Mapping alternative polyadenylation data with AppEnD. We analyzed PAS-Seq data (control and experimental sample) that was provided by Ami Ashar-Patel and Melissa Moore, as well as A-Seq data from Petar Grozdanov and Clint Macdonald. The PAS-Seq data were generated by sequencing from the 5′ end through the poly(A) tail (defined by the 20-nt dT primer used) into the anchor 3′ adapter incorporated as part of dT-priming for cDNA synthesis. The sites of poly(A) addition were mapped genome-wide. (A,B) The polyadenylation sites discovered by running AppEnD on PAS-Seq data. Poly(A) sites utilized by the EBAG9 gene (A) and the NET1 gene (B) are shown. The data from the control sample are at the top and the data for the experimental sample are at the bottom. Note that EBAG9 shows differential polyadenylation between control and experimental samples. (C) Example of PAS-Seq data. We required the length of the oligo(A) sequence that is nontemplated to be at least 4 nt to identify a polyadenylation site. Often the sequence of the 3′ adapter contained errors, presumably as a result of reading through the long oligo(A) stretch. (D,E) Poly(A) sites discovered using AppEnD on A-Seq data. The gene in (D) shows differential polyadenylation between control and experimental samples. (F,G) Examples of A-Seq reads containing (F) false positive poly(A) site due to adventitious priming and (G) a true poly(A) site.
