Pan-cancer pseudogene RNA analysis reveals a regulatory network promoting cancer cell proliferation

Pseudogenes have long been thought to be nonfunctional due to their lack of protein-coding ability, compared to protein-coding genes (PCG). Actually, pseudogenes can transcribe functional long non-coding RNAs (lncRNAs) involved in cancer development and progression. These lncRNAs may regulate mRNA levels as competing endogenous RNAs (ceRNAs). However, a systematic pan-cancer analysis of pseudogene ceRNA networks is still lacking. Here, we curated 9455 pseudogenes and constructed ceRNA networks for 14 cancer types. We discovered that pseudogenes in ceRNA networks were the most cancer type-specific and that ~ 20% cePCGs were cancer hallmark genes, supporting the close relationship between pseudogenes and cancer. Notably, in breast cancer (BRCA), we found that the ceRNA subnetwork of ZNF252P/AL390726.5 (pseudogenes)-miRNAs-ESR1 may enhance the proto-oncogene MYC and we experimentally validated oncogenic effects of the two pseudogenes. Moreover, pseudogene-based subtyping of BRCA tumors revealed a new subtype featured by immunoglobulin pseudogenes. Collectively, our findings could pave the way for more pseudogene research in cancer. We provided our results in a user-friendly database, cePseudo, available at http://bioinfo-sysu.com/cePseudo2 .