Landscape of antisense genes in the human genome and identification of new human hepatic antisense RNAs by long-read sequencing.

Background: Protein-coding genes have been considered the functional part of the genome, although they represent only 2% of the genome. In contrast, more than 90% of the genome produces non-coding RNA (ncRNA), including antisense (AS) genes, a type of long non-coding genes (encoding transcripts > 200 nucleotides) located on the opposite strand of coding genes. Therefore, antisense RNA (asRNA) can be complementary to the counterpart sense RNA, supporting a regulatory role with potential pathogenic consequences, as their deregulation has been associated with cardiovascular disease, cancer, and diabetes.

Results: We performed an in-depth review of AS genes in Ensembl and evaluated the expression of AS genes in human liver by third-generation RNA sequencing methods. Currently, 1656 AS genes overlapping with 1556 sense genes have been identified in the human genome. Coding genes with antisense counterparts were significantly larger and had higher transcriptional activity than genes without antisense counterparts. RNA nanopore sequencing of 15 human livers identified 185 transcripts (55 novel) from 150 known AS genes, and 1316 transcripts from 807 sense genes, with 111 sense-antisense pairs. Sense transcripts showed higher expression than antisense transcripts. Remarkably, RNA nanopore sequencing of human livers identified 146 transcripts (67 novel) corresponding to 118 possible novel AS genes.

Conclusion: This study reveals the landscape of AS genes in the human genome and demonstrates the power of long-read RNA sequencing to identify novel transcripts and even novel AS genes, exploring the relationship between sense and AS gene expression as well.