(HSD)-D-3, a dataset of Homo Sapiens Splice regions, and its extraction procedure from a major public database

摘要

The aim of this work is to describe a cleaning procedure of GenBank data, producing material to train and to assess the prediction accuracy of computational approaches for gene characterization. A procedure (GenBank2HS(3)D) has been defined, producing a dataset ((HSD)-D-3 - Homo Sapiens Splice Sites Dataset) of Homo Sapiens Splice regions extracted from GenBank (Rel.123 at this time). It selects, from the complete GenBank Primate Division, entries of Human Nuclear DNA according with several assessed criteria; then it extracts exons and introns from these entries (actually 4523 + 3802). Donor and acceptor sites are then extracted as windows of 140 nucleotides around each splice site (3799 + 3799). After discarding windows not including canonical GT-AG junctions (65 + 74), including insufficient data (not enough material for a 140 nucleotide window) (686 + 589), including not ACCT bases (29 + 30), and redundant (218 + 226), the remaining windows (2796 + 2880) are reported in the dataset. Finally, windows of false splice sites axe selected by searching canonical GT-AG pairs in not splicing positions (271937 + 332 296). The false sites in a range +/- 60 from a true splice site axe marked as proximal. HS3D, release 1.2 at this time, is available at the Web server of the University of Sannio: http://www.sci.unisannio.it/docenti/rampone/. [References: 29]