A computational scan for U12-dependent introns in the human genome sequence

Abstract

U12-dependent introns are found in small numbers in most eukaryotic genomes, but their scarcity makes accurate characterisation of their properties challenging. A computational search for U12-dependent introns was performed using the draft version of the human genome sequence. Human expressed sequences confirmed 404 U12-dependent introns within the human genome, a 6-fold increase over the total number of non-redundant U12-dependent introns previously identified in all genomes. Although most of these introns had AT-AC or GT-AG terminal dinucleotides, small numbers of introns with a surprising diversity of termini were found, suggesting that many of the non-canonical introns found in the human genome may be variants of U12-dependent introns and, thus, spliced by the minor spliceosome. Comparisons with U2-dependent introns revealed that the U12-dependent intron set lacks the 'short intron' peak characteristic of U2-dependent introns. Analysis of this U12-dependent intron set confirmed reports of a biased distribution of U12-dependent introns in the genome and allowed the identification of several alternative splicing events as well as a surprising number of apparent splicing errors. This new larger reference set of U12-dependent introns will serve as a resource for future studies of both the properties and evolution of the U12 spliceosome.

Figure 1

Length of U12- and U2-dependent introns. The length of 168 U12-dependent introns and 11 402 RefSeq-confirmed U2-dependent introns <1 kb in length are plotted. Grey bars represent the counts of U12-dependent introns grouped into 50-bp wide bins, while the black line represents the frequency of U2-dependent introns grouped into 10-bp wide bins.

Figure 2

Branch site to acceptor site distance for U12-dependent introns. The distance between the branch site and the acceptor site is plotted for 108 AT-AC U12-dependent introns (black bars) and 275 GT-AG U12-dependent introns (grey bars).