAs compared to mRNAs in healthy individuals

In this study we used human primary granulocytes to assess natural variability of lncRNA expression. We used polyA+ granulocyte RNA-seq from 10 healthy individuals to define granulocyte lncRNA transcriptome, which was not available before. We annotated 6,249 lncRNA transcripts forming 1,591 lncRNA loci, contributing 268 novel loci to the human genome annotation. We show that focusing on granulocytes allows identification of less well expressed, less efficiently spliced and more granulocyte specific lncRNAs. We used Ribosomal depleted granulocyte RNA-seq from 7 healthy individuals sampled in 3 replicates to estimate reproducibility and variability of lncRNA expression and found that though being as well reproducible between replicates, lncRNAs are significantly more variable in their expression than mRNAs. We confirmed this finding in publicly available Geuvadis lymphoblastoid cell line (LCL) RNA-seq dataset from 462 individuals and also used this dataset to show that the number of identified lncRNA increases with the number of donors analyzed. Overall, we showed that high variability of lncRNAs is an important feature that distinguishes them from mRNAs and influences their identification strategy, making it necessary to include as many individuals as possible into the identification effort for every given cell type in order to comprehensively annotate all the lncRNAs in the human genome.