POLA-SEQ: A PIPELINE FOR THE ANALYSIS OF MRNA ISOFORMS FROM ALTERNATIVE POLYADENYLATION IN CANCER

GARCÍA SOLÁ M; MEDINA MV; COSO OA; ABBA MC; KORDON EC

Congreso; Reunión Anual SAIC; 2017

More than half of human genes contain alternative polyadenylation (pA) sites, which allow the creation of various 3? untranslated regions (3?UTRs). Alternative pA sites can be localized in internal introns or exons, allowing the formation of different protein isoforms, or they may be localized in the 3?UTR resulting in transcripts with variable 3? UTRs encoding the same protein. It has been reported that cancer cells often express substantial amounts of mRNA isoforms with shorter 3?UTRs than differentiated non transformed cells. It is also known than higher gene expression is tightly linked to cellular proliferation and, generally, short 3?UTR isoforms are more abundant in proliferating cells. Here, we report a pipeline named PolA-Seq to analyze the differential expression of genes than present alternati- ve pA sites based on RNA-seq data. This new tool allows the use of transcriptomic public data for searching specific mRNA isoforms involved in tumor development. The pipeline consists in the imple- mentation of Galaxy tools and its processing capability, R/Biocon- ductor and deeptools packages, for data analysis and visualization. This pipeline allowed us to identify different 3?UTRs of previously reported and unreported genes with alternative polyadenylation based on RNA-Seq data. With the use of the deeptools packages, we took an alignment of reads and generated a coverage track for each 3?UTR in the different genes of interest and compared their levels between normal differentiated mammary glands and mammary tumors. We have successfully tested this pipeline comparing the le- vels of short and long Beta 1 integrins mRNA isoforms in tumors Vs lactating mammary glands from RET transgenic mice. Our results confirm the overexpression of the shorter form, which has been previously discovered by our group, in mammary tumors. In addition, preliminary data allowed us to verify this pattern in other relevant genes with alternative pA sites.